Visual Feature Encoding for GNNs on Road Networks

TL;DR

This paper introduces a method combining vision backbone networks with graph neural networks to improve road network classification by encoding satellite imagery, demonstrating the benefits of fine-tuning on remote sensing datasets.

Contribution

It presents a novel architecture integrating state-of-the-art vision models with GNNs for road network analysis, including a transfer learning approach with remote sensing data.

Findings

Visual feature encoders outperform low-level features.

Fine-tuning on remote sensing datasets improves GNN performance.

Pretraining on NWPU-RESISC45 enhances classification accuracy.

Abstract

In this work, we present a novel approach to learning an encoding of visual features into graph neural networks with the application on road network data. We propose an architecture that combines state-of-the-art vision backbone networks with graph neural networks. More specifically, we perform a road type classification task on an Open Street Map road network through encoding of satellite imagery using various ResNet architectures. Our architecture further enables fine-tuning and a transfer-learning approach is evaluated by pretraining on the NWPU-RESISC45 image classification dataset for remote sensing and comparing them to purely ImageNet-pretrained ResNet models as visual feature encoders. The results show not only that the visual feature encoders are superior to low-level visual features, but also that the fine-tuning of the visual feature encoder to a general remote sensing dataset such as NWPU-RESISC45 can further improve the performance of a GNN on a machine learning task like road type classification.