IR-Flow: Bridging Discriminative and Generative Image Restoration via Rectified Flow

2026-04-21

AI Summary

Existing single-step discriminative image restoration methods often suffer from detail loss, while generative approaches are hindered by low sampling efficiency and the entanglement between noise and residual components. This work introduces Rectified Flow to image restoration for the first time, establishing a unified framework that models an approximately linear transport trajectory from degraded to clean images through multi-scale data distribution flows and cumulative velocity fields. To enhance few-step restoration performance, the method incorporates multi-step consistency constraints. By synergistically combining the strengths of both discriminative and generative paradigms, the proposed approach enables highly efficient inference, achieving state-of-the-art quantitative and perceptual quality with only a few sampling steps on tasks such as deraining, denoising, and raindrop removal, while also demonstrating improved robustness to out-of-distribution degradations.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Multimodal Learning
• Planning, Routing, and Scheduling: Replanning and Plan Repair

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

In image restoration, single-step discriminative mappings often lack fine details via expectation learning, whereas generative paradigms suffer from inefficient multi-step sampling and noise-residual coupling. To address this dilemma, we propose IR-Flow, a novel image restoration method based on Rectified Flow that serves as a unified framework bridging the gap between discriminative and generative paradigms. Specifically, we first construct multilevel data distribution flows, which expand the ability of models to learn from and adapt to various levels of degradation. Subsequently, cumulative velocity fields are proposed to learn transport trajectories across varying degradation levels, guiding intermediate states toward the clean target, while a multi-step consistency constraint is presented to enforce trajectory coherence and boost few-step restoration performance. We show that directly establishing a linear transport flow between degraded and clean image domains not only enables fast inference but also improves adaptability to out-of-distribution degradations. Extensive evaluations on deraining, denoising and raindrop removal tasks demonstrate that IR-Flow achieves competitive quantitative results with only a few sampling steps, offering an efficient and flexible framework that maintains an excellent distortion-perception balance. Our code is available at https://github.com/fanzh03/IR-Flow.

Problem

• image restoration
• discriminative methods
• generative models
• sampling efficiency
• noise-residual coupling

Innovation

• Rectified Flow
• image restoration
• multi-step consistency
• velocity field
• distortion-perception tradeoff