PHATNet: A Physics-guided Haze Transfer Network for Domain-adaptive Real-world Image Dehazing

2025-07-20

AI Summary

Real-world image dehazing models often suffer from significant performance degradation on unseen domains. To address this, we propose the Physics-guided Haze Transfer Network (PHATNet), which leverages the atmospheric scattering model to synthesize domain-specific hazy images from source-domain haze-free images—thereby constructing a tailored fine-tuning set for adaptive model optimization. Our method introduces two novel losses: Haze-Transfer-Consistency loss, enforcing structural consistency in haze transfer, and Content-Leakage loss, explicitly decoupling haze characteristics from intrinsic image content. PHATNet integrates physical priors, domain adaptation, and adversarial training into a unified framework. Extensive experiments demonstrate that PHATNet consistently enhances the performance of multiple state-of-the-art dehazing models across diverse real-world dehazing benchmarks, achieving superior domain adaptation capability and establishing new state-of-the-art results.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Diffusion Models for Vision
• Application Domains: Transportation

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Responsible Web: Algorithmic accountability and transparency on the web

Abstract

Image dehazing aims to remove unwanted hazy artifacts in images. Although previous research has collected paired real-world hazy and haze-free images to improve dehazing models' performance in real-world scenarios, these models often experience significant performance drops when handling unseen real-world hazy images due to limited training data. This issue motivates us to develop a flexible domain adaptation method to enhance dehazing performance during testing. Observing that predicting haze patterns is generally easier than recovering clean content, we propose the Physics-guided Haze Transfer Network (PHATNet) which transfers haze patterns from unseen target domains to source-domain haze-free images, creating domain-specific fine-tuning sets to update dehazing models for effective domain adaptation. Additionally, we introduce a Haze-Transfer-Consistency loss and a Content-Leakage Loss to enhance PHATNet's disentanglement ability. Experimental results demonstrate that PHATNet significantly boosts state-of-the-art dehazing models on benchmark real-world image dehazing datasets.

Problem

• Enhancing real-world image dehazing via domain adaptation
• Transferring haze patterns to improve model generalization
• Mitigating performance drops on unseen hazy images

Innovation

• Physics-guided Haze Transfer Network for adaptation
• Haze pattern transfer to create fine-tuning sets
• Haze-Transfer-Consistency and Content-Leakage Losses