生成之前先对齐图像
Align Images Before You Generate
CorrAdapter 用扩散模型中间特征自建跨图像对应关系，并通过旁路分支只聚合匹配区域的信息，为多视图/视频扩散提供了一种免训练、与基线无关的时空一致性增强手段，可直接插入现有 transformer 主干。
原文依据 CorrAdapter designs a bypass branch for transformer blocks in the multi-image diffusion model
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论多图像扩散模型能够生成多视图或视频等图像，以描述静态或动态场景，但纹理与结构漂移依然存在，严重削弱了时空一致性。解决这一问题仍然困难，尤其是在纯生成式推理过程中没有任何外部几何或语义先验的情况下。本文提出 CorrAdapter，一种即插即用的适配器，它发现并利用多图像扩散自身的一种内在特性，在所有输出图像真正被生成之前就完成对齐。具体而言，CorrAdapter 为多图像扩散模型中的 Transformer 模块设计了一条旁路分支，其中包含一个原生对应关系构建器，从扩散模型的中间特征中建立可靠的对应关系；以及一个对齐区域聚合器，仅整合来自匹配区域的消息，以避免模糊的信息交互。以原生对应关系作为引导，CorrAdapter 无需任何辅助输入即可增强时空一致性，并且保持免训练、与基线无关，从而能够无缝泛化到各类生成任务。此外，我们还提供了一种可选的训练方案，以探索进一步提升的可能性。在静态多视图生成和动态视频生成上的实验表明，CorrAdapter 在强基线之上持续提升时空一致性与感知质量，为几何上忠实的多图像扩散提供了一种简单而通用的即插即用方法。代码见 https://github.com/SuhZhang/CorrAdapter。
英文原文摘要 1631 字符
Multi-image diffusion models can generate images like multi-views or videos to describe static or dynamic scenes, yet texture and structure drift persist, severely undermining the spatiotemporal consistency. Addressing this issue remains challenging, especially without any external geometric or semantic priors during the pure generative inference. In this paper, we introduce CorrAdapter, a plug-and-play adapter that discovers and exploits an innate property of the multi-image diffusion itself, aligning all output images before they are in fact generated. Specifically, CorrAdapter designs a bypass branch for transformer blocks in the multi-image diffusion model, encompassing a native correspondence constructor that builds reliable correspondences from the diffusion model's intermediate features, and an aligned area aggregator that integrates messages from only matching regions to avoid ambiguous information interactions. Given the native correspondences as guidance, CorrAdapter can enhance spatiotemporal consistency without any auxiliary inputs, and remains training-free and baseline-agnostic, which enables it to generalize seamlessly to various generation tasks. Additionally, we provide an optional training scheme to explore further-improved possibilities. Experiments on both static multi-view generation and dynamic video generation show that CorrAdapter consistently improves spatiotemporal consistency and perceptual quality over strong baselines, offering a simple yet versatile drop-in approach to geometrically faithful multi-image diffusion. Code is available at https://github.com/SuhZhang/CorrAdapter.