Frequency Domain-Based Diffusion Model for Unpaired Image Dehazing

2025-07-01

AI Summary

In unsupervised image dehazing, conventional contrastive learning often introduces irrelevant content and neglects haze-specific degradation patterns—particularly in the frequency domain’s amplitude spectrum. To address this, we propose the first end-to-end dehazing framework integrating diffusion modeling into frequency-domain reconstruction, jointly repairing the amplitude spectrum and optimizing the phase spectrum. Our key contributions are: (1) a novel amplitude-spectrum diffusion model that explicitly characterizes haze-induced degradation distributions; (2) an amplitude residual encoder (ARE) to mitigate domain shift between hazy and clear amplitude spectra; and (3) an attention-driven phase correction module (PCM) to suppress phase-related artifacts. Extensive experiments on both synthetic and real-world datasets demonstrate significant improvements over state-of-the-art methods, achieving superior dehazing quality and enhanced detail fidelity.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

Unpaired image dehazing has attracted increasing attention due to its flexible data requirements during model training. Dominant methods based on contrastive learning not only introduce haze-unrelated content information, but also ignore haze-specific properties in the frequency domain (ie,~haze-related degradation is mainly manifested in the amplitude spectrum). To address these issues, we propose a novel frequency domain-based diffusion model, named ours, for fully exploiting the beneficial knowledge in unpaired clear data. In particular, inspired by the strong generative ability shown by Diffusion Models (DMs), we tackle the dehazing task from the perspective of frequency domain reconstruction and perform the DMs to yield the amplitude spectrum consistent with the distribution of clear images. To implement it, we propose an Amplitude Residual Encoder (ARE) to extract the amplitude residuals, which effectively compensates for the amplitude gap from the hazy to clear domains, as well as provide supervision for the DMs training. In addition, we propose a Phase Correction Module (PCM) to eliminate artifacts by further refining the phase spectrum during dehazing with a simple attention mechanism. Experimental results demonstrate that our ours outperforms other state-of-the-art methods on both synthetic and real-world datasets.

Problem

• Unpaired image dehazing without haze-unrelated content interference
• Addressing haze-specific frequency domain properties for better reconstruction
• Eliminating artifacts in phase spectrum during dehazing process

Innovation

• Frequency domain-based diffusion model for dehazing
• Amplitude Residual Encoder compensates amplitude gap
• Phase Correction Module refines phase spectrum