SN-Graph: a Minimalist 3D Object Representation for Classification

TL;DR

This paper introduces SN-Graph, a minimalist 3D object representation using internal spheres and graph structures, which improves classification accuracy especially with fewer nodes and arbitrary rotations.

Contribution

The paper presents a novel 3D object representation method using internal spheres and graph neural networks, outperforming existing methods in accuracy and robustness.

Findings

SN-Graph achieves higher classification accuracy on ModelNet40.

It performs well with fewer nodes and arbitrary rotations.

The method is effective for 3D object analysis using GNNs.

Abstract

Using deep learning techniques to process 3D objects has achieved many successes. However, few methods focus on the representation of 3D objects, which could be more effective for specific tasks than traditional representations, such as point clouds, voxels, and multi-view images. In this paper, we propose a Sphere Node Graph (SN-Graph) to represent 3D objects. Specifically, we extract a certain number of internal spheres (as nodes) from the signed distance field (SDF), and then establish connections (as edges) among the sphere nodes to construct a graph, which is seamlessly suitable for 3D analysis using graph neural network (GNN). Experiments conducted on the ModelNet40 dataset show that when there are fewer nodes in the graph or the tested objects are rotated arbitrarily, the classification accuracy of SN-Graph is significantly higher than the state-of-the-art methods.