Sphere-Guided Training of Neural Implicit Surfaces

TL;DR

This paper introduces a sphere-guided training method for neural implicit surfaces that improves reconstruction quality by efficiently excluding empty scene regions during volumetric ray marching, leading to more detailed 3D reconstructions.

Contribution

The authors propose a novel coarse sphere-based surface representation that enhances neural implicit surface training by reducing unnecessary computations and improving detail fidelity.

Findings

Improved reconstruction quality across multiple datasets.

Enhanced efficiency in volumetric ray marching.

Consistent performance gains in various implicit surface models.

Abstract

In recent years, neural distance functions trained via volumetric ray marching have been widely adopted for multi-view 3D reconstruction. These methods, however, apply the ray marching procedure for the entire scene volume, leading to reduced sampling efficiency and, as a result, lower reconstruction quality in the areas of high-frequency details. In this work, we address this problem via joint training of the implicit function and our new coarse sphere-based surface reconstruction. We use the coarse representation to efficiently exclude the empty volume of the scene from the volumetric ray marching procedure without additional forward passes of the neural surface network, which leads to an increased fidelity of the reconstructions compared to the base systems. We evaluate our approach by incorporating it into the training procedures of several implicit surface modeling methods and observe uniform improvements across both synthetic and real-world datasets. Our codebase can be accessed via the project page: https://andreeadogaru.github.io/SphereGuided