SolidGS: Consolidating Gaussian Surfel Splatting for Sparse-View Surface Reconstruction

TL;DR

SolidGS introduces a novel consolidation of Gaussian splats with a solid kernel function, geometrical regularization, and monocular normal estimation, significantly enhancing sparse-view surface reconstruction quality over existing methods.

Contribution

The paper proposes SolidGS, a new Gaussian splatting approach that improves surface reconstruction from sparse views by consolidating Gaussians with a solid kernel and regularization techniques.

Findings

Outperforms existing Gaussian splatting methods on DTU, Tanks-and-Temples, LLFF datasets.

Achieves higher quality surface reconstruction from sparse multi-view images.

Demonstrates robustness and accuracy improvements in surface geometry reconstruction.

Abstract

Gaussian splatting has achieved impressive improvements for both novel-view synthesis and surface reconstruction from multi-view images. However, current methods still struggle to reconstruct high-quality surfaces from only sparse view input images using Gaussian splatting. In this paper, we propose a novel method called SolidGS to address this problem. We observed that the reconstructed geometry can be severely inconsistent across multi-views, due to the property of Gaussian function in geometry rendering. This motivates us to consolidate all Gaussians by adopting a more solid kernel function, which effectively improves the surface reconstruction quality. With the additional help of geometrical regularization and monocular normal estimation, our method achieves superior performance on the sparse view surface reconstruction than all the Gaussian splatting methods and neural field methods on the widely used DTU, Tanks-and-Temples, and LLFF datasets.