From Blobs to Spokes: High-Fidelity Surface Reconstruction via Oriented Gaussians

2026-04-08

AI Summary

This work addresses the challenge that 3D Gaussian Splatting lacks an explicit global geometric field, making it difficult to extract complete, watertight surface meshes. To overcome this limitation, the authors propose a Gaussian-oriented closed-surface modeling approach that introduces learnable oriented normals for each Gaussian, enabling the construction of a closed-form occupancy field and a normal field. The method incorporates a consistency loss and a specialized densification strategy, alongside a region-adaptive Primal Adaptive Meshing algorithm and an improved differentiable rasterizer. Evaluated on the DTU and Tanks and Temples benchmarks, the approach achieves state-of-the-art performance, faithfully reconstructing fine and thin structures—such as bicycle spokes—and producing compact, geometrically complete, watertight meshes.

Abstract

3D Gaussian Splatting (3DGS) has revolutionized fast novel view synthesis, yet its opacity-based formulation makes surface extraction fundamentally difficult. Unlike implicit methods built on Signed Distance Fields or occupancy, 3DGS lacks a global geometric field, forcing existing approaches to resort to heuristics such as TSDF fusion of blended depth maps. Inspired by the Objects as Volumes framework, we derive a principled occupancy field for Gaussian Splatting and show how it can be used to extract highly accurate watertight meshes of complex scenes. Our key contribution is to introduce a learnable oriented normal at each Gaussian element and to define an adapted attenuation formulation, which leads to closed-form expressions for both the normal and occupancy fields at arbitrary locations in space. We further introduce a novel consistency loss and a dedicated densification strategy to enforce Gaussians to wrap the entire surface by closing geometric holes, ensuring a complete shell of oriented primitives. We modify the differentiable rasterizer to output depth as an isosurface of our continuous model, and introduce Primal Adaptive Meshing for Region-of-Interest meshing at arbitrary resolution. We additionally expose fundamental biases in standard surface evaluation protocols and propose two more rigorous alternatives. Overall, our method Gaussian Wrapping sets a new state-of-the-art on DTU and Tanks and Temples, producing complete, watertight meshes at a fraction of the size of concurrent work-recovering thin structures such as the notoriously elusive bicycle spokes.

Problem

• 3D Gaussian Splatting
• surface reconstruction
• occupancy field
• watertight mesh
• geometric representation

Innovation

• Oriented Gaussians
• Occupancy Field
• Surface Reconstruction
• Watertight Meshing
• 3D Gaussian Splatting