DPF-Net: Physical Imaging Model Embedded Data-Driven Underwater Image Enhancement

2025-03-16

AI Summary

Underwater images suffer from severe color distortion, low contrast, and blurred details due to wavelength-dependent light absorption and scattering. To address these challenges, we propose DPF-Net, a two-stage physically guided enhancement network. In the first stage, a novel physics-guided parameter estimation module jointly learns key underwater imaging parameters—including depth, lighting conditions, and medium attenuation coefficients—by leveraging physical priors. In the second stage, a degradation-consistent enhancement network is constructed based on the estimated parameters, incorporating parameter-embedded feature fusion and multi-constraint joint optimization. Crucially, we introduce a weak-reference loss trained across the entire dataset, substantially reducing reliance on scarce high-quality ground-truth images. Our method synergistically integrates physical modeling with data-driven learning. Extensive experiments demonstrate state-of-the-art performance on multiple benchmarks, achieving significant improvements in color fidelity, contrast restoration, and fine-grained texture recovery. Code and pretrained models are publicly available.

Technology Category

• Computer Vision: Low Level & Physics-based Vision
• Search and Optimization: Learning to Search
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Responsible Web: Data and user privacy-enhancing technologies for the Web
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Due to the complex interplay of light absorption and scattering in the underwater environment, underwater images experience significant degradation. This research presents a two-stage underwater image enhancement network called the Data-Driven and Physical Parameters Fusion Network (DPF-Net), which harnesses the robustness of physical imaging models alongside the generality and efficiency of data-driven methods. We first train a physical parameter estimate module using synthetic datasets to guarantee the trustworthiness of the physical parameters, rather than solely learning the fitting relationship between raw and reference images by the application of the imaging equation, as is common in prior studies. This module is subsequently trained in conjunction with an enhancement network, where the estimated physical parameters are integrated into a data-driven model within the embedding space. To maintain the uniformity of the restoration process amid underwater imaging degradation, we propose a physics-based degradation consistency loss. Additionally, we suggest an innovative weak reference loss term utilizing the entire dataset, which alleviates our model's reliance on the quality of individual reference images. Our proposed DPF-Net demonstrates superior performance compared to other benchmark methods across multiple test sets, achieving state-of-the-art results. The source code and pre-trained models are available on the project home page: https://github.com/OUCVisionGroup/DPF-Net.

Problem

• Enhance underwater images degraded by light absorption and scattering.
• Integrate physical imaging models with data-driven methods for better results.
• Reduce reliance on individual reference images using a novel loss term.

Innovation

• Combines physical imaging models with data-driven methods
• Uses synthetic datasets for physical parameter estimation
• Introduces physics-based degradation consistency loss