PriorNet: A Novel Lightweight Network with Multidimensional Interactive Attention for Efficient Image Dehazing

TL;DR

PriorNetAddressing the challenge of dehazing with limited resources, this work proposes a lightweight network that directly estimates a per-pixel scaling factor $K(x)$ under the atmospheric scattering model to recover haze-free images. The method introduces Multidimensional Interactive Attention (MIA), fusing channel and a novel spatial attention to efficiently capture global haze cues without hefty parameter costs, complemented by a uniform 5x5 convolutional feature extractor and skip connections. The learning objective combines $Loss_{mse}$ with a perceptual term $Loss_{perception}$ (weighted by $\beta=0.1$) to preserve detail and realism, and the model achieves about $18$ Kb parameters while delivering strong generalization across datasets such as $\text{Haze4K}$ and $\text{Hazy extunderscore NYU extunderscore DepthV2}$, outperforming several baselines in PSNR/SSIM and downstream recognition tasks. This work demonstrates that attention-driven, physically grounded, and highly compact designs can realize effective dehazing suitable for edge devices, with broad implications for preprocessing in real-world vision systems.

Abstract

Hazy images degrade visual quality, and dehazing is a crucial prerequisite for subsequent processing tasks. Most current dehazing methods rely on neural networks and face challenges such as high computational parameter pressure and weak generalization capabilities. This paper introduces PriorNet--a novel, lightweight, and highly applicable dehazing network designed to significantly improve the clarity and visual quality of hazy images while avoiding excessive detail extraction issues. The core of PriorNet is the original Multi-Dimensional Interactive Attention (MIA) mechanism, which effectively captures a wide range of haze characteristics, substantially reducing the computational load and generalization difficulties associated with complex systems. By utilizing a uniform convolutional kernel size and incorporating skip connections, we have streamlined the feature extraction process. Simplifying the number of layers and architecture not only enhances dehazing efficiency but also facilitates easier deployment on edge devices. Extensive testing across multiple datasets has demonstrated PriorNet's exceptional performance in dehazing and clarity restoration, maintaining image detail and color fidelity in single-image dehazing tasks. Notably, with a model size of just 18Kb, PriorNet showcases superior dehazing generalization capabilities compared to other methods. Our research makes a significant contribution to advancing image dehazing technology, providing new perspectives and tools for the field and related domains, particularly emphasizing the importance of improving universality and deployability.

Figures (4)

• Figure 1: Architecture of PriorNet
• Figure 2: Architecture of the MIA
• Figure 3: Architecture of Convolutional Feature Extraction Module
• Figure 4: Visual Results of Dehazing Methods.