Shape Classification of Building Information Models using Neural Networks
I. Evangelou, N. Vitsas, G. Papaioannou, M. Georgioudakis, A. Chatzisymeon (2021). Shape Classification of Building Information Models using Neural Networks, Eurographics Workshop on 3D Object Retrieval (3DOR).
Abstract: The Building InformationModelling (BIM) procedure introduces specifications and data exchange formatswidely used by the construction industry to describe functional and geometricelements of building structures in the design, planning, cost estimation andconstruction phases of large civil engineering projects. In this paper weexplain how to apply a modern, low-parameter, neural-network-basedclassification solution to the automatic geometric BIM element labeling, whichis becoming an increasingly important task in software solutions for theconstruction industry. The network is designed so that it extracts featuresregarding general shape, scale and aspect ratio of each BIM element and beextremely fast during training and prediction. We evaluate our networkarchitecture on a real BIM dataset and showcase accuracy that is difficult toachieve with a generic 3D shape classification network.