Swimm3R：结合介质感知SfM的水下三维重建方法
Swimm3R: Splatting with Medium-aware SfM for Underwater 3D Reconstruction
浏览论文内容
中文总结 AI 辅助
该研究提出Swimm3R框架，结合介质感知SfM与水下Beta高斯溅射，建立巴巴多斯水下视频数据集，在水下三维重建中提升了PSNR与下游定位性能。
中文摘要 AI 辅助
我们提出了Swimm3R，这是一个将介质感知运动恢复结构（Structure-from-Motion, SfM）与水下Beta高斯溅射（Underwater Beta Splatting）相结合的统一框架，以解决水下三维重建中由散射和衰减导致的失效问题。Swimm3R将空气中的几何先验知识提炼为前馈骨干网络，并使用物理头来回归水下图像形成参数、相机位姿以及恢复的点云。此外，我们引入了水下Beta高斯溅射，它通过Beta基元与散射感知几何梯度扩展了高斯溅射，以实现稳定的水下几何表示。我们还建立了巴巴多斯水下视频数据集（Barbados underwater video dataset），以证明我们的方法在具有挑战性的水下环境中的有效性。在该数据集上，Swimm3R在具有挑战性的散射条件下能稳健地恢复水下场景结构，生成连贯的海底几何结构。利用这些预测的点云，所提出的水下Beta高斯溅射相较于WaterSplatting将平均峰值信噪比（PSNR）提高了1.47分贝，同时下游定位性能在RRA@15和RTA@15指标上分别提升了2.0和2.4个百分点。
英文摘要
We propose Swimm3R, a unified framework that combines medium-aware structure-from-motion (SfM) with Underwater Beta Splatting to address scattering- and attenuation-induced failures in underwater 3D reconstruction. Swimm3R distills in-air geometric priors into a feed-forward backbone and uses a physics head to regress underwater image-formation parameters, camera poses, and restored point clouds. Additionally, we introduce Underwater Beta Splatting, which extends Gaussian splatting with Beta primitives and scattering-aware geometric gradients for stable underwater geometry representation. We further establish the Barbados underwater video dataset to demonstrate the effectiveness of our method in challenging underwater environments. On this dataset, Swimm3R robustly recovers underwater scene structure under challenging scattering conditions, yielding coherent seafloor geometry. Using these predicted point clouds, the proposed Underwater Beta Splatting improves average PSNR by $1.47$ dB over WaterSplatting while increasing downstream localization performance by $2.0$ and $2.4$ percentage points in RRA@15 and RTA@15, respectively.
发表机构
- University of Minnesota–Twin Cities（明尼苏达大学双城分校）
- Minnesota Robotics Institute (MnRI)（明尼苏达机器人研究所）
机构由 AI 辅助整理，请以论文原文为准。
