Scholarlink

<p>In underwater images, most useful features are occluded by water. The extent of the occlusion depends on imaging geometry and can vary even across a sequence of burst images. As a result, 3D reconstruction methods robust on in-air scenes, like Neural Radiance Field methods (NeRFs) or 3D Gaussian Splatting (3DGS), can degrade substantially on underwater scenes. While a recent underwater adaptation of NeRFs achieved high-quality results, it is impractically slow: reconstruction takes hours and its rendering rate, in frames per second (FPS), is less than 1. Here, we present a new method that takes only a few minutes for reconstruction and renders novel underwater scenes at 140 FPS. Named <i>Gaussian Splashing</i>, our method unifies the strengths and speed of 3DGS with an image formation model for capturing scattering, introducing innovations in the rendering and depth estimation procedures and in the 3DGS loss function. Across existing datasets and a new dataset we collected, Gaussian Splashing preserves competitive reconstruction quality while improving rendering speed, and it is particularly effective for distant scene details affected by backscatter.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Gaussian splashing enables direct volumetric rendering underwater

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月28日

查看全文
  • Nir Mualem,
  • Roy Amoyal,
  • Oren Freifeld,
  • Derya Akkaynak

摘要

In underwater images, most useful features are occluded by water. The extent of the occlusion depends on imaging geometry and can vary even across a sequence of burst images. As a result, 3D reconstruction methods robust on in-air scenes, like Neural Radiance Field methods (NeRFs) or 3D Gaussian Splatting (3DGS), can degrade substantially on underwater scenes. While a recent underwater adaptation of NeRFs achieved high-quality results, it is impractically slow: reconstruction takes hours and its rendering rate, in frames per second (FPS), is less than 1. Here, we present a new method that takes only a few minutes for reconstruction and renders novel underwater scenes at 140 FPS. Named Gaussian Splashing, our method unifies the strengths and speed of 3DGS with an image formation model for capturing scattering, introducing innovations in the rendering and depth estimation procedures and in the 3DGS loss function. Across existing datasets and a new dataset we collected, Gaussian Splashing preserves competitive reconstruction quality while improving rendering speed, and it is particularly effective for distant scene details affected by backscatter.