Multi-Modal Phishing Website Detection with Real-Time Rendering Signals and TLS Fingerprints

Multi-Modal Phishing Website Detection with Real-Time Rendering Signals and TLS Fingerprints | 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 Multi-Modal Phishing Website Detection with Real-Time Rendering Signals and TLS Fingerprints Emily K. Dawson, Sophie L. Cartwright, James R. Whitfield This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9460394/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 Phishing website detection often relies solely on lexical or HTML features, which makes classifiers fragile against obfuscated URLs and template-based page cloning. This study develops a multi-modal detection framework that combines URL lexical embeddings, DOM structural features, page rendering signals (such as font entropy and visual layout similarity), and TLS certificate fingerprints. We build a dataset of 950,000 URLs, including 160,000 confirmed phishing instances collected from browser telemetry and public feeds over nine months. Character-level CNNs encode URLs, while a gradient boosting model integrates DOM and TLS features. A small Siamese CNN compares rendered screenshots with a benign-template bank to capture near-duplicate phishing pages. The framework achieves an AUC of 0.987, recall of 95.4%, and reduces false positives by 19.3% compared with a strong lexical-only baseline. Online experiments in a proxy-based deployment show that median detection latency remains below 20 ms per request. The results indicate that combining transport-layer fingerprints and rendering behavior yields robust, real-time phishing detection suitable for production environments. Artificial Intelligence and Machine Learning phishing website detection multi-modal learning TLS fingerprinting visual similarity gradient boosting CNN Full Text Additional Declarations The authors declare no competing interests. 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-9460394","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":625650614,"identity":"251b4bc5-5c74-4fd1-b65c-425be9b92ff9","order_by":0,"name":"Emily K. Dawson","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABCklEQVRIiWNgGAWjYBACAxQe45//zPxA+sAD4rU0MLNLNgC1JJCihd/gAJCBT4s5+/FnD34w3JGTj0h//OHjDjZp42uHHwJtsZPTbcCuxbInx9ywh+GZseGNHDPJmWd4jM1upxkAtSQbmx3A4bADOWwSPAyHEzfOyGFj5mGTSDa7nQDSciBxGy4t558/k/zDcLh+44z0x5//sBnUb56d/gG/lhsJZtJAWxLkJRIMpBnbEpgNpHMI2HLjjZm0DMNhww08b8wke84cYJa4nVNwIMEAj1/Opz+TfMNwWF6+HRhiPyoOMPPPTt/84UOFnRwuLWDA+A8UDqhG4VEOA/INRCgaBaNgFIyCkQkAqsFjRsj/Z5wAAAAASUVORK5CYII=","orcid":"","institution":"School of Informatics, The University of Edinburgh, Edinburgh EH8 9AB, United Kingdom","correspondingAuthor":true,"prefix":"","firstName":"Emily","middleName":"K.","lastName":"Dawson","suffix":""},{"id":625650615,"identity":"97187775-533b-4737-91cf-0335e53a69cf","order_by":1,"name":"Sophie L. Cartwright","email":"","orcid":"","institution":"School of Informatics, The University of Edinburgh, Edinburgh EH8 9AB, United Kingdom","correspondingAuthor":false,"prefix":"","firstName":"Sophie","middleName":"L.","lastName":"Cartwright","suffix":""},{"id":625650616,"identity":"07bd9ddf-9396-4357-b04a-2b4cce6bde0e","order_by":2,"name":"James R. Whitfield","email":"","orcid":"","institution":"School of Informatics, The University of Edinburgh, Edinburgh EH8 9AB, United Kingdom","correspondingAuthor":false,"prefix":"","firstName":"James","middleName":"R.","lastName":"Whitfield","suffix":""}],"badges":[],"createdAt":"2026-04-19 08:14:08","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-9460394/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9460394/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":107488618,"identity":"159136ff-0e41-4b8e-ba5a-d716862633b9","added_by":"auto","created_at":"2026-04-22 02:45:19","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":250067,"visible":true,"origin":"","legend":"","description":"","filename":"01MultiModalPhishingWebsiteDetectionwithRealTimeRenderingSignalsandTLSFingerprints.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9460394/v1_covered_3fe64bb3-9a4a-4c06-9573-fc3858a48f9e.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eMulti-Modal Phishing Website Detection with Real-Time Rendering Signals and TLS Fingerprints\u003c/strong\u003e\u003c/p\u003e","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"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":"phishing website detection, multi-modal learning, TLS fingerprinting, visual similarity, gradient boosting, CNN","lastPublishedDoi":"10.21203/rs.3.rs-9460394/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9460394/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003ePhishing website detection often relies solely on lexical or HTML features, which makes classifiers fragile against obfuscated URLs and template-based page cloning. This study develops a multi-modal detection framework that combines URL lexical embeddings, DOM structural features, page rendering signals (such as font entropy and visual layout similarity), and TLS certificate fingerprints. We build a dataset of 950,000 URLs, including 160,000 confirmed phishing instances\u0026nbsp;collected from browser telemetry and public feeds over nine months. Character-level CNNs encode URLs, while a gradient boosting model integrates DOM and TLS features. A small Siamese CNN compares rendered screenshots with a benign-template bank to capture near-duplicate phishing pages. The framework achieves an AUC of 0.987, recall of 95.4%, and reduces false positives by 19.3%\u0026nbsp;compared with a strong lexical-only baseline. Online experiments in a proxy-based deployment show that median detection latency remains below 20 ms\u0026nbsp;per request. The results indicate that combining transport-layer fingerprints and rendering behavior yields robust, real-time phishing detection suitable for production environments.\u003c/p\u003e","manuscriptTitle":"Multi-Modal Phishing Website Detection with Real-Time Rendering Signals and TLS Fingerprints","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-21 10:51:06","doi":"10.21203/rs.3.rs-9460394/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":"83c3f505-cd0e-483a-8dc8-6b72ae5a9b86","owner":[],"postedDate":"April 21st, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":66592195,"name":"Artificial Intelligence and Machine Learning"}],"tags":[],"updatedAt":"2026-04-21T10:51:06+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-21 10:51:06","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9460394","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9460394","identity":"rs-9460394","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}