RPD-Diff: Region-Adaptive Physics-Guided Diffusion Model for Visibility Enhancement under Dense and Non-Uniform Haze

2025-08-23

AI Summary

To address severe information degradation and strong spatial heterogeneity in single-image dehazing under dense, non-uniform haze—leading to insufficient conditioning and poor adaptability in conventional diffusion models—this paper proposes a region-adaptive, physics-guided diffusion model. Our method integrates atmospheric scattering priors into the diffusion process to formulate a physically grounded intermediate-state objective; designs a haze-aware denoising timestep predictor for block-wise, dynamic adjustment of denoising strength; and introduces a transmission-map cross-attention mechanism to enhance modeling of spatial haze distribution. Evaluated on four real-world datasets, our approach achieves state-of-the-art performance, significantly improving fine-detail recovery and color fidelity compared to existing methods.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Search and Optimization: Learning to Search
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Security and Privacy: Data transparency and provenance
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Single-image dehazing under dense and non-uniform haze conditions remains challenging due to severe information degradation and spatial heterogeneity. Traditional diffusion-based dehazing methods struggle with insufficient generation conditioning and lack of adaptability to spatially varying haze distributions, which leads to suboptimal restoration. To address these limitations, we propose RPD-Diff, a Region-adaptive Physics-guided Dehazing Diffusion Model for robust visibility enhancement in complex haze scenarios. RPD-Diff introduces a Physics-guided Intermediate State Targeting (PIST) strategy, which leverages physical priors to reformulate the diffusion Markov chain by generation target transitions, mitigating the issue of insufficient conditioning in dense haze scenarios. Additionally, the Haze-Aware Denoising Timestep Predictor (HADTP) dynamically adjusts patch-specific denoising timesteps employing a transmission map cross-attention mechanism, adeptly managing non-uniform haze distributions. Extensive experiments across four real-world datasets demonstrate that RPD-Diff achieves state-of-the-art performance in challenging dense and non-uniform haze scenarios, delivering high-quality, haze-free images with superior detail clarity and color fidelity.

Problem

• Enhancing visibility under dense non-uniform haze
• Addressing insufficient conditioning in diffusion dehazing
• Managing spatially varying haze distributions adaptively

Innovation

• Physics-guided Intermediate State Targeting strategy
• Haze-Aware Denoising Timestep Predictor mechanism
• Region-adaptive diffusion model for non-uniform haze