TurboEdit: Text-Based Image Editing Using Few-Step Diffusion Models

TL;DR

TurboEdit introduces a fast, three-step diffusion-based method for text-based image editing, addressing artifacts and enhancing editing strength through noise schedule adjustments and pseudo-guidance.

Contribution

The paper proposes a novel, efficient approach for text-based image editing using minimal diffusion steps, with insights into failure modes and solutions for artifacts and editing strength.

Findings

Enables effective image editing with as few as three diffusion steps.

Identifies noise mismatch as a cause of artifacts and proposes a shifted noise schedule.

Introduces pseudo-guidance to increase editing strength without artifacts.

Abstract

Diffusion models have opened the path to a wide range of text-based image editing frameworks. However, these typically build on the multi-step nature of the diffusion backwards process, and adapting them to distilled, fast-sampling methods has proven surprisingly challenging. Here, we focus on a popular line of text-based editing frameworks - the ``edit-friendly'' DDPM-noise inversion approach. We analyze its application to fast sampling methods and categorize its failures into two classes: the appearance of visual artifacts, and insufficient editing strength. We trace the artifacts to mismatched noise statistics between inverted noises and the expected noise schedule, and suggest a shifted noise schedule which corrects for this offset. To increase editing strength, we propose a pseudo-guidance approach that efficiently increases the magnitude of edits without introducing new artifacts. All in all, our method enables text-based image editing with as few as three diffusion steps, while providing novel insights into the mechanisms behind popular text-based editing approaches.