Akademik Veri Yönetim Sistemi
6. International Mediterranean Scientific Research Congress, Rome, İtalya, 13 - 15 Ağustos 2024, ss.145-161
Representation in architecture is a visual thinking activity that enables architects to concretize their
design processes. Architecture can express itself through representational tools and repeat the design
process many times to achieve more efficient results. While sketches, perspective drawings, and models
are traditionally accepted forms of representation, new means of representation have emerged with the
rapid development of technology. In the 21st century, the emergence of generative artificial intelligence
technology is becoming a prominent technology for architectural practice, where visual images are of
great importance in design. Generative artificial intelligence can contribute to the field by producing
and developing an architectural form that accurately conveys prompts. This study conducted research
on how generative artificial intelligence tools can be used as architectural representation tools. As a field
study, Italo Calvino's Invisible Cities was found valuable as it is a descriptive work about many different
imaginary cities. The prompts written to be entered into the artificial intelligence tool included
descriptions from the book and concepts underlined in the representations produced by previous
designers. The study realized text-to-image and image-to-image productions through Ideogram,
Leonardo, and Adobe Firefly tools. A reference was taken from a previously realized sketch
representation in the image-to-image production. The sketch chosen as a reference was preferred
because it has the potential to be completed creatively by artificial intelligence. With the research, a
discussion on how user-friendly artificial intelligence tools, which are publicly available, accessible,
and free options, can take place as an architectural representation tool is realized.
Keywords: Representation in architecture, Visualization with Artificial Intelligence, Text to image
generation, Image to image generation