Neural Turtle Graphics for Modeling City Road Layouts
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 represents road segm...
Main Authors: | , , , , , , , , |
---|---|
Other Authors: | |
Format: | Article |
Language: | English |
Published: |
Institute of Electrical and Electronics Engineers (IEEE)
2021
|
Online Access: | https://hdl.handle.net/1721.1/130552 |
_version_ | 1811085230322221056 |
---|---|
author | Chu, Hang Li, Daiqing Acuna, David Kar, Amlan Shugrina, Maria Wei, Xinkai Liu, Ming-Yu Torralba, Antonio Fidler, Sanja |
author2 | Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science |
author_facet | Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Chu, Hang Li, Daiqing Acuna, David Kar, Amlan Shugrina, Maria Wei, Xinkai Liu, Ming-Yu Torralba, Antonio Fidler, Sanja |
author_sort | Chu, Hang |
collection | MIT |
description | 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 represents 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 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 a part 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. |
first_indexed | 2024-09-23T13:05:43Z |
format | Article |
id | mit-1721.1/130552 |
institution | Massachusetts Institute of Technology |
language | English |
last_indexed | 2024-09-23T13:05:43Z |
publishDate | 2021 |
publisher | Institute of Electrical and Electronics Engineers (IEEE) |
record_format | dspace |
spelling | mit-1721.1/1305522022-10-01T12:59:40Z Neural Turtle Graphics for Modeling City Road Layouts Chu, Hang Li, Daiqing Acuna, David Kar, Amlan Shugrina, Maria Wei, Xinkai Liu, Ming-Yu Torralba, Antonio Fidler, Sanja Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science 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 represents 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 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 a part 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. 2021-05-04T14:20:38Z 2021-05-04T14:20:38Z 2020-02 2019-10 2021-04-15T17:18:41Z Article http://purl.org/eprint/type/ConferencePaper 9781728148038 2380-7504 https://hdl.handle.net/1721.1/130552 Chu, Hang et al. "Neural Turtle Graphics for Modeling City Road Layouts." 2019 IEEE/CVF International Conference on Computer Vision, October-November 2019, Seoul, South Korea, Institute of Electrical and Electronics Engineers, February 2020. © 2019 IEEE en http://dx.doi.org/10.1109/iccv.2019.00462 2019 IEEE/CVF International Conference on Computer Vision Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf Institute of Electrical and Electronics Engineers (IEEE) arXiv |
spellingShingle | Chu, Hang Li, Daiqing Acuna, David Kar, Amlan Shugrina, Maria Wei, Xinkai Liu, Ming-Yu Torralba, Antonio Fidler, Sanja Neural Turtle Graphics for Modeling City Road Layouts |
title | Neural Turtle Graphics for Modeling City Road Layouts |
title_full | Neural Turtle Graphics for Modeling City Road Layouts |
title_fullStr | Neural Turtle Graphics for Modeling City Road Layouts |
title_full_unstemmed | Neural Turtle Graphics for Modeling City Road Layouts |
title_short | Neural Turtle Graphics for Modeling City Road Layouts |
title_sort | neural turtle graphics for modeling city road layouts |
url | https://hdl.handle.net/1721.1/130552 |
work_keys_str_mv | AT chuhang neuralturtlegraphicsformodelingcityroadlayouts AT lidaiqing neuralturtlegraphicsformodelingcityroadlayouts AT acunadavid neuralturtlegraphicsformodelingcityroadlayouts AT karamlan neuralturtlegraphicsformodelingcityroadlayouts AT shugrinamaria neuralturtlegraphicsformodelingcityroadlayouts AT weixinkai neuralturtlegraphicsformodelingcityroadlayouts AT liumingyu neuralturtlegraphicsformodelingcityroadlayouts AT torralbaantonio neuralturtlegraphicsformodelingcityroadlayouts AT fidlersanja neuralturtlegraphicsformodelingcityroadlayouts |