Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Artificiale Rilievo GAN-Generated Architectural Sculptural Relief
This paper describes “Artificiale Rilievo”, the first work of architectural sculptural relief produced by a generative adversarial network (GAN). Technically, the authors present novel methods developed for the generation of three-dimensional sculptural designs using a pseudo-3d description of form based on vector displacement maps (VDMs). Our approach improves on existing methods by expanding the range of possible forms, and suggests broad application in ornamental architectural design. Conceptually, the artistic work described here brings tiling geometries found in contemporary architectural ornament into dialog with forms drawn from the Western architectural canon, and reflects on the dataset as a retrograde influence in the otherwise avant-garde field of creative AI. In contrast with other AI-driven tools that center efficiency at the expense of expressiveness and authorial jurisdiction, the methods described here stand as an alternative approach to the application of machine learning in architectural design. Negotiating the “uncanny” boundary of individually-recognizable forms within a differentiated field, the piece materializes an animated walk through the latent space of a GAN in the solidity of cast bronze.
Artificiale Rilievo GAN-Generated Architectural Sculptural Relief
This paper describes “Artificiale Rilievo”, the first work of architectural sculptural relief produced by a generative adversarial network (GAN). Technically, the authors present novel methods developed for the generation of three-dimensional sculptural designs using a pseudo-3d description of form based on vector displacement maps (VDMs). Our approach improves on existing methods by expanding the range of possible forms, and suggests broad application in ornamental architectural design. Conceptually, the artistic work described here brings tiling geometries found in contemporary architectural ornament into dialog with forms drawn from the Western architectural canon, and reflects on the dataset as a retrograde influence in the otherwise avant-garde field of creative AI. In contrast with other AI-driven tools that center efficiency at the expense of expressiveness and authorial jurisdiction, the methods described here stand as an alternative approach to the application of machine learning in architectural design. Negotiating the “uncanny” boundary of individually-recognizable forms within a differentiated field, the piece materializes an animated walk through the latent space of a GAN in the solidity of cast bronze.
Artificiale Rilievo GAN-Generated Architectural Sculptural Relief
Gengnagel, Christoph (Herausgeber:in) / Baverel, Olivier (Herausgeber:in) / Betti, Giovanni (Herausgeber:in) / Popescu, Mariana (Herausgeber:in) / Thomsen, Mette Ramsgaard (Herausgeber:in) / Wurm, Jan (Herausgeber:in) / Steinfeld, Kyle (Autor:in) / Tebbecke, Titus (Autor:in) / Grigoriadis, Georgieos (Autor:in) / Zhou, David (Autor:in)
Design Modelling Symposium Berlin ; 2022 ; Berlin, Germany
Towards Radical Regeneration ; Kapitel: 12 ; 133-148
18.09.2022
16 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
| British Library Online Contents | 2007
British Library Online Contents | 2000
| Online Contents | 1996
The Mausoleum at Halicarnassus: sculptural decoration and architectural background
| British Library Conference Proceedings | 1997