Inverse-and-Edit: Effective and Fast Image Editing by Cycle Consistency Models
This addresses the problem of slow and limited editing capabilities in diffusion models for image generation tasks, offering a more efficient solution.
The paper tackles the computational intensity and poor inversion quality of diffusion models for image editing by introducing a cycle-consistency optimization with consistency models, achieving high-quality editing in just four steps and matching or surpassing full-step diffusion models in performance.
Recent advances in image editing with diffusion models have achieved impressive results, offering fine-grained control over the generation process. However, these methods are computationally intensive because of their iterative nature. While distilled diffusion models enable faster inference, their editing capabilities remain limited, primarily because of poor inversion quality. High-fidelity inversion and reconstruction are essential for precise image editing, as they preserve the structural and semantic integrity of the source image. In this work, we propose a novel framework that enhances image inversion using consistency models, enabling high-quality editing in just four steps. Our method introduces a cycle-consistency optimization strategy that significantly improves reconstruction accuracy and enables a controllable trade-off between editability and content preservation. We achieve state-of-the-art performance across various image editing tasks and datasets, demonstrating that our method matches or surpasses full-step diffusion models while being substantially more efficient. The code of our method is available on GitHub at https://github.com/ControlGenAI/Inverse-and-Edit.