Image Classification with Hierarchical Multigraph Networks

TL;DR

This paper explores the use of Hierarchical Multigraph Networks, a type of GCN, for image classification, demonstrating that they can outperform traditional CNNs on standard datasets by leveraging irregular inputs and multirelational data.

Contribution

The paper introduces best practices for designing GCNs for image classification, showing they can surpass CNN performance on several benchmark datasets.

Findings

GCNs can outperform CNNs on MNIST, CIFAR-10, and PASCAL datasets.

Hierarchical Multigraph Networks effectively utilize irregular image inputs.

Multirelational modeling enhances image classification accuracy.

Abstract

Graph Convolutional Networks (GCNs) are a class of general models that can learn from graph structured data. Despite being general, GCNs are admittedly inferior to convolutional neural networks (CNNs) when applied to vision tasks, mainly due to the lack of domain knowledge that is hardcoded into CNNs, such as spatially oriented translation invariant filters. However, a great advantage of GCNs is the ability to work on irregular inputs, such as superpixels of images. This could significantly reduce the computational cost of image reasoning tasks. Another key advantage inherent to GCNs is the natural ability to model multirelational data. Building upon these two promising properties, in this work, we show best practices for designing GCNs for image classification; in some cases even outperforming CNNs on the MNIST, CIFAR-10 and PASCAL image datasets.