GoodDrag: Towards Good Practices for Drag Editing with Diffusion Models

TL;DR

GoodDrag introduces a novel diffusion-based framework with an AlDD approach and motion supervision to enhance drag editing stability and quality, supported by new datasets and metrics.

Contribution

It presents a new AlDD framework and motion supervision for drag editing, along with a benchmark dataset and quality metrics, advancing the state-of-the-art.

Findings

GoodDrag outperforms existing methods in quality and stability.

The new dataset Drag100 enables better benchmarking.

Proposed metrics accurately assess drag editing quality.

Abstract

In this paper, we introduce GoodDrag, a novel approach to improve the stability and image quality of drag editing. Unlike existing methods that struggle with accumulated perturbations and often result in distortions, GoodDrag introduces an AlDD framework that alternates between drag and denoising operations within the diffusion process, effectively improving the fidelity of the result. We also propose an information-preserving motion supervision operation that maintains the original features of the starting point for precise manipulation and artifact reduction. In addition, we contribute to the benchmarking of drag editing by introducing a new dataset, Drag100, and developing dedicated quality assessment metrics, Dragging Accuracy Index and Gemini Score, utilizing Large Multimodal Models. Extensive experiments demonstrate that the proposed GoodDrag compares favorably against the state-of-the-art approaches both qualitatively and quantitatively. The project page is https://gooddrag.github.io.