3DFacilities: Annotated 3D reconstructions of building facilities

Thomas Czerniawski, Fernanda Leite

Research output: Chapter in Book/Report/Conference proceedingConference contribution

23 Scopus citations

Abstract

Scan-to-BIM is the process of converting 3D reconstructions into building information models (BIM). Currently, it involves manual tracing of point clouds by human users in BIM authoring tools, with some automation functionality available for walls, floors, windows, doors, and piping. Emerging semantic segmentation methods demonstrate a level of versatility that could extend the capabilities of automated Scan-to-BIM well past the limited existing object categories. The accuracy of supervised deep learning methods in the context of 3D scene segmentation has experienced rapid improvement over the past year due to the recent availability of large, annotated datasets of indoor spaces. Unfortunately, the semantic object categories in the available datasets do not cover many essential BIM object categories, such as heating, ventilation and air-conditioning (HVAC), and plumbing systems. In an effort to leverage the success of deep learning for Scan-to-BIM, we present 3DFacilities, an annotated dataset of 3D reconstructions of building facilities. The dataset contains over 11,000 individual RGB-D frames comprising 50 scene reconstructions annotated with 3D camera poses and per-vertex and per-pixel annotations. Our dataset is available at https://thomasczerniawski.com/3dfacilities/.

Original languageEnglish (US)
Title of host publicationAdvanced Computing Strategies for Engineering - 25th EG-ICE International Workshop 2018, Proceedings
EditorsBernd Domer, Ian F. Smith
PublisherSpringer Verlag
Pages186-200
Number of pages15
ISBN (Print)9783319916347
DOIs
StatePublished - 2018
Externally publishedYes
Event25th Workshop of the European Group for Intelligent Computing in Engineering, EG-ICE 2018 - Lausanne, Switzerland
Duration: Jun 10 2018Jun 13 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10863 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference25th Workshop of the European Group for Intelligent Computing in Engineering, EG-ICE 2018
Country/TerritorySwitzerland
CityLausanne
Period6/10/186/13/18

Keywords

  • Building information modeling
  • Computer vision
  • Deep learning

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science

Fingerprint

Dive into the research topics of '3DFacilities: Annotated 3D reconstructions of building facilities'. Together they form a unique fingerprint.

Cite this