Throughout two separate courses, the use of AI in architecture was researched, explored, and put to the test. The first experiment was research, taking black and white images taken of the Bechert residence by architect Alfred Preis and recoloring them using AI tools and comparative analysis. The second experiment was creating a database of precast concrete facades, and training a GAN (generative adversarial network) to generate images of new precast facades. The third and final experiment was to create an actual precast facade using a public domain text-to-image database (Midjourney) and illustrate it as a digital model, while simultaneously creating and studying database of all precast buildings in North America.
Ex. 01: Recolored images of the Bechert Residence by Alfred Preis
Ex. 02: Original compiled database of precast buildings for use by StyleGAN
Resultant images generated by the StyleGAN, ran over numerous iterations.
Ex. 03A: Precast case study, American Concrete Institute by Minoru Yamasaki in Farmington Hills, Michigan
Ex. 03B: Text-to-image tool used with verbiage extracted from precast concrete research
Ex. 03C: Using depth mapping to transform AI imagery into digital components, translated into a precast facade
