Neural-IMLS: Self-supervised Implicit Moving Least-Squares Network for Surface Reconstruction

TL;DR

Neural-IMLS introduces a self-supervised neural network that learns a noise-resistant signed distance function for accurate surface reconstruction from noisy, unoriented point clouds, outperforming existing methods.

Contribution

It proposes a novel self-supervised framework combining MLP and IMLS for robust surface reconstruction from raw point clouds.

Findings

Reconstructs faithful shapes on noisy and incomplete data.

Outperforms existing methods on synthetic and real scans.

Produces accurate surface approximations with mutual regularization.

Abstract

Surface reconstruction is very challenging when the input point clouds, particularly real scans, are noisy and lack normals. Observing that the Multilayer Perceptron (MLP) and the implicit moving least-square function (IMLS) provide a dual representation of the underlying surface, we introduce Neural-IMLS, a novel approach that directly learns the noise-resistant signed distance function (SDF) from unoriented raw point clouds in a self-supervised fashion. We use the IMLS to regularize the distance values reported by the MLP while using the MLP to regularize the normals of the data points for running the IMLS. We also prove that at the convergence, our neural network, benefiting from the mutual learning mechanism between the MLP and the IMLS, produces a faithful SDF whose zero-level set approximates the underlying surface. We conducted extensive experiments on various benchmarks, including synthetic scans and real scans. The experimental results show that {\em Neural-IMLS} can reconstruct faithful shapes on various benchmarks with noise and missing parts. The source code can be found at~\url{https://github.com/bearprin/Neural-IMLS}.