Search papers, labs, and topics across Lattice.
AquaFlow introduces a monocular Gaussian Splatting framework tailored for underwater streaming reconstruction, addressing challenges like light attenuation and scattering that hinder camera pose tracking and scene geometry. By fine-tuning a 3D vision foundation model on extensive underwater datasets and employing a medium-guided incremental Gaussian initialization strategy, AquaFlow enhances both pose and pointmap estimation. Evaluated across 62 diverse underwater trajectories, it achieves state-of-the-art performance, reducing localization error by 13.2% and improving PSNR by 4.74 dB over existing methods.
Underwater reconstruction just got a major upgrade, with AquaFlow slashing localization error by over 13% while enhancing image quality.
Recent monocular 3D Gaussian Splatting (3DGS) streaming reconstruction methods have achieved impressive performance by balancing reconstruction quality and efficiency. However, extending these frameworks to underwater scenes remains challenging due to severe visual degradation, such as light attenuation and scattering, which degrades camera pose tracking and distorts scene geometry. To address these challenges, we propose AquaFlow, a monocular Gaussian Splatting streaming reconstruction framework for efficient and high-fidelity underwater reconstruction. Specifically, AquaFlow fine-tunes a 3D vision foundation model on large-scale underwater data for robust pose and pointmap estimation, and introduces a medium-guided incremental Gaussian initialization strategy for streaming mapping. Furthermore, we develop a streaming-compatible hybrid scene representation that integrates structured, distance-conditioned neural Gaussians with a physics-inspired optical model to compensate for underwater image formation effects, enabling accurate scene reconstruction. We evaluate AquaFlow on a comprehensive dataset of 62 diverse underwater trajectories, collected from both public benchmarks and in-the-wild web videos across various scales. Extensive experiments demonstrate that AquaFlow achieves state-of-the-art tracking and rendering performance, reducing average localization error by 13.2% and improving PSNR by 4.74 dB compared to WaterSplat-SLAM.
