Deep Point Cloud Normal Estimation via Triplet Learning

TL;DR

This paper introduces a novel two-phase deep learning approach for normal estimation on 3D point clouds, improving accuracy at sharp features and robustness to noise by learning local patch representations with triplet networks.

Contribution

It proposes a triplet learning-based feature encoding method combined with a simple normal regression network, enhancing normal estimation accuracy and feature preservation.

Findings

Better preservation of sharp features in normal estimation.

Improved accuracy on CAD-like shapes.

Smaller network size with competitive performance.

Abstract

Normal estimation on 3D point clouds is a fundamental problem in 3D vision and graphics. Current methods often show limited accuracy in predicting normals at sharp features (e.g., edges and corners) and less robustness to noise. In this paper, we propose a novel normal estimation method for point clouds. It consists of two phases: (a) feature encoding which learns representations of local patches, and (b) normal estimation that takes the learned representation as input and regresses the normal vector. We are motivated that local patches on isotropic and anisotropic surfaces have similar or distinct normals, and that separable features or representations can be learned to facilitate normal estimation. To realise this, we first construct triplets of local patches on 3D point cloud data, and design a triplet network with a triplet loss for feature encoding. We then design a simple network with several MLPs and a loss function to regress the normal vector. Despite having a smaller network size compared to most other methods, experimental results show that our method preserves sharp features and achieves better normal estimation results on CAD-like shapes.