Fine-gained Zero-shot Video Sampling

TL;DR

The paper introduces $\\mathcal{ZS}^2$, a zero-shot video sampling method that generates high-quality, fine-grained videos from image diffusion models without training, outperforming some supervised approaches.

Contribution

It presents a novel zero-shot video sampling algorithm that leverages dependency noise and temporal momentum attention to produce detailed videos from image models.

Findings

Achieves state-of-the-art zero-shot video generation performance.

Outperforms some recent supervised methods.

Enables high-quality, fine-grained video synthesis from images.

Abstract

Incorporating a temporal dimension into pretrained image diffusion models for video generation is a prevalent approach. However, this method is computationally demanding and necessitates large-scale video datasets. More critically, the heterogeneity between image and video datasets often results in catastrophic forgetting of the image expertise. Recent attempts to directly extract video snippets from image diffusion models have somewhat mitigated these problems. Nevertheless, these methods can only generate brief video clips with simple movements and fail to capture fine-grained motion or non-grid deformation. In this paper, we propose a novel Zero-Shot video Sampling algorithm, denoted as $\mathcal{ZS}^2$, capable of directly sampling high-quality video clips from existing image synthesis methods, such as Stable Diffusion, without any training or optimization. Specifically, $\mathcal{ZS}^2$ utilizes the dependency noise model and temporal momentum attention to ensure content consistency and animation coherence, respectively. This ability enables it to excel in related tasks, such as conditional and context-specialized video generation and instruction-guided video editing. Experimental results demonstrate that $\mathcal{ZS}^2$ achieves state-of-the-art performance in zero-shot video generation, occasionally outperforming recent supervised methods. Homepage: \url{https://densechen.github.io/zss/}.