[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-vision-transformers-beat-cnns-at-catching-deepfakes":10,"sections":34},{"siteName":4,"siteTagline":5,"publisherName":4,"contactEmail":6},"The Revision","Tech news, decoded.","editor@therevision.news",{"gaMeasurementId":8,"adsenseClientId":9},"G-ZW2MV82GYR","ca-pub-8533917693782264",{"article":11},{"id":12,"slug":13,"title":14,"dek":15,"body_md":16,"tags_json":17,"published_at":18,"created_at":19,"updated_at":20,"status":21,"review_note":22,"review_notes":23,"image_url":22,"persona_id":22,"persona_name":22,"section":24,"tags":25,"sources":29,"feedback":33,"feedback_at":22,"cost_usd":33,"total_tokens":33},4219,"vision-transformers-beat-cnns-at-catching-deepfakes","Vision Transformers Beat CNNs at Catching Deepfakes","An ensemble of fine-tuned vision transformers hit 96.77% AUC on a tough deepfake benchmark, outpacing the previous best detector by over seven points.","A research team has built a deepfake image detector that generalizes better than existing tools — and won a major competition doing it.\n\nThe approach combines three fine-tuned vision transformers — DINOv2, AIMv2, and OpenCLIP's ViT-L\u002F14 — into an ensemble trained on the DF-Wild dataset, which was released for the IEEE SP Cup 2025 challenge. That dataset was chosen specifically because it covers a wide range of manipulation and generation techniques, the kind of variety that tends to break detectors trained on narrower data. The ensemble hit an AUC of 96.77% and an Equal Error Rate of 9% on the DF-Wild test set, beating the prior state-of-the-art detector, Effort, by 7.05 percentage points in AUC and 8 points in EER. The solution took first place at ICASSP 2025.\n\nGeneralization is the core problem in deepfake detection right now. Most classifiers learn to spot artifacts from whatever generator produced their training data, then fall apart when a new model enters the picture. Vision transformers, with their attention-based global feature extraction, appear better suited to capturing manipulation signals that survive across different generation techniques than convolutional networks trained on spatial features alone. A 7-point AUC gap over the previous best is a meaningful margin, not a rounding error.\n\nThe field has been here before — a new detector posts strong benchmark numbers, then the generative models move on and the gap closes. Whether this ensemble holds up against the next wave of generators is the real test, and that answer won't come from a leaderboard.","[\"deepfakes\",\"computer vision\",\"ai\",\"security\"]","2026-07-07T04:00:00.000Z","2026-07-07T20:13:14.582Z","2026-07-07T20:13:17.486Z","published",null,[],"ai",[26,27,24,28],"deepfakes","computer vision","security",[30],{"name":31,"url":32},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2604.17376",0,{"sections":35},[36,40,44,49,54,59,64,69,74,78,83,87,92,97],{"name":37,"slug":24,"count":38,"latest_published_at":39},"AI",2590,"2026-07-16T04:00:00.000Z",{"name":41,"slug":28,"count":42,"latest_published_at":43},"Security",294,"2026-07-15T19:59:48.000Z",{"name":45,"slug":46,"count":47,"latest_published_at":48},"Deals","deals",179,"2026-06-29T20:02:07.000Z",{"name":50,"slug":51,"count":52,"latest_published_at":53},"Policy","policy",158,"2026-07-16T00:02:48.000Z",{"name":55,"slug":56,"count":57,"latest_published_at":58},"Hardware","hardware",122,"2026-07-14T19:46:26.000Z",{"name":60,"slug":61,"count":62,"latest_published_at":63},"Consumer Tech","consumer-tech",93,"2026-07-13T13:20:48.000Z",{"name":65,"slug":66,"count":67,"latest_published_at":68},"Software","software",70,"2026-07-13T19:52:25.000Z",{"name":70,"slug":71,"count":72,"latest_published_at":73},"Science","science",66,"2026-07-10T10:29:37.000Z",{"name":75,"slug":76,"count":77,"latest_published_at":18},"Dev Tools","dev-tools",59,{"name":79,"slug":80,"count":81,"latest_published_at":82},"Gaming","gaming",41,"2026-07-09T04:00:00.000Z",{"name":84,"slug":85,"count":81,"latest_published_at":86},"Startups","startups","2026-06-29T20:55:50.000Z",{"name":88,"slug":89,"count":90,"latest_published_at":91},"General","general",29,"2026-07-10T22:28:58.000Z",{"name":93,"slug":94,"count":95,"latest_published_at":96},"Reviews","reviews",20,"2026-06-24T12:00:01.000Z",{"name":98,"slug":99,"count":100,"latest_published_at":101},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]