R-PointHop: A Green, Accurate, and Unsupervised Point Cloud Registration Method

TL;DR

R-PointHop is an unsupervised, efficient, and rotation-invariant 3D point cloud registration method that leverages hierarchical local features and local reference frames to achieve accurate alignment with minimal training.

Contribution

It introduces a novel unsupervised registration approach using local reference frames and hierarchical features, reducing model size and training time compared to deep learning methods.

Findings

Effective on multiple datasets including 3DMatch, ModelNet40, and Stanford Bunny.

Achieves smaller registration errors than deep learning methods.

Significantly reduces model size and training time.

Abstract

Inspired by the recent PointHop classification method, an unsupervised 3D point cloud registration method, called R-PointHop, is proposed in this work. R-PointHop first determines a local reference frame (LRF) for every point using its nearest neighbors and finds local attributes. Next, R-PointHop obtains local-to-global hierarchical features by point downsampling, neighborhood expansion, attribute construction and dimensionality reduction steps. Thus, point correspondences are built in hierarchical feature space using the nearest neighbor rule. Afterwards, a subset of salient points with good correspondence is selected to estimate the 3D transformation. The use of the LRF allows for invariance of the hierarchical features of points with respect to rotation and translation, thus making R-PointHop more robust at building point correspondence, even when the rotation angles are large. Experiments are conducted on the 3DMatch, ModelNet40, and Stanford Bunny datasets, which demonstrate the effectiveness of R-PointHop for 3D point cloud registration. R-PointHop's model size and training time are an order of magnitude smaller than those of deep learning methods, and its registration errors are smaller, making it a green and accurate solution. Our codes are available on GitHub.