Enhancing Underwater Images via Adaptive Semantic-aware Codebook Learning

2026-02-11

IEEE Transactions on Geoscience and Remote Sensing

AI Summary

Underwater image enhancement is challenging due to the absence of ground-truth references and spatially non-uniform degradation across semantic regions, which often leads to color distortion and detail loss when applying global processing strategies. To address this, this work proposes SUCode, a novel network that introduces, for the first time, a semantic-aware pixel-wise discrete codebook to adaptively model heterogeneous degradations. The method employs a three-stage training strategy to mitigate pseudo-label contamination and integrates a channel attention module (GCAM) with a frequency-aware feature fusion mechanism (FAFF). Extensive experiments demonstrate that SUCode significantly outperforms state-of-the-art methods across multiple benchmark datasets, achieving leading performance on both full-reference and no-reference evaluation metrics.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Semi-Supervised Learning
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Underwater Image Enhancement (UIE) is an ill-posed problem where natural clean references are not available, and the degradation levels vary significantly across semantic regions. Existing UIE methods treat images with a single global model and ignore the inconsistent degradation of different scene components. This oversight leads to significant color distortions and loss of fine details in heterogeneous underwater scenes, especially where degradation varies significantly across different image regions. Therefore, we propose SUCode (Semantic-aware Underwater Codebook Network), which achieves adaptive UIE from semantic-aware discrete codebook representation. Compared with one-shot codebook-based methods, SUCode exploits semantic-aware, pixel-level codebook representation tailored to heterogeneous underwater degradation. A three-stage training paradigm is employed to represent raw underwater image features to avoid pseudo ground-truth contamination. Gated Channel Attention Module (GCAM) and Frequency-Aware Feature Fusion (FAFF) jointly integrate channel and frequency cues for faithful color restoration and texture recovery. Extensive experiments on multiple benchmarks demonstrate that SUCode achieves state-of-the-art performance, outperforming recent UIE methods on both reference and no-reference metrics. The code will be made public available at https://github.com/oucailab/SUCode.

Problem

• Underwater Image Enhancement
• ill-posed problem
• semantic regions
• degradation variation
• color distortion

Innovation

• Semantic-aware codebook
• Underwater image enhancement
• Adaptive representation
• Frequency-aware fusion
• Gated Channel Attention