Neural Turtle Graphics for Modeling City Road Layouts
This addresses the problem of generating and analyzing urban road networks for applications like city planning and simulation, though it is incremental as it builds on existing generative models for graphs.
The paper tackles the problem of modeling city road layouts by proposing Neural Turtle Graphics (NTG), a sequential generative model for spatial graphs, and shows that it outperforms existing approaches on diverse metrics and achieves state-of-the-art performance on the SpaceNet dataset for aerial road parsing.
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.