[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-a07a31d6-3a63-4452-b898-02a6e575682e":3},{"id":4,"title":5,"summary":6,"original_url":7,"source_id":8,"tags":9,"published_at":23,"created_at":24,"modified_at":25,"is_published":26,"publish_type":27,"image_url":13,"view_count":28},"a07a31d6-3a63-4452-b898-02a6e575682e","Vera：Netflix × Caltech 把视频编辑拆成「编辑层 + 原视频」，Mixture-of-Transformers 守住内容保真","Vera 是 Netflix 与加州理工学院联合发布的分层扩散视频编辑框架，arXiv:2606.23610，2026 年 6 月 22 日上线。核心创新在于把\"要编辑的内容\"和\"要保留的像素\"分开建模：编辑层、alpha 蒙版、合成视频各自由独立 DiT 编码，再通过联合自注意力实现跨层一致性。配合 48.6 万帧分层训练集（合成 + 真实单目标 + 多目标带阴影\u002F反射\u002F遮挡），Vera-14B 在 PSNR\u002FSSIM\u002FLPIPS 等内容保护指标上对 VACE、Ditto、Lucy-Edit 等开源基线实现 2-3 倍量级领先，用户研究也获得显著偏好优势。这套思路首次把\"可控编辑\"和\"像素级保护\"统一在单一生成框架内。","https:\u002F\u002Farxiv.org\u002Fabs\u002F2606.23610","7437aeb9-930c-4866-a2e9-48003c1a792b",[10,14,17,20],{"id":11,"name":12,"slug":12,"description":13,"color":13},"7b67033c-19e6-4052-a626-e681bba64c7a","diffusion",null,{"id":15,"name":16,"slug":16,"description":13,"color":13},"499f4b56-819d-49a3-9609-33e775143b86","multimodal",{"id":18,"name":19,"slug":19,"description":13,"color":13},"4f214978-cac1-4f39-aa4b-f92a0d0934b7","transformer",{"id":21,"name":22,"slug":22,"description":13,"color":13},"ebe5dcd1-46b1-4298-b8c2-8e0e2f456e56","video-generation","2026-06-24T04:00:00Z","2026-06-24T04:08:58.153889Z","2026-06-24T04:08:58.153899Z",true,"agent",3]