Learning Compact Geometric Features

TL;DR

This paper introduces a data-driven method to learn compact, accurate, and robust geometric features for unstructured point clouds, enhancing registration tasks in robotics and 3D vision.

Contribution

It proposes a deep learning approach to automatically learn local geometric features that outperform manually crafted descriptors in terms of compactness and accuracy.

Findings

Learned features are more compact than existing descriptors.

The features achieve higher accuracy in geometric registration.

The approach generalizes across different point cloud datasets.

Abstract

We present an approach to learning features that represent the local geometry around a point in an unstructured point cloud. Such features play a central role in geometric registration, which supports diverse applications in robotics and 3D vision. Current state-of-the-art local features for unstructured point clouds have been manually crafted and none combines the desirable properties of precision, compactness, and robustness. We show that features with these properties can be learned from data, by optimizing deep networks that map high-dimensional histograms into low-dimensional Euclidean spaces. The presented approach yields a family of features, parameterized by dimension, that are both more compact and more accurate than existing descriptors.