Geometric Attention for Prediction of Differential Properties in 3D Point Clouds

TL;DR

This paper introduces a geometric attention mechanism for predicting differential geometric properties like normals and feature lines in 3D point clouds, enhancing surface reconstruction and meshing.

Contribution

It proposes a novel learnable geometric attention method to better select neighborhoods and incorporate geometric relations in point cloud analysis.

Findings

Improves normal vector prediction accuracy

Enhances feature line extraction from raw point clouds

Demonstrates effectiveness in geometry processing tasks

Abstract

Estimation of differential geometric quantities in discrete 3D data representations is one of the crucial steps in the geometry processing pipeline. Specifically, estimating normals and sharp feature lines from raw point cloud helps improve meshing quality and allows us to use more precise surface reconstruction techniques. When designing a learnable approach to such problems, the main difficulty is selecting neighborhoods in a point cloud and incorporating geometric relations between the points. In this study, we present a geometric attention mechanism that can provide such properties in a learnable fashion. We establish the usefulness of the proposed technique with several experiments on the prediction of normal vectors and the extraction of feature lines.