Concentric Spherical GNN for 3D Representation Learning

TL;DR

This paper introduces a multi-resolution spherical GNN architecture that effectively learns 3D representations invariant to orientation, improving classification accuracy on rotated 3D data.

Contribution

The authors propose a hierarchical multi-resolution spherical GNN that handles both mesh and point cloud data, including an efficient method to map point clouds to spherical images.

Findings

Improved accuracy on 3D classification with rotated data

Effective handling of both mesh and point cloud inputs

Enhanced spherical convolution techniques

Abstract

Learning 3D representations that generalize well to arbitrarily oriented inputs is a challenge of practical importance in applications varying from computer vision to physics and chemistry. We propose a novel multi-resolution convolutional architecture for learning over concentric spherical feature maps, of which the single sphere representation is a special case. Our hierarchical architecture is based on alternatively learning to incorporate both intra-sphere and inter-sphere information. We show the applicability of our method for two different types of 3D inputs, mesh objects, which can be regularly sampled, and point clouds, which are irregularly distributed. We also propose an efficient mapping of point clouds to concentric spherical images, thereby bridging spherical convolutions on grids with general point clouds. We demonstrate the effectiveness of our approach in improving state-of-the-art performance on 3D classification tasks with rotated data.