A substantial part of architectural and urban design involves processing of compositional interdependencies and contexts. This article attempts to isolate the problem of spatial composition from the broader category of synthetic image processing. The capacity of deep convolutional neural networks for recognition and utilization of complex compositional principles has been demonstrated and evaluated under three scenarios varying in scope and approach. The proposed method reaches 95.1%–98.5% efficiency in the generation of context-fitting spatial composition. The technique can be applied for the extraction of compositional principles from the architectural, urban, or artistic contexts and may facilitate the design-related decision making by complementing the required expert analysis.

No abstract

Data s craping” i s a t erm usually used in Web browsing to refer to the automated process of data extraction from websites or interfaces designed for human use. Currently, nearly two thirds of Net traffic are generated by bots rather than humans. Similarly, Deep Convolutional Neural Networks (CNNs) can be used as artificial agents scraping cities for relevant contexts. The convolutional filters, which distinguish CNNs from the Fully-connected Neural Networks (FNNs), make them very promising candidates for feature detection in the abundant and easily accessible smart-city data consisting of GIS and BIM models, as well as satellite imagery and sensory outputs. These new, convolutional city users could roam the abstract, digitized spaces of our cities to provide insight into the architectural and urban contexts relevant to design and management processes. This article presents the results of a query of the state-of-the-art applications of Convolutional Neural Networks as architectural “city scrapers” and proposes a new, experimental framework for utilization of CNNs in context scraping in urban scale.