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

TL;DR

This paper tackles the challenge of reconstructing accurate surfaces from sparse-view images by introducing SolidGS, a consolidated Gaussian surfel representation. SolidGS enforces solid, view-consistent geometry through a global solidness factor and augments optimization with geometric regularizations and monocular normal priors, along with virtual-view supervision. The method achieves state-of-the-art results for sparse-view surface reconstruction and novel-view synthesis on DTU, Tanks-and-Temples, and LLFF, while maintaining fast training times. The work highlights the practical impact of consolidating Gaussian primitives and leveraging priors to enable high-fidelity geometry from minimal input, with robust performance across datasets and lighting/viewpoint conditions.

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.

Figures (12)

• Figure 1: Overview of SolidGS. We present SolidGS, which reconstructs a consolidated Gaussian field from sparse inputs. Given only three input views, our approach enables high-precision and detailed mesh extraction, and high-quality novel view synthesis, achieved within just three minutes.
• Figure 2: Illustration of SolidGS. (a) The Gausian functions $\mathcal{G}_i$ of the original 3DGS and SolidGS in \ref{['eq:solidgaussian']}. (b) Visualization of ray-plane intersections and blended depths $d_j, j \in \{1,2\}$ for two types of Gaussians. In the original Gaussian, as rays deviate from the Gaussian center, the depth is blended with the background Gaussians, leading to inconsistent multi-view geometry in 3D space. Meanwhile, our SolidGS is opaque in most of its effective areas, giving out consistent geometry regardless of view directions.
• Figure 3: SolidGS Framework. With 3 input views, we initialize the camera pose with COLMAP and the point clouds with MVSFormer. Virtual views are generated by linear interpolation between pairs of training views to provide additional geometric regularization. We represent the scene with our SolidGS and gradually enhance the solidity during training. SolidGS are optimized with photometric loss, monocular normal loss, and geometric regularization (normal consistency loss and depth distortion loss).
• Figure 4: Qualitative Mesh Results on DTU Dataset. We show the reconstructed meshes with the closest input view for reference. Meshes are fused using the TSDF + Marching Cube method for explicit methods including PGSR chen2024pgsr, 2DGS huang20242d, MVPGS xu2024mvpgs, and our method. NeuSurf huang2024neusurf maintains an SDF, from which the mesh is extracted using the Marching Cube.
• Figure 5: Qualitative Mesh Results on TNT Dataset. All meshes are extracted using TSDF + Marching Cube from Gaussians.