Restricted accessResearch articleFirst published online 2022-6
Machine learning and complex compositional principles in architecture: Application of convolutional neural networks for generation of context-dependent spatial compositions
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.
ChingFDK. Principles. Architecture–Form, Space and Order. 3rd ed. New Jersey, NY: John Willey & Sons, Inc.; 2007, pp. 338–402.
2.
TamkeMNicholasPZwierzyckiM. Machine learning for architectural design: practices and infrastructure. Int J Architectural Comput2018; 16(2): 123–143.
3.
FukushimaK. Neocognitron: a self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position. Biol Cybernetics1980; 36: 193–202.
4.
LeCunYBoserBDenkerJS, et al.Back propagation applied to handwritten zip code recognition, Neural Comput1989; 1(4): 541–551
5.
KrizhevskyASutskeverIHintonGE. ImageNet Classification with Deep Convolutional Neural Networks. Adv Neural Inf Process Syst2012; 25(2): 84–90.
6.
SchmittG. Information architecture: Basics of CAAD and its future. 1st ed. Basel, Switzerland: Birkhäuser, 1999.
7.
SaggioA. The IT revolution in architecture. Thoughts on a paradigm shift. 5th ed. Lulu.com, 2013.
8.
SłykJ. The origin of information architecture. Warsaw, Poland: Oficyna Wydawnicza Politechniki Warszawskiej, 2012.
9.
GeroJS. Artificial intelligence in computer-aided design: Progress and prognosis. Computer-Aided Des1996; 28(3): 153–154.
10.
RameshAPavlovMGohG, et al.Zero-shot text-to-image generation. arXiv Preprint. arXiv:2102.12092 2021
11.
ChenMRadfordAChildR, et al.Generative pretraining from pixels. Proc 37th Int Conf Machine Learn2020; 119: 1691–1703.
12.
MordvintsevAOlahCTykaM. Inceptionism: Going deeper into neural networks. Google Research Blog, 2015.
13.
GatysALEckerSABethgeM. A neural algorithm of artistic style. J Vis2016; 16(326).
14.
SilvestreJIkedaYGuenaF. Artificial imagination of architecture with deep convolutional neural network. CAADRIA2016: 881–890.
15.
SzegedyCLiuWJiaY, et al.Going deeper with convolutions. IEEE CVPR2015: 1–9.
16.
DengJDongWSocherR, et al.Image net: A large-scale hierarchical image database. IEEE Conference Computer Vision and Pattern RecognitionJune, 2009, Miami, Florida: 248–255.
17.
TakizawaAFurutaA. 3D spatial analysis method with first-person viewpoint by deep convolutional neural network with omnidirectional RGB and depth Images. Conf Proc 35th eCAADe Conf2017; 2: 693–702.
18.
GoodfellowIJPouget-AbadieJMirzaM, et al.Generative adversarial networks. arXiv Preprint. arXiv 2014.
19.
MostafaAAl-KaseemB. Fill in the blanks: Deep convolutional generative adversarial networks to investigate the virtual design space of historical islamic patterns. 9th ASCAAD Conf Proc2021: 614–621.
20.
NgJMYKheanNMaddenD. Optimising image classification: Implementation of convolutional neural network algorithms to distinguish between plans and sections within the Architectural, Engineering and Construction (AEC) industry. In: Intelligent & Informed - 24th CAADRIA Conference, April, 2019, Victoria University of Wellington, pp. 795–804.
21.
LongJShelhamerEDarrellT. Fully convolutional networks for semantic segmentation. IEEE CVPR2015: 3431–3440.
22.
XiaoYChenSIkedaY, et al.Automatic recognition and segmentation of architectural elements from 2D drawings by convolutional neural network. Proc 25th CAADRIA Conf2020; 1: 843–852.
23.
EisenstadtVLangenhanCAlthoffKD. Generation of floor plan variations with convolutional neural networks and case-based reasoning-An approach for transformative adaptation of room configurations within a framework for support of early conceptual design phases. Proc 37th eCAADe 23rd SIGraDi ConferenceDecember 1, 2019; 2: 79–84.
24.
PengWZhangFNagakuraT. Machines’ perception of space: Employing 3D isovist methods and a convolutional neural network in architectural space classification. ACADIA2017: 474–481.
25.
BenediktM. To take hold of space: Isovists and isovist fields. Environ Plann B Plann Des1979; 6(1): 47–65.
26.
HintonGERumelhartDEWilliamsRJ. Learning representations by back-propagating errors. Nature1986; 323(9): 533–536.
27.
HuberPJ. Robust estimation of location parameter. Ann Stat1964; 53(1): 73–101.
28.
StrivastavaNHintonGKrizhevskyA, et al.Dropout: A simple way to prevent neural networks from overfitting. J Machine Learn Res2014; 15: 1929–1958.
29.
DosovitskiyABeyerLKolesnikovA, et al.An image is worth 16x16 Words: Transformers for image recognition at scale. arXiv Preprint. arXiv 2020.
30.
CaronMTouvronHMisraI, et al.Emerging properties in self-supervised vision transformers. arXiv Preprint. arXiv 2021.
