DTG-Restore：面向生成式视频超分辨率的免训练扩散精炼
DTG-Restore: Training-Free Diffusion Refinement for Generative Video Super-Resolution
提出解耦时间引导（DTG），在采样时让无条件分支在更干净的扩散时间步上评估，为视频复原提供可复用的无训练先验利用手段，可迁移到其他基于扩散的复原任务。
原文依据 Our proposed Decoupled Time Guidance (DTG) evaluates the unconditional branch at a cleaner diffusion timestep
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论近期视频扩散模型的进展实现了显著的生成保真度，然而利用这些先验进行复原仍受限于标准无分类器引导中条件分支与无条件分支之间的强耦合。我们引入一个免训练框架，通过在时间上解耦这些信号来增强失真和低分辨率视频。我们提出的解耦时间引导（DTG）在更干净的扩散时间步上评估无条件分支，提供一种前瞻先验，在保持几何结构的同时抑制扭曲内容的复制。这种时间偏置在整个采样过程中退火，使模型无需重新训练即可从结构校正过渡到细节精炼。与任何现成的复原模块以即插即用方式结合，我们的方法在AI生成和真实世界视频中均提升了感知连贯性并恢复了合理的结构。为便于评估，我们构建了GenWarp480，一个由多种文本到视频模型合成的4,400个失真480p视频的基准。GenWarp480聚焦于典型的生成退化，如扭曲的人脸、身体错位和空间伪影，为评估对生成错误的鲁棒性提供了专门构建的测试平台。大量实验表明，我们的方法在无需任何模型训练的情况下，在结构保真度和时序稳定性方面取得了显著改进。
英文原文摘要 1445 字符
Recent progress in video diffusion models has enabled remarkable generative fidelity, yet leveraging these priors for restoration remains limited by the strong coupling between conditional and unconditional branches in standard classifier-free guidance. We introduce a training-free framework that enhances distorted and low-resolution videos by decoupling these signals in time. Our proposed Decoupled Time Guidance (DTG) evaluates the unconditional branch at a cleaner diffusion timestep, providing a lookahead prior that preserves geometry while suppressing replication of warped content. This temporal bias is annealed throughout sampling, allowing the model to transition from structure correction to detail refinement without retraining. Combined with any off-the-shelf restoration module in a plug-and-play manner, our approach improves perceptual coherence and restores plausible structure in AI-generated and real-world videos alike. To facilitate evaluation, we curate GenWarp480, a benchmark of 4,400 distorted 480p videos synthesized from diverse text-to-video models. GenWarp480 focuses on characteristic generative degradations such as warped faces, body misalignments, and spatial artifacts, providing a purpose-built testbed for assessing robustness to generative errors. Extensive experiments demonstrate that our method achieves significant improvements in structural fidelity and temporal stability without any model training.