License: Creative Commons Attribution 3.0 Unported license (CC BY 3.0)
When quoting this document, please refer to the following
DOI: 10.4230/LIPIcs.GISCIENCE.2018.39
URN: urn:nbn:de:0030-drops-93672
URL: http://dagstuhl.sunsite.rwth-aachen.de/volltexte/2018/9367/
Go to the corresponding LIPIcs Volume Portal

Kim, Seongyong ; Park, Seula ; Yu, Kiyun

Application of Style Transfer in the Vectorization Process of Floorplans (Short Paper)

pdf-format:
LIPIcs-GISCIENCE-2018-39.pdf (0.5 MB)

Abstract

As the market for indoor spatial information burgeons, the construction of indoor spatial databases consequently gain attention. Since floorplans are portable records of buildings, they are an indispensable source for the efficient construction of indoor environments. However, as previous research on floorplan information retrieval usually targeted specific formats, a system for constructing spatial information must include heuristic refinement steps. This study aims to convert diverse floorplans into an integrated format using the style transfer by deep networks. Our deep networks mimic a robust perception of human that recognize the cell structure of floorplans under various formats. The integrated format ensures that unified post-processing steps are required to the vectorization of floorplans. Through this process, indoor spatial information is constructed in a pragmatic way, using a plethora of architectural floorplans.

BibTeX - Entry

@InProceedings{kim_et_al:LIPIcs:2018:9367,
  author =	{Seongyong Kim and Seula Park and Kiyun Yu},
  title =	{{Application of Style Transfer in the Vectorization Process of Floorplans (Short Paper)}},
  booktitle =	{10th International Conference on Geographic Information  Science (GIScience 2018)},
  pages =	{39:1--39:6},
  series =	{Leibniz International Proceedings in Informatics (LIPIcs)},
  ISBN =	{978-3-95977-083-5},
  ISSN =	{1868-8969},
  year =	{2018},
  volume =	{114},
  editor =	{Stephan Winter and Amy Griffin and Monika Sester},
  publisher =	{Schloss Dagstuhl--Leibniz-Zentrum fuer Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{http://drops.dagstuhl.de/opus/volltexte/2018/9367},
  URN =		{urn:nbn:de:0030-drops-93672},
  doi =		{10.4230/LIPIcs.GISCIENCE.2018.39},
  annote =	{Keywords: Floorplan, Vectorising, Style Transfer, Generative Adversarial Networks}
}

Keywords: Floorplan, Vectorising, Style Transfer, Generative Adversarial Networks
Collection: 10th International Conference on Geographic Information Science (GIScience 2018)
Issue Date: 2018
Date of publication: 02.08.2018

DROPS-Home | Fulltext Search | Imprint | Privacy Published by LZI