A computational framework for aesthetic preferences in architecture using computer vision and artificial neural networks

Research output: Contribution to journalArticleResearchpeer review

Authors

External Research Organisations

  • Universidade Presbiteriana Mackenzie
View graph of relations

Details

Original languageEnglish
JournalInternational Journal of Architectural Computing
Publication statusE-pub ahead of print - 10 Sept 2024

Abstract

This paper introduces a computational aesthetics framework utilizing computer vision (CV) and artificial neural networks (ANN) to predict the aesthetic preferences of groups of people for architecture. It relies on part-to-whole theories from aesthetics and cognitive psychology. A survey of a group of people on preferences of images is held to record an average hedonic response (AHR). CV algorithms MSER and SAM recognize parts in images. Birkhoff’s aesthetic measure formula is adapted by employing the number of parts and their connections. These quantities are used as input layers of an ANN, and the AHR is the target output. The ANN evaluates images to output a predicted hedonic response (PHR), which is tested as a criterion in parametric design space navigation and in mapping the latent space of GANs. We conclude that such a framework is a heuristic method for better understanding the design and latent spaces and exploring designs.

Keywords

    Artificial neural networks, computational aesthetics, computer vision, design space navigation, empirical aesthetics, hedonic response, heuristics, latent space map, quantitative aesthetics

ASJC Scopus subject areas

Cite this

A computational framework for aesthetic preferences in architecture using computer vision and artificial neural networks. / Sardenberg, Victor; Guatelli, Igor; Becker, Mirco.
In: International Journal of Architectural Computing, 10.09.2024.

Research output: Contribution to journalArticleResearchpeer review

Sardenberg, V., Guatelli, I., & Becker, M. (2024). A computational framework for aesthetic preferences in architecture using computer vision and artificial neural networks. International Journal of Architectural Computing. Advance online publication. https://doi.org/10.1177/14780771241279350
Sardenberg V, Guatelli I, Becker M. A computational framework for aesthetic preferences in architecture using computer vision and artificial neural networks. International Journal of Architectural Computing. 2024 Sept 10. Epub 2024 Sept 10. doi: 10.1177/14780771241279350
Sardenberg, Victor ; Guatelli, Igor ; Becker, Mirco. / A computational framework for aesthetic preferences in architecture using computer vision and artificial neural networks. In: International Journal of Architectural Computing. 2024.
Download
@article{91e2c9c1a5254e55b31c5344e3fd78d6,
title = "A computational framework for aesthetic preferences in architecture using computer vision and artificial neural networks",
abstract = "This paper introduces a computational aesthetics framework utilizing computer vision (CV) and artificial neural networks (ANN) to predict the aesthetic preferences of groups of people for architecture. It relies on part-to-whole theories from aesthetics and cognitive psychology. A survey of a group of people on preferences of images is held to record an average hedonic response (AHR). CV algorithms MSER and SAM recognize parts in images. Birkhoff{\textquoteright}s aesthetic measure formula is adapted by employing the number of parts and their connections. These quantities are used as input layers of an ANN, and the AHR is the target output. The ANN evaluates images to output a predicted hedonic response (PHR), which is tested as a criterion in parametric design space navigation and in mapping the latent space of GANs. We conclude that such a framework is a heuristic method for better understanding the design and latent spaces and exploring designs.",
keywords = "Artificial neural networks, computational aesthetics, computer vision, design space navigation, empirical aesthetics, hedonic response, heuristics, latent space map, quantitative aesthetics",
author = "Victor Sardenberg and Igor Guatelli and Mirco Becker",
note = "Publisher Copyright: {\textcopyright} The Author(s) 2024.",
year = "2024",
month = sep,
day = "10",
doi = "10.1177/14780771241279350",
language = "English",

}

Download

TY - JOUR

T1 - A computational framework for aesthetic preferences in architecture using computer vision and artificial neural networks

AU - Sardenberg, Victor

AU - Guatelli, Igor

AU - Becker, Mirco

N1 - Publisher Copyright: © The Author(s) 2024.

PY - 2024/9/10

Y1 - 2024/9/10

N2 - This paper introduces a computational aesthetics framework utilizing computer vision (CV) and artificial neural networks (ANN) to predict the aesthetic preferences of groups of people for architecture. It relies on part-to-whole theories from aesthetics and cognitive psychology. A survey of a group of people on preferences of images is held to record an average hedonic response (AHR). CV algorithms MSER and SAM recognize parts in images. Birkhoff’s aesthetic measure formula is adapted by employing the number of parts and their connections. These quantities are used as input layers of an ANN, and the AHR is the target output. The ANN evaluates images to output a predicted hedonic response (PHR), which is tested as a criterion in parametric design space navigation and in mapping the latent space of GANs. We conclude that such a framework is a heuristic method for better understanding the design and latent spaces and exploring designs.

AB - This paper introduces a computational aesthetics framework utilizing computer vision (CV) and artificial neural networks (ANN) to predict the aesthetic preferences of groups of people for architecture. It relies on part-to-whole theories from aesthetics and cognitive psychology. A survey of a group of people on preferences of images is held to record an average hedonic response (AHR). CV algorithms MSER and SAM recognize parts in images. Birkhoff’s aesthetic measure formula is adapted by employing the number of parts and their connections. These quantities are used as input layers of an ANN, and the AHR is the target output. The ANN evaluates images to output a predicted hedonic response (PHR), which is tested as a criterion in parametric design space navigation and in mapping the latent space of GANs. We conclude that such a framework is a heuristic method for better understanding the design and latent spaces and exploring designs.

KW - Artificial neural networks

KW - computational aesthetics

KW - computer vision

KW - design space navigation

KW - empirical aesthetics

KW - hedonic response

KW - heuristics

KW - latent space map

KW - quantitative aesthetics

UR - http://www.scopus.com/inward/record.url?scp=85203713801&partnerID=8YFLogxK

U2 - 10.1177/14780771241279350

DO - 10.1177/14780771241279350

M3 - Article

AN - SCOPUS:85203713801

JO - International Journal of Architectural Computing

JF - International Journal of Architectural Computing

SN - 1478-0771

ER -

By the same author(s)