[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-spackle-fixes-the-blind-spots-in-ai-generated-3d-scenes":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},8262,"spackle-fixes-the-blind-spots-in-ai-generated-3d-scenes","Spackle Fixes the Blind Spots in AI-Generated 3D Scenes","A new technique patches the blurry gaps that appear when AI stretches a single photo into a wide-angle 3D view, without slowing down rendering.","A new method called Spackle tackles a stubborn problem in AI-generated 3D scenes: filling in the parts of a photo the camera never captured.\n\nResearchers describe Spackle, a residual learning framework for single-image novel view synthesis, the technique that lets software generate a photorealistic 3D view of a scene from just one photo. Existing hybrid systems that pair feedforward 3D Gaussian Splatting with diffusion models run into what the paper calls capacity competition: they only get a fixed budget of Gaussians to work with, so rendering newly visible areas steals resources from the parts of the scene the camera already saw, and quality drops as the requested viewpoint strays further from the original shot. Spackle works in three steps: it first predicts a baseline set of 3D Gaussian attributes from the input view, then automatically flags the regions that came out poorly, and finally trains a second, residual set of Gaussians focused only on fixing those weak spots. At render time, the system blends the baseline and residual Gaussians into one output.\n\nThis is a narrow but real bottleneck for anything that needs to spin a single photo into a usable 3D view - AR previews, robotics simulation, even quick property or product walkthroughs. Instead of scaling up the whole model or throwing more compute at every pixel, Spackle spends extra effort only where the reconstruction is weakest, which is the kind of targeted fix that tends to survive contact with real products.\n\nThe paper reports state-of-the-art results on large-view-deviation cases, but that's Spackle's own benchmark suite - worth watching for independent comparisons before anyone builds a product roadmap around it.","[\"ai\",\"computer-vision\",\"3d-rendering\",\"gaussian-splatting\"]","2026-09-28T04:00:00.000Z","2026-09-28T20:18:04.130Z","2026-09-28T20:18:11.040Z","published",null,[],"ai",[24,26,27,28],"computer-vision","3d-rendering","gaussian-splatting",[30],{"name":31,"url":32},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.30941",0,{"sections":35},[36,40,45,50,55,60,65,70,75,80,85,90,94,99],{"name":37,"slug":24,"count":38,"latest_published_at":39},"AI",4877,"2026-09-28T16:14:54.000Z",{"name":41,"slug":42,"count":43,"latest_published_at":44},"Security","security",764,"2026-09-28T12:00:00.000Z",{"name":46,"slug":47,"count":48,"latest_published_at":49},"Policy","policy",401,"2026-09-28T15:14:48.000Z",{"name":51,"slug":52,"count":53,"latest_published_at":54},"Deals","deals",268,"2026-09-28T14:00:00.000Z",{"name":56,"slug":57,"count":58,"latest_published_at":59},"Hardware","hardware",191,"2026-09-28T15:45:00.000Z",{"name":61,"slug":62,"count":63,"latest_published_at":64},"Science","science",154,"2026-09-28T13:19:18.000Z",{"name":66,"slug":67,"count":68,"latest_published_at":69},"Consumer Tech","consumer-tech",138,"2026-09-28T14:13:01.000Z",{"name":71,"slug":72,"count":73,"latest_published_at":74},"Software","software",91,"2026-09-25T20:55:00.000Z",{"name":76,"slug":77,"count":78,"latest_published_at":79},"Dev Tools","dev-tools",85,"2026-09-28T13:00:00.000Z",{"name":81,"slug":82,"count":83,"latest_published_at":84},"Startups","startups",78,"2026-09-28T15:29:21.000Z",{"name":86,"slug":87,"count":88,"latest_published_at":89},"Gaming","gaming",48,"2026-09-25T18:35:21.000Z",{"name":91,"slug":92,"count":88,"latest_published_at":93},"General","general","2026-09-26T17:02:42.000Z",{"name":95,"slug":96,"count":97,"latest_published_at":98},"Reviews","reviews",30,"2026-09-24T20:07:31.000Z",{"name":100,"slug":101,"count":102,"latest_published_at":103},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]