DeltaConv: Anisotropic Operators for Geometric Deep Learning on Point Clouds

TL;DR

DeltaConv introduces anisotropic convolution layers for 3D point-cloud data that leverage vector calculus operators, enabling explicit directional processing and improving performance on benchmarks.

Contribution

The paper proposes DeltaConv, a novel anisotropic convolution layer for point clouds that separates scalar and vector streams using geometric operators, enhancing directional feature learning.

Findings

Matches or surpasses state-of-the-art on benchmarks

Speeds up training and inference

Robust and easy to implement

Abstract

Learning from 3D point-cloud data has rapidly gained momentum, motivated by the success of deep learning on images and the increased availability of 3D~data. In this paper, we aim to construct anisotropic convolution layers that work directly on the surface derived from a point cloud. This is challenging because of the lack of a global coordinate system for tangential directions on surfaces. We introduce DeltaConv, a convolution layer that combines geometric operators from vector calculus to enable the construction of anisotropic filters on point clouds. Because these operators are defined on scalar- and vector-fields, we separate the network into a scalar- and a vector-stream, which are connected by the operators. The vector stream enables the network to explicitly represent, evaluate, and process directional information. Our convolutions are robust and simple to implement and match or improve on state-of-the-art approaches on several benchmarks, while also speeding up training and inference.