UniEdit-Flow: Unleashing Inversion and Editing in the Era of Flow Models

2025-04-17

AI Summary

To address the lack of efficient image inversion and editing methods for flow matching models, this paper introduces Uni-Inv—a tuning-free, model-agnostic inversion framework—and Uni-Edit—a delayed-injection region-based editing method—both specifically designed for flow models. Leveraging the geometric properties of flow trajectories, Uni-Inv employs a predictor-corrector algorithm coupled with trajectory-aligned inversion to achieve high-fidelity reconstruction. Uni-Edit integrates region masking guidance with a delayed-injection mechanism to ensure precise, controllable editing within target regions while strongly preserving unedited areas. Extensive evaluation across diverse flow-based generative models demonstrates that our approach significantly outperforms diffusion-model adaptation strategies in editing quality, achieving both high reconstruction accuracy and robust editing performance under low computational overhead. This work establishes a general, efficient paradigm for controllable generation with flow matching models.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Calibration & Uncertainty Quantification
• Natural Language Processing: Generation

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Flow matching models have emerged as a strong alternative to diffusion models, but existing inversion and editing methods designed for diffusion are often ineffective or inapplicable to them. The straight-line, non-crossing trajectories of flow models pose challenges for diffusion-based approaches but also open avenues for novel solutions. In this paper, we introduce a predictor-corrector-based framework for inversion and editing in flow models. First, we propose Uni-Inv, an effective inversion method designed for accurate reconstruction. Building on this, we extend the concept of delayed injection to flow models and introduce Uni-Edit, a region-aware, robust image editing approach. Our methodology is tuning-free, model-agnostic, efficient, and effective, enabling diverse edits while ensuring strong preservation of edit-irrelevant regions. Extensive experiments across various generative models demonstrate the superiority and generalizability of Uni-Inv and Uni-Edit, even under low-cost settings. Project page: https://uniedit-flow.github.io/

Problem

• Inversion and editing methods for flow models are ineffective
• Straight-line trajectories in flow models challenge diffusion-based approaches
• Propose tuning-free, model-agnostic framework for flow model inversion and editing

Innovation

• Predictor-corrector framework for flow models
• Uni-Inv inversion method for accurate reconstruction
• Uni-Edit region-aware robust image editing