cs.CVNov 18, 2025

A Generative Data Framework with Reliable Supervision for Underwater Image Restoration and Enhancement

Authors: Yufeng Tian, Yifan Chen, Zhe Sun, Libang Chen, Mingyu Dou, Jijun Lu, Ye Zheng, Xuelong Li

Organizations: Institute of Artificial Intelligence (TeleAI), China Telecom, Shanghai, 200232, China · Department of Automation, Shanghai Jiao Tong University, Shanghai, 200240, China · School of Artificial Intelligence, Optics and ElectroNics (iOPEN), Northwestern Polytechnical University, Xi’an, 710072, China · Guangdong Provincial Key Laboratory of Quantum Metrology and Sensing, School of Physics and Astronomy, Sun Yat-Sen University (Zhuhai Campus), Zhuhai, 519082, China · 

Abstract

Learning-based underwater image restoration and enhancement rely heavily on high-quality paired training data. However, obtaining clear reference images for real underwater scenes remains challenging. Existing datasets often use the outputs of enhancement algorithms as references, which may introduce uncertainty into supervision targets. To address this issue, we propose a generative data framework that preserves clear source images as restoration references and synthesizes corresponding underwater inputs through unpaired image-to-image translation. Using representative underwater appearances to guide the translation, we construct UWNature from natural photographs. We systematically examine the reliability of the generated supervision through its agreement with real underwater appearances, structural consistency, and depth-related degradation characteristics. Experiments across 6 restoration architectures and 3 test sets evaluate its effectiveness through objective metrics, human perceptual assessment, and the correspondence between synthetic degradation and learned restoration behavior. The results demonstrate competitive restoration performance and favorable perceptual quality on real underwater images. Further analysis reveals that greater spatial plausibility of synthetic degradation does not necessarily improve restoration performance, highlighting the importance of evaluating both the intrinsic properties of synthetic data and their effectiveness for model learning. These findings support a scalable framework for constructing and evaluating reliable paired supervision for underwater image restoration and enhancement. The generated datasets are available at: https://github.com/tianyufengcn/SynUIEDatasets.

Figures & tables

Explore similar work

CardsList
  1. Rethinking Underwater Image Enhancement: Learning through Text-Guided Degradation Curriculum

    Oct 21, 2026Hongri Tian, Wenxue Li, Tian Ye +7Underwater Image EnhancementImage Restoration

  2. WaterGen: Decoupling Scene and Medium in Underwater Image Generation

    Jun 30, 2026Jiayi Wu, Tianfu Wang, Tianyi Xiong +6Synthetic Data AugmentationImage Generation

  3. Multi-level Constrained Learning for Underwater Image Enhancement

    Oct 29, 2026Wangtao Bao, Liping Xie, Kanjian Zhang +1Underwater Image EnhancementImage Restoration