FRAG: Frequency Adapting Group for Diffusion Video Editing

TL;DR

FRAG is a novel method that improves diffusion-based video editing by preserving high-frequency details, leading to more consistent and natural edits without additional training.

Contribution

FRAG introduces a receptive field branch that enhances high-frequency component preservation in diffusion models, improving video editing quality in a model-agnostic way.

Findings

Reduces blurring and flickering in edited videos

Improves consistency and fidelity of video edits

Validated on TGVE and DAVIS benchmarks

Abstract

In video editing, the hallmark of a quality edit lies in its consistent and unobtrusive adjustment. Modification, when integrated, must be smooth and subtle, preserving the natural flow and aligning seamlessly with the original vision. Therefore, our primary focus is on overcoming the current challenges in high quality edit to ensure that each edit enhances the final product without disrupting its intended essence. However, quality deterioration such as blurring and flickering is routinely observed in recent diffusion video editing systems. We confirm that this deterioration often stems from high-frequency leak: the diffusion model fails to accurately synthesize high-frequency components during denoising process. To this end, we devise Frequency Adapting Group (FRAG) which enhances the video quality in terms of consistency and fidelity by introducing a novel receptive field branch to preserve high-frequency components during the denoising process. FRAG is performed in a model-agnostic manner without additional training and validates the effectiveness on video editing benchmarks (i.e., TGVE, DAVIS).