COVE: Unleashing the Diffusion Feature Correspondence for Consistent Video Editing

TL;DR

COVE introduces a novel diffusion feature correspondence method for consistent, high-quality video editing that leverages inherent diffusion features, a sliding-window similarity strategy, and token merging to improve efficiency and temporal coherence without additional training.

Contribution

The paper presents a new diffusion feature correspondence approach for video editing, enabling temporal consistency and efficiency without extra training or optimization.

Findings

Achieves state-of-the-art performance in various video editing scenarios.

Outperforms existing methods both quantitatively and qualitatively.

Efficiently reduces GPU memory usage and accelerates editing process.

Abstract

Video editing is an emerging task, in which most current methods adopt the pre-trained text-to-image (T2I) diffusion model to edit the source video in a zero-shot manner. Despite extensive efforts, maintaining the temporal consistency of edited videos remains challenging due to the lack of temporal constraints in the regular T2I diffusion model. To address this issue, we propose COrrespondence-guided Video Editing (COVE), leveraging the inherent diffusion feature correspondence to achieve high-quality and consistent video editing. Specifically, we propose an efficient sliding-window-based strategy to calculate the similarity among tokens in the diffusion features of source videos, identifying the tokens with high correspondence across frames. During the inversion and denoising process, we sample the tokens in noisy latent based on the correspondence and then perform self-attention within them. To save GPU memory usage and accelerate the editing process, we further introduce the temporal-dimensional token merging strategy, which can effectively reduce redundancy. COVE can be seamlessly integrated into the pre-trained T2I diffusion model without the need for extra training or optimization. Extensive experiment results demonstrate that COVE achieves the start-of-the-art performance in various video editing scenarios, outperforming existing methods both quantitatively and qualitatively. The code will be release at https://github.com/wangjiangshan0725/COVE.