3D-UIR: 3D Gaussian for Underwater 3D Scene Reconstruction via Physics-Based Appearance-Medium Decouplin

2025-05-27

AI Summary

Underwater novel view synthesis suffers from non-uniform medium attenuation caused by water’s scattering and absorption—violating the homogeneous-medium assumption underlying conventional volumetric rendering, leading to artifacts and appearance inconsistency in methods like 3D Gaussian Splatting (3DGS). To address this, we propose the first physics-driven appearance–medium disentanglement framework tailored for underwater scenes. Our method introduces an underwater imaging model-guided appearance embedding and explicit medium representation; incorporates distance-aware pseudo-depth supervision to jointly optimize geometric fidelity and scale consistency; and integrates depth regularization with medium-aware scale penalization. Evaluated on multiple real-world underwater datasets, our approach achieves significant improvements in PSNR and SSIM, while enabling more accurate geometry reconstruction and physically consistent appearance recovery.

Technology Category

• Computer Vision: 3D Computer Vision
• Search and Optimization: Mixed Discrete/Continuous Search
• Reasoning under Uncertainty: Uncertainty Representations

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Novel view synthesis for underwater scene reconstruction presents unique challenges due to complex light-media interactions. Optical scattering and absorption in water body bring inhomogeneous medium attenuation interference that disrupts conventional volume rendering assumptions of uniform propagation medium. While 3D Gaussian Splatting (3DGS) offers real-time rendering capabilities, it struggles with underwater inhomogeneous environments where scattering media introduce artifacts and inconsistent appearance. In this study, we propose a physics-based framework that disentangles object appearance from water medium effects through tailored Gaussian modeling. Our approach introduces appearance embeddings, which are explicit medium representations for backscatter and attenuation, enhancing scene consistency. In addition, we propose a distance-guided optimization strategy that leverages pseudo-depth maps as supervision with depth regularization and scale penalty terms to improve geometric fidelity. By integrating the proposed appearance and medium modeling components via an underwater imaging model, our approach achieves both high-quality novel view synthesis and physically accurate scene restoration. Experiments demonstrate our significant improvements in rendering quality and restoration accuracy over existing methods. The project page is available at href{https://bilityniu.github.io/3D-UIR}{https://bilityniu.github.io/3D-UIR

Problem

• Addresses underwater scene reconstruction challenges from light-media interactions
• Overcomes inhomogeneous medium attenuation in 3D Gaussian Splatting
• Decouples object appearance from water effects via physics-based modeling

Innovation

• Physics-based appearance-medium decoupling framework
• Appearance embeddings for medium representation
• Distance-guided optimization with pseudo-depth supervision