FrameForge：利用视频基础模型先验实现统一图像生成
原标题：Paper page - FrameForge: Harnessing Priors of Video Foundation Models for Unified Image Generation
AI 摘要
研究者提出 FrameForge，一种利用预训练视频基础模型先验的统一条件图像生成方法。该方法通过混合时间专家（MoTE）动态选择不同虚拟间隔的时间 RoPE 专家，并用时间推理链（CoTR）建模条件到目标的潜在中间状态转换。在 DreamBench、COCO2017 和 GEdit-Bench 上的实验显示，其在主体驱动生成、canny 到图像生成和图像编辑任务上优于对比方法。
正文节选
FrameForge: Harnessing Priors of Video Foundation Models for Unified Image Generation Abstract FrameForge enables adaptive conditional image generation by modeling variable condition-target relationships through dynamic temporal experts and intermediate reasoning pathways in pretrained video models. Unified conditional image generation requires a single model to preserve different aspects of a visual input while producing a target image that follows textual guidance. Pretrained video models prov