WorldWarp：基於非同步影片擴散的三維幾何傳播

WorldWarp: Propagating 3D Geometry with Asynchronous Video Diffusion

December 22, 2025

作者: Hanyang Kong, Xingyi Yang, Xiaoxu Zheng, Xinchao Wang

cs.AI

摘要

生成具備長程幾何一致性的影片存在一個根本難題：一方面，一致性要求嚴格遵循像素空間中的3D幾何關係，而另一方面，最先進的生成模型卻在相機條件化的潛在空間中運作最為高效。這種斷層導致現有方法在處理遮擋區域和複雜相機軌跡時面臨困難。為彌合此鴻溝，我們提出WorldWarp框架，該框架將3D結構錨點與2D生成修飾器相結合。為建立幾何基礎，WorldWarp透過高斯潑濺（3DGS）技術維護一個線上3D幾何快取。透過將歷史內容顯式扭曲至新視角，此快取作為結構支架，確保每個新幀都遵循先前的幾何關係。然而，靜態扭曲不可避免地會因遮擋而產生空洞與瑕疵。我們採用專為「填充與修訂」目標設計的時空擴散（ST-Diff）模型來解決此問題。我們的關鍵創新在於時空可變的噪聲調度機制：空白區域接受完整噪聲以觸發生成，而扭曲區域則接受部分噪聲以實現精修。透過在每一步動態更新3D快取，WorldWarp能在影片片段間維持一致性。因此，該方法透過以3D邏輯引導結構、同時以擴散邏輯完善紋理，實現了業界頂尖的擬真度。專案頁面：https://hyokong.github.io/worldwarp-page/。

English

Generating long-range, geometrically consistent video presents a fundamental dilemma: while consistency demands strict adherence to 3D geometry in pixel space, state-of-the-art generative models operate most effectively in a camera-conditioned latent space. This disconnect causes current methods to struggle with occluded areas and complex camera trajectories. To bridge this gap, we propose WorldWarp, a framework that couples a 3D structural anchor with a 2D generative refiner. To establish geometric grounding, WorldWarp maintains an online 3D geometric cache built via Gaussian Splatting (3DGS). By explicitly warping historical content into novel views, this cache acts as a structural scaffold, ensuring each new frame respects prior geometry. However, static warping inevitably leaves holes and artifacts due to occlusions. We address this using a Spatio-Temporal Diffusion (ST-Diff) model designed for a "fill-and-revise" objective. Our key innovation is a spatio-temporal varying noise schedule: blank regions receive full noise to trigger generation, while warped regions receive partial noise to enable refinement. By dynamically updating the 3D cache at every step, WorldWarp maintains consistency across video chunks. Consequently, it achieves state-of-the-art fidelity by ensuring that 3D logic guides structure while diffusion logic perfects texture. Project page: https://hyokong.github.io/worldwarp-page/{https://hyokong.github.io/worldwarp-page/}.