Neighbourhood-Insensitive Point Cloud Normal Estimation Network

TL;DR

This paper presents a self-attention-based neural network for point cloud normal estimation that adapts to large neighborhoods, achieving state-of-the-art accuracy with smaller, faster models.

Contribution

A novel self-attention mechanism with learnable temperature enables effective normal estimation across large neighborhoods, outperforming existing methods.

Findings

Achieves 94.1% accuracy in normal estimation

Model is 25x smaller and 12x faster than previous methods

Normal estimation improves ICP convergence speed

Abstract

We introduce a novel self-attention-based normal estimation network that is able to focus softly on relevant points and adjust the softness by learning a temperature parameter, making it able to work naturally and effectively within a large neighbourhood range. As a result, our model outperforms all existing normal estimation algorithms by a large margin, achieving 94.1% accuracy in comparison with the previous state of the art of 91.2%, with a 25x smaller model and 12x faster inference time. We also use point-to-plane Iterative Closest Point (ICP) as an application case to show that our normal estimations lead to faster convergence than normal estimations from other methods, without manually fine-tuning neighbourhood range parameters. Code available at https://code.active.vision.