Eta Inversion: Designing an Optimal Eta Function for Diffusion-based Real Image Editing
This addresses the challenge of achieving faithful and source-resembling edits in diffusion-based image editing, representing an incremental improvement over current methods.
The paper tackles the problem of diffusion inversion for real image editing by designing a time- and region-dependent eta function, resulting in a method that outperforms existing strategies in quantitative and qualitative assessments.
Diffusion models have achieved remarkable success in the domain of text-guided image generation and, more recently, in text-guided image editing. A commonly adopted strategy for editing real images involves inverting the diffusion process to obtain a noisy representation of the original image, which is then denoised to achieve the desired edits. However, current methods for diffusion inversion often struggle to produce edits that are both faithful to the specified text prompt and closely resemble the source image. To overcome these limitations, we introduce a novel and adaptable diffusion inversion technique for real image editing, which is grounded in a theoretical analysis of the role of $η$ in the DDIM sampling equation for enhanced editability. By designing a universal diffusion inversion method with a time- and region-dependent $η$ function, we enable flexible control over the editing extent. Through a comprehensive series of quantitative and qualitative assessments, involving a comparison with a broad array of recent methods, we demonstrate the superiority of our approach. Our method not only sets a new benchmark in the field but also significantly outperforms existing strategies.