Image Classification with Hierarchical Multigraph Networks
This work addresses the challenge of making GCNs competitive with CNNs in vision tasks, which could reduce computational costs for image reasoning, though it appears incremental as it builds on existing GCN properties.
The paper tackled the problem of applying Graph Convolutional Networks (GCNs) to image classification by leveraging their ability to handle irregular inputs like superpixels and model multirelational data, resulting in best practices that sometimes outperform CNNs on MNIST, CIFAR-10, and PASCAL datasets.
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.