HazeFlow: Revisit Haze Physical Model as ODE and Non-Homogeneous Haze Generation for Real-World Dehazing

2025-09-19

AI Summary

Real-world single-image dehazing faces two key challenges: (1) domain shift arising from the scarcity of paired training data, and (2) limited expressivity of the conventional atmospheric scattering model (ASM) in characterizing complex, spatially varying haze distributions. To address these, we reformulate ASM as an ordinary differential equation (ODE), casting dehazing as continuous trajectory learning from hazy to clear images. Building upon the Rectified Flow framework, we propose a non-uniform generative dehazing method enabling efficient single-step inference. Furthermore, to better capture realistic haze statistics, we introduce a haze distribution simulation strategy grounded in Markov Chain Brownian Motion (MCBM). Extensive experiments on multiple real-world benchmarks demonstrate state-of-the-art performance, significantly narrowing the gap between synthetic and real domains while improving generalization and restoration fidelity.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Search and Optimization: Sampling/Simulation-based Search
• Machine Learning: Deep Generative Models & Autoencoders

Application Category

• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications

Abstract

Dehazing involves removing haze or fog from images to restore clarity and improve visibility by estimating atmospheric scattering effects. While deep learning methods show promise, the lack of paired real-world training data and the resulting domain gap hinder generalization to real-world scenarios. In this context, physics-grounded learning becomes crucial; however, traditional methods based on the Atmospheric Scattering Model (ASM) often fall short in handling real-world complexities and diverse haze patterns. To solve this problem, we propose HazeFlow, a novel ODE-based framework that reformulates ASM as an ordinary differential equation (ODE). Inspired by Rectified Flow (RF), HazeFlow learns an optimal ODE trajectory to map hazy images to clean ones, enhancing real-world dehazing performance with only a single inference step. Additionally, we introduce a non-homogeneous haze generation method using Markov Chain Brownian Motion (MCBM) to address the scarcity of paired real-world data. By simulating realistic haze patterns through MCBM, we enhance the adaptability of HazeFlow to diverse real-world scenarios. Through extensive experiments, we demonstrate that HazeFlow achieves state-of-the-art performance across various real-world dehazing benchmark datasets.

Problem

• Addressing real-world dehazing generalization with limited paired training data
• Reformulating atmospheric scattering model as ODE for better physical grounding
• Generating non-homogeneous haze patterns to overcome domain gap issues

Innovation

• Reformulates haze model as ODE trajectory
• Uses single inference step for dehazing
• Generates non-homogeneous haze via Markov Chain