DGNet: Dynamic Gradient-Guided Network for Water-Related Optics Image Enhancement

TL;DR

DGNet introduces a dynamic gradient-guided approach with feature restoration and frequency smoothing modules, significantly improving underwater image enhancement by addressing complex degradation factors and noise.

Contribution

The paper proposes a novel dynamic gradient mechanism and specialized modules for underwater image enhancement, enhancing robustness and generalization over previous methods.

Findings

Achieved PSNR of 25.6dB and SSIM of 0.93 on UIEB dataset.

Outperformed existing state-of-the-art methods in underwater image quality.

Demonstrated efficiency with reduced parameters and inference time.

Abstract

Underwater image enhancement (UIE) is a challenging task due to the complex degradation caused by underwater environments. To solve this issue, previous methods often idealize the degradation process, and neglect the impact of medium noise and object motion on the distribution of image features, limiting the generalization and adaptability of the model. Previous methods use the reference gradient that is constructed from original images and synthetic ground-truth images. This may cause the network performance to be influenced by some low-quality training data. Our approach utilizes predicted images to dynamically update pseudo-labels, adding a dynamic gradient to optimize the network's gradient space. This process improves image quality and avoids local optima. Moreover, we propose a Feature Restoration and Reconstruction module (FRR) based on a Channel Combination Inference (CCI) strategy and a Frequency Domain Smoothing module (FRS). These modules decouple other degradation features while reducing the impact of various types of noise on network performance. Experiments on multiple public datasets demonstrate the superiority of our method over existing state-of-the-art approaches, especially in achieving performance milestones: PSNR of 25.6dB and SSIM of 0.93 on the UIEB dataset. Its efficiency in terms of parameter size and inference time further attests to its broad practicality. The code will be made publicly available.