Semantic Image Inversion and Editing using Rectified Stochastic Differential Equations
This work addresses the problem of efficient and effective image inversion and editing for users of generative models, offering a novel approach that improves upon existing methods without requiring expensive training or optimization.
The paper tackles the problem of inverting and editing real images using generative models, specifically addressing challenges in diffusion models by proposing a method based on rectified flows and dynamic optimal control. It achieves state-of-the-art performance in zero-shot inversion and editing, with human evaluations confirming user preference.
Generative models transform random noise into images; their inversion aims to transform images back to structured noise for recovery and editing. This paper addresses two key tasks: (i) inversion and (ii) editing of a real image using stochastic equivalents of rectified flow models (such as Flux). Although Diffusion Models (DMs) have recently dominated the field of generative modeling for images, their inversion presents faithfulness and editability challenges due to nonlinearities in drift and diffusion. Existing state-of-the-art DM inversion approaches rely on training of additional parameters or test-time optimization of latent variables; both are expensive in practice. Rectified Flows (RFs) offer a promising alternative to diffusion models, yet their inversion has been underexplored. We propose RF inversion using dynamic optimal control derived via a linear quadratic regulator. We prove that the resulting vector field is equivalent to a rectified stochastic differential equation. Additionally, we extend our framework to design a stochastic sampler for Flux. Our inversion method allows for state-of-the-art performance in zero-shot inversion and editing, outperforming prior works in stroke-to-image synthesis and semantic image editing, with large-scale human evaluations confirming user preference.