BiFM: Bidirectional Flow Matching for Few-Step Image Editing and Generation

TL;DR

BiFM introduces a unified bidirectional flow matching framework that enhances few-step image editing and generation by jointly learning inversion and generation, leading to improved quality and scalability.

Contribution

The paper proposes BiFM, a novel model that jointly learns image generation and inversion with bidirectional flow matching, improving efficiency and generalization in few-step image editing and generation.

Findings

Outperforms existing few-step methods in image editing quality

Enables one-step inversion with high accuracy

Seamlessly integrates into existing diffusion frameworks

Abstract

Recent diffusion and flow matching models have demonstrated strong capabilities in image generation and editing by progressively removing noise through iterative sampling. While this enables flexible inversion for semantic-preserving edits, few-step sampling regimes suffer from poor forward process approximation, leading to degraded editing quality. Existing few-step inversion methods often rely on pretrained generators and auxiliary modules, limiting scalability and generalization across different architectures. To address these limitations, we propose BiFM (Bidirectional Flow Matching), a unified framework that jointly learns generation and inversion within a single model. BiFM directly estimates average velocity fields in both ``image $\to$ noise" and ``noise $\to$ image" directions, constrained by a shared instantaneous velocity field derived from either predefined schedules or pretrained multi-step diffusion models. Additionally, BiFM introduces a novel training strategy using continuous time-interval supervision, stabilized by a bidirectional consistency objective and a lightweight time-interval embedding. This bidirectional formulation also enables one-step inversion and can integrate seamlessly into popular diffusion and flow matching backbones. Across diverse image editing and generation tasks, BiFM consistently outperforms existing few-step approaches, achieving superior performance and editability.