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

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}.

References in corpus (9)

• Point2Mesh: A Self-Prior for Deformable Meshes
• Thingi10K: A Dataset of 10,000 3D-Printing Models
• Neural Dual Contouring
• Iterative Poisson Surface Reconstruction (iPSR) for Unoriented Points
• Dual Octree Graph Networks for Learning Adaptive Volumetric Shape Representations
• Kaolin: A PyTorch Library for Accelerating 3D Deep Learning Research
• Orienting Point Clouds with Dipole Propagation
• Surface Reconstruction from Point Clouds: A Survey and a Benchmark
• A Survey and Benchmark of Automatic Surface Reconstruction from Point Clouds

Cited by in corpus (2)

• PPSURF: Combining Patches and Point Convolutions for Detailed Surface Reconstruction
• LFS-Aware Surface Reconstruction from Unoriented 3D Point Clouds