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.

The paper describes a research experiment of incorporating quantitative aesthetic evaluation and feeding the metric back into a parametric model to steer the search within the design space for a high-ranking design solution. The experiment is part of a longer-standing interest and research in quantitative aesthetics. A web platform inspired by dating apps was developed to retrieve an aesthetic score of images (drawings and photographs of architectural projects). The app and scoring system was tested for functionality against an existing dataset of aesthetic measure (triangles, polygon nets). In the actual experiment, an evolutionary algorithm generated images of design candidates (phenotypes) and used the aesthetic score retrieved by the ``crowd'' of app users as a fitness function for the next generation/population. The research is in the tradition of empirical aesthetics of G. T. Fechner (Fechner, 1876), using a web app to crowdsource aesthetic scores and using these to evolve design candidates. The paper describes how the system is set up and presents its results in four distinct exercises.

This paper presents a comparison of different workflows for mobile robotic fabrication using modular building blocks. Different localization, locomotion, and interlocking building systems strategies are tested and compared. The work is influenced by related research into ecosystems of building parts, design software, and builder robots to digitize the construction work. For localization, it compares LIDARs, reacTIVision, and ArUco markers. As a mobile platform, a MIR100 robot platform, a 3.3 m linear axis, and a manual trolly are used. Interlocking components such as wood slates, custom-made bricks, and interlocking wood building blocks are used. The research is in the field of collective robotic construction (CRC) using bespoke robots designed in tandem with specific discrete building blocks.

A method for building slabs composed of discrete short timber slats is proposed. It involves simulating its principal stresses to locate and orient each building component. Those are assembled by a robotic arm in collaboration with a worker using an Augmented Reality head-mounted device. The prototype serves as a proof of concept and proved structurally viable, and future work involves interlocking the slats using halved joints and implementing a strategy to reuse wood production waste components as stock material.

A computational framework is introduced to map aesthetic categories, highlighting distinctions between human-created competition entries and AI-generated designs. Assessment of architectural aesthetics has traditionally been subjective; however, computational methods now enable the systematic analysis and visualization of design variations. Using 14 aesthetic parameters, the framework applies Principal Component Analysis (PCA) to create visual maps of aesthetic relationships. Two hypotheses are discussed: (1) There is a hegemony of the aesthetic category of the beautiful in mainstream architectural competitions, and (2) that generative models like Stable Diffusion can produce visuals belonging to other aesthetic categories. The computational aesthetic framework is applied to verify both. The first hypothesis is disproven by revealing that human-created designs demonstrate significant aesthetic variation and unpredictability, while the second is confirmed by demonstrating the capacity to produce images in 16 other categories beyond the beautiful. The maps of visual relationships group images by aesthetic categories, encouraging designers to explore and enhance underpopulated areas. The computational aesthetics framework is used to analytically and quantitatively support arguments regarding architectural aesthetics.

This paper discusses Augmented Reality on smartphones as a tool for building complex structures without utilizing CNC technology while showcasing architectural ...