Point Cloud Upsampling and Normal Estimation using Deep Learning for Robust Surface Reconstruction

TL;DR

This paper introduces a deep learning method for upsampling low-density, noisy point clouds with normals to produce smoother, more complete point clouds that enable better surface reconstruction.

Contribution

A novel deep learning architecture that jointly estimates point positions and normals for improved surface reconstruction from sparse point clouds.

Findings

Enhanced point cloud density and quality.

More accurate surface normals and smoother reconstructions.

Closer approximation to ground truth surfaces.

Abstract

The reconstruction of real-world surfaces is on high demand in various applications. Most existing reconstruction approaches apply 3D scanners for creating point clouds which are generally sparse and of low density. These points clouds will be triangulated and used for visualization in combination with surface normals estimated by geometrical approaches. However, the quality of the reconstruction depends on the density of the point cloud and the estimation of the surface normals. In this paper, we present a novel deep learning architecture for point cloud upsampling that enables subsequent stable and smooth surface reconstruction. A noisy point cloud of low density with corresponding point normals is used to estimate a point cloud with higher density and appendant point normals. To this end, we propose a compound loss function that encourages the network to estimate points that lie on a surface including normals accurately predicting the orientation of the surface. Our results show the benefit of estimating normals together with point positions. The resulting point cloud is smoother, more complete, and the final surface reconstruction is much closer to ground truth.