Prior-guided Hierarchical Harmonization Network for Efficient Image Dehazing
Authors:
Preprint
Abstract
Image dehazing is a crucial task that involves the enhancement of degraded images to recover their sharpness and textures. While vision Transformers have exhibited impressive results in diverse dehazing tasks, their quadratic complexity and lack of dehazing priors pose significant drawbacks for real-world applications. In this paper, guided by triple priors, Bright Channel Prior (BCP), Dark Channel Prior (DCP), and Histogram Equalization (HE), we propose a \textit{P}rior-\textit{g}uided Hierarchical \textit{H}armonization Network (PGH$^2$Net) for image dehazing. PGH$^2$Net is built upon the UNet-like architecture with an efficient encoder and decoder, consisting of two module types: (1) Prior aggregation module that injects B/DCP and selects diverse contexts with gating attention. (2) Feature harmonization modules that subtract low-frequency components from spatial and channel aspects and learn more informative feature distributions to equalize the feature maps.Keywords
All Related Versions
This publication has 43 references indexed in Scilit:
- DehazeMamba: large multi-modal model guided single image dehazing via mambaVisual Intelligence, 2025
- EENet: An effective and efficient network for single image dehazingPattern Recognition, 2024
- MambaIR: A Simple Baseline for Image Restoration with State-Space Model2024
- Simple Baselines for Image Restoration2022
- Deep image prior based defense against adversarial examplesPattern Recognition, 2022
- DCA-CycleGAN: Unsupervised single image dehazing using Dark Channel Attention optimized CycleGANJournal of Visual Communication and Image Representation, 2022
- Defocus Deblurring Using Dual-Pixel Data2020
- CBAM: Convolutional Block Attention Module2018
- Image Deblurring via Extreme Channels Prior2017
- Fast Haze Removal for Nighttime Image Using Maximum Reflectance Prior2017