VideoCanvas: Unified Video Completion from Arbitrary Spatiotemporal Patches via In-Context Conditioning

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Cai, Minghong, Wang, Qiulin, Ye, Zongli, Liu, Wenze, Liu, Quande, Ye, Weicai, Wang, Xintao, Wan, Pengfei, Gai, Kun, Yue, Xiangyu
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866911723904040960
author Cai, Minghong
Wang, Qiulin
Ye, Zongli
Liu, Wenze
Liu, Quande
Ye, Weicai
Wang, Xintao
Wan, Pengfei
Gai, Kun
Yue, Xiangyu
author_facet Cai, Minghong
Wang, Qiulin
Ye, Zongli
Liu, Wenze
Liu, Quande
Ye, Weicai
Wang, Xintao
Wan, Pengfei
Gai, Kun
Yue, Xiangyu
contents Existing controllable video generation methods are typically designed for rigid, task-specific settings, such as first-frame image-to-video, inpainting, or interpolation, treating spatio-temporal control as a set of isolated problems. We formalize a unified task, arbitrary spatio-temporal video completion, where a model generates a coherent video from user-specified patches placed at any spatial location and timestamp. However, realizing such a unified framework within modern latent video diffusion models is non-trivial: causal video VAEs compress multiple frames into a single latent slot, making frame-level conditioning fundamentally ill-posed, and directly feeding sparsely populated, zero-padded video inputs into the VAE leads to severe out-of-distribution artifacts. To address these challenges, we propose VideoCanvas, a simple yet effective framework that adapts the In-Context Conditioning paradigm to arbitrary spatio-temporal completion without modifying or retraining the VAE. Our key idea is a hybrid conditioning strategy that decouples spatial and temporal control: spatially, we encode zero-padded full-frame canvases in image mode to keep VAE inputs in-distribution, and temporally we use Temporal RoPE Interpolation to assign each condition a continuous fractional index in the latent sequence for precise frame-level alignment. To evaluate this capability, we develop VideoCanvasBench, the first benchmark for arbitrary spatio-temporal video completion, covering both intra-scene fidelity and inter-scene creativity. Extensive experiments demonstrate that VideoCanvas achieves state-of-the-art performance across a diverse range of video generation tasks under a single, unified framework.
format Preprint
id arxiv_https___arxiv_org_abs_2510_08555
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle VideoCanvas: Unified Video Completion from Arbitrary Spatiotemporal Patches via In-Context Conditioning
Cai, Minghong
Wang, Qiulin
Ye, Zongli
Liu, Wenze
Liu, Quande
Ye, Weicai
Wang, Xintao
Wan, Pengfei
Gai, Kun
Yue, Xiangyu
Computer Vision and Pattern Recognition
Existing controllable video generation methods are typically designed for rigid, task-specific settings, such as first-frame image-to-video, inpainting, or interpolation, treating spatio-temporal control as a set of isolated problems. We formalize a unified task, arbitrary spatio-temporal video completion, where a model generates a coherent video from user-specified patches placed at any spatial location and timestamp. However, realizing such a unified framework within modern latent video diffusion models is non-trivial: causal video VAEs compress multiple frames into a single latent slot, making frame-level conditioning fundamentally ill-posed, and directly feeding sparsely populated, zero-padded video inputs into the VAE leads to severe out-of-distribution artifacts. To address these challenges, we propose VideoCanvas, a simple yet effective framework that adapts the In-Context Conditioning paradigm to arbitrary spatio-temporal completion without modifying or retraining the VAE. Our key idea is a hybrid conditioning strategy that decouples spatial and temporal control: spatially, we encode zero-padded full-frame canvases in image mode to keep VAE inputs in-distribution, and temporally we use Temporal RoPE Interpolation to assign each condition a continuous fractional index in the latent sequence for precise frame-level alignment. To evaluate this capability, we develop VideoCanvasBench, the first benchmark for arbitrary spatio-temporal video completion, covering both intra-scene fidelity and inter-scene creativity. Extensive experiments demonstrate that VideoCanvas achieves state-of-the-art performance across a diverse range of video generation tasks under a single, unified framework.
title VideoCanvas: Unified Video Completion from Arbitrary Spatiotemporal Patches via In-Context Conditioning
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2510.08555