Between Images and Built Form: Automating the Recognition of Standardised Building Components Using Deep Learning

Abstract

Building on the richness of recent contributions in the field, this paper presents a state-of-the-art CNN analysis method for automating the recognition of standardised building components in modern heritage buildings. At the turn of the twentieth century manufactured building components became widely advertised for specification by architects. Consequently, a form of standardisation across various typologies began to take place. During this era of rapid economic and industrialised growth, many forms of public building were erected. This paper seeks to demonstrate a method for informing the recognition of such elements using deep learning to recognise 'families' of elements across a range of buildings in order to retrieve and recognise their technical specifications from the contemporary trade literature. The method is illustrated through the case of Carnegie Public Libraries in the UK, which provides a unique but ubiquitous platform from which to explore the potential for the automated recognition of manufactured standard architectural components. The aim of enhancing this knowledge base is to use the degree to which these were standardised originally as a means to inform and so support their ongoing care but also that of many other contemporary buildings. Although these libraries are numerous, they are maintained at a local level and as such, their shared challenges for maintenance remain unknown to one another. Additionally, this paper presents a methodology to indirectly retrieve useful indicators and semantics, relating to emerging HBIM families, by applying deep learning to a varied range of architectural imagery.

Networking Historic Environmental Standards to Address Modern Challenges for Sustainable Conservation in HBIM

“…Previous work has confirmed this hypothesis through a comparative test of the performance (measured class-by-class using the relative F1-scores) of 3 different machine learning architectures in the classification of different ventilation building components [70]. The results indicated that a Feature Pyramid Network classification model (FPN, Figure 3) outperforms a high-end Traditional Machine Learning model (TML) composed of SIFT, K-means clustering and standard RF, by an average 26% on all the 4 examined classes of objects.…”

Conservation Digitalization through Heritage Building Information Modelling (HBIM): Online Database Theoretical Review

“…Deep learning allows machines to learn from pixels to classifiers. Each layered layer with a similar structure performing different transformation functions pulls features from layers below and above in a directly connected way [33]. In Digital Cultural Heritage (DCH), semantic segmentation of 3D Point Clouds with Deep Learning (DL) techniques can help recognise historical architectural elements at an adequate level of detail, speeding up the process of modelling historical buildings for developing BIM models from survey data, called HBIM (Historical Building Information Modeling) [28].…”

Shape Grammar Libraries of European Classical Architectural Elements for Historic Bim

“…In addition, the incorporation of regulations, such as the system of rates for buildings in London which determined eighteenth century window sizes, wall thicknesses and storey heights for the purposes of taxation and fire protection, may also enhance the capacity of heritage-related parametric library objects. As might the deployment of deep learning in relation to the information contained within historic building catalogues (Prizeman 2016;2019, Pezzica et al 2019. This latter notion has previously been explored in the development of visual matching techniques for early 20 th century Carnegie Library buildings (Prizeman et al 2018).…”