Three-Dimensional Information Retrieval (3DIR): A Graph Theoretic Formulation for Exploiting 3D Geometry and Model Topology in Information Retrieval

Three-Dimensional Information Retrieval (3DIR): A Graph Theoretic Formulation for Exploiting 3D Geometry and Model Topology in Information Retrieval

Peter Demian (Loughborough University, UK), Kirti Ruikar (Loughborough University, UK) and Anne Morris (Loughborough University, UK)
Copyright: © 2018 |Pages: 21
DOI: 10.4018/978-1-5225-5625-1.ch005


The 3DIR project investigated the use of 3D visualization to formulate queries, compute the relevance of information items, and visualize search results. Workshops identified the user needs. Based on these, a graph theoretic formulation was created to inform the emerging system architecture. A prototype was developed. This enabled relationships between 3D objects to be used to widen a search. An evaluation of the prototype demonstrated that a tight coupling between text-based retrieval and 3D models could enhance information retrieval but add an extra layer of complexity.
Chapter Preview

Building design, construction and operation are information intensive activities. For example, even over a decade ago in the UK construction industry, on average, one computer-aided design (CAD) document was produced for every 9 m2 of building floor space (Gray and Hughes 2001). Several researchers (Leslie, 1996; Veeramani and Russell, 2000; Ugwu, 2005) have reported the problem of “information overload” in the construction sector.

BIMs are following this general trend and becoming more information-rich. Regarding volumes of information specifically in BIMs, Demian and Walters (2014) identified BIM platforms as a particularly favourable communication medium in construction, compared to extranets, email and Enterprise Resource Planning systems. Charalambous et al. (2013) reported the advantages of BIM over documents and extranets. Although no absolute measures of the quantities of information were found, the implication from studies such as those is that BIMs are increasingly information-rich.

Information retrieval techniques have been used in construction to retrieve reusable designs (‎Demian and Fruchter 2005). Beyond text, Brilakis and Soibelman (‎2008) automatically identify particular features in construction site photographs with a view subsequently to using information retrieval techniques to manage collections of photographs. Bridging textual and geometric content, Caldas et al. ‎(2002) propose techniques for automatically classifying construction documents based on project CAD components. Lin and Soibelman (2009) augment standard information retrieval techniques with formal representations of domain knowledge to improve the performance of a search engine for online product information. Rezgui ‎(2006) similarly uses domain knowledge to formulate an ontology that informs the indexing and retrieval of construction content. These studies demonstrate how standard retrieval computations can be complemented when applied to building design and construction.

Complete Chapter List

Search this Book: