Mastering Spatial Graph Prediction of Road Networks

TL;DR

This paper introduces a reinforcement learning framework for predicting road networks from satellite images by iteratively refining graph structures, leading to improved accuracy and robustness, especially under occlusions.

Contribution

The paper presents a novel RL-based method for spatial graph prediction that captures complex topological information beyond standard supervised techniques.

Findings

Enhanced performance on benchmark datasets

Superior reasoning about graph topology with tree-based search

Robust predictions under occlusions using synthetic benchmarks

Abstract

Accurately predicting road networks from satellite images requires a global understanding of the network topology. We propose to capture such high-level information by introducing a graph-based framework that simulates the addition of sequences of graph edges using a reinforcement learning (RL) approach. In particular, given a partially generated graph associated with a satellite image, an RL agent nominates modifications that maximize a cumulative reward. As opposed to standard supervised techniques that tend to be more restricted to commonly used surrogate losses, these rewards can be based on various complex, potentially non-continuous, metrics of interest. This yields more power and flexibility to encode problem-dependent knowledge. Empirical results on several benchmark datasets demonstrate enhanced performance and increased high-level reasoning about the graph topology when using a tree-based search. We further highlight the superiority of our approach under substantial occlusions by introducing a new synthetic benchmark dataset for this task.