$π$-SUB: A Physics-Informed Synthetic Underwater Benchmark Dataset for Underwater Image Enhancement

2026-08-11

AI Summary

Existing synthetic underwater image datasets exhibit a significant domain gap from real-world scenes, limiting the training and evaluation of underwater image enhancement (UIE) methods. To address this, this work proposes a high-fidelity synthesis framework grounded in an extended physical model that jointly accounts for depth-dependent downwelling irradiance, bio-optically resolved absorption, scattering characteristics across all Jerlov water types, and independently controllable residual effects. The framework generates paired datasets spanning diverse water types and depths. Experimental results demonstrate that the synthesized images achieve a 46% lower FID than Syrea, yield UIQM improvements of 4.18%–9.46% across four state-of-the-art UIE models, and reduce NIQE by 23.98%–48.78%, substantially enhancing photorealism and generalization capability.

Abstract

This paper presents $π$-SUB, a physics-informed framework for generating synthetic underwater benchmark datasets that bridges the synthetic-to-real gap for Underwater Image Enhancement (UIE). The proposed framework extends the classical underwater image formation model by incorporating depth-dependent downwelling irradiance, biologically resolved absorption, and environmental scattering across all ten Jerlov water types, together with independently controllable residual phenomena. Using this framework, the $π$-SUB dataset consists of paired synthetic underwater-reference images spanning shallow-to-deep and coastal-to-oceanic environments. Extensive simulation studies have been carried out to evaluate $π$-SUB along two criteria namely hyper-realism and generalizability. For hyper-realism, $π$-SUB attains a global Frechet Inception Distance (FID) that is 46% lower than Syrea. For generalizability, four state-of-the-art UIE architectures (FUnIE-GAN, Pix2Pix, PUIE-Net, and Phaseformer) are used for comparative evaluation of $π$-SUB. These models were independently trained on six datasets including one real and five synthetic datasets and tested on six real-world benchmarks datasets. Across four UIE architectures and six real benchmark datasets, $π$-SUB improves UIQM by 4.18% over PHISWID (next best) and 9.46% over Syrea (next best), while reducing NIQE by 48.78% and 23.98%, respectively. These results establish $π$-SUB as a hyper-realistic and generalizable benchmark for developing the next generation of underwater image enhancement methods. The code and dataset are available at https://github.com/airl-iisc/pi-SUB

Problem

• Underwater Image Enhancement
• Synthetic Benchmark Dataset
• Synthetic-to-Real Gap
• Physics-Informed Modeling
• Generalizability

Innovation

• physics-informed
• underwater image enhancement
• synthetic benchmark dataset
• Jerlov water types
• hyper-realism