IntrinsicWeather：本征空间中的可控天气编辑
IntrinsicWeather: Controllable Weather Editing in Intrinsic Space
把天气编辑从像素空间搬到本征空间（材质/几何/光照图），用本征图作为中间表示提升可控性；其本征图感知注意力与 CLIP 空间天气提示词插值可直接迁移到其他条件图像编辑与逆渲染任务。
原文依据 an intrinsic map-aware attention mechanism that improves spatial correspondence and decomposition quality in large outdoor scenes
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论本文提出 IntrinsicWeather，一个在本征空间中进行可控天气编辑的扩散框架。该框架包含两个基于扩散先验的组件：逆渲染器从输入图像中估计材质属性、场景几何与光照，并将其表示为本征图；前向渲染器则利用这些几何与材质图，结合描述特定天气条件的文本提示词生成最终图像。与传统像素空间编辑方法相比，本征图提升了可控性。我们提出一种本征图感知注意力机制，改善了大型户外场景中的空间对应关系与分解质量。在前向渲染中，我们利用天气提示词的 CLIP 空间插值实现细粒度天气控制。我们还引入了一个合成数据集与一个真实世界数据集，分别包含 3.8 万张与 1.8 万张各种天气条件下的图像，且每张图像均带有本征图标注。IntrinsicWeather 优于最优的像素空间编辑方法、天气复原方法以及基于渲染的方法，并在自动驾驶等下游任务上展现出潜力，可提升检测与分割在恶劣天气场景下的鲁棒性。
英文原文摘要 1270 字符
We present IntrinsicWeather, a diffusion-based framework for controllable weather editing in intrinsic space. Our framework includes two components based on diffusion priors: an inverse renderer that estimates material properties, scene geometry, and lighting as intrinsic maps from an input image, and a forward renderer that utilizes these geometry and material maps along with a text prompt that describes specific weather conditions to generate a final image. The intrinsic maps enhance controllability compared to traditional pixel-space editing approaches. We propose an intrinsic map-aware attention mechanism that improves spatial correspondence and decomposition quality in large outdoor scenes. For forward rendering, we leverage CLIP-space interpolation of weather prompts to achieve fine-grained weather control. We also introduce a synthetic and a real-world dataset, containing 38k and 18k images under various weather conditions, each with intrinsic map annotations. IntrisicWeather outperforms state-of-the-art pixel-space editing approaches, weather restoration methods, and rendering-based methods, showing promise for downstream tasks such as autonomous driving, enhancing the robustness of detection and segmentation in challenging weather scenarios.