Neural Turtle Graphics for Modeling City Road Layouts

TL;DR

Neural Turtle Graphics (NTG) is a new generative model for spatial graphs that effectively models city road layouts, allowing for style control and high-quality synthesis, outperforming existing methods.

Contribution

Introduction of NTG, a neural network-based sequential generative model for city road layouts, with applications in synthesis and aerial road parsing.

Findings

NTG outperforms existing approaches on diverse metrics

Enables style control and sketch-based synthesis of road layouts

Achieves state-of-the-art results on the SpaceNet dataset

Abstract

We propose Neural Turtle Graphics (NTG), a novel generative model for spatial graphs, and demonstrate its applications in modeling city road layouts. Specifically, we represent the road layout using a graph where nodes in the graph represent control points and edges in the graph represent road segments. NTG is a sequential generative model parameterized by a neural network. It iteratively generates a new node and an edge connecting to an existing node conditioned on the current graph. We train NTG on Open Street Map data and show that it outperforms existing approaches using a set of diverse performance metrics. Moreover, our method allows users to control styles of generated road layouts mimicking existing cities as well as to sketch parts of the city road layout to be synthesized. In addition to synthesis, the proposed NTG finds uses in an analytical task of aerial road parsing. Experimental results show that it achieves state-of-the-art performance on the SpaceNet dataset.