Voxurf: Voxel-based Efficient and Accurate Neural Surface Reconstruction

TL;DR

Voxurf introduces a voxel-based neural surface reconstruction method that significantly improves efficiency and accuracy by addressing prior limitations in fine detail recovery and spatial coherence.

Contribution

The paper proposes Voxurf, a novel voxel-based approach with a two-stage training, dual color network, and hierarchical features to enhance surface reconstruction quality and speed.

Findings

Voxurf outperforms previous methods in reconstruction quality.

Voxurf achieves a 20x faster training speed on the DTU benchmark.

The approach effectively captures fine geometric details.

Abstract

Neural surface reconstruction aims to reconstruct accurate 3D surfaces based on multi-view images. Previous methods based on neural volume rendering mostly train a fully implicit model with MLPs, which typically require hours of training for a single scene. Recent efforts explore the explicit volumetric representation to accelerate the optimization via memorizing significant information with learnable voxel grids. However, existing voxel-based methods often struggle in reconstructing fine-grained geometry, even when combined with an SDF-based volume rendering scheme. We reveal that this is because 1) the voxel grids tend to break the color-geometry dependency that facilitates fine-geometry learning, and 2) the under-constrained voxel grids lack spatial coherence and are vulnerable to local minima. In this work, we present Voxurf, a voxel-based surface reconstruction approach that is both efficient and accurate. Voxurf addresses the aforementioned issues via several key designs, including 1) a two-stage training procedure that attains a coherent coarse shape and recovers fine details successively, 2) a dual color network that maintains color-geometry dependency, and 3) a hierarchical geometry feature to encourage information propagation across voxels. Extensive experiments show that Voxurf achieves high efficiency and high quality at the same time. On the DTU benchmark, Voxurf achieves higher reconstruction quality with a 20x training speedup compared to previous fully implicit methods. Our code is available at https://github.com/wutong16/Voxurf.