Generative early architectural visualizations: incorporating architect’s style-trained models

• DOI: 10.1093/jcde/qwae065
• Publication Date: 2024-07-15T14:58:52Z
• Authors: Jin-Kook Lee; Youngjin Yoo; Seung Hyun Cha
• Keywords: Architectural style; Generative Design

Abstract This study introduces a novel approach to architectural visualization using generative artificial intelligence (AI), particularly emphasizing text-to-image technology, remarkably improve the process right from initial design phase within architecture, engineering, and construction industry. By creating more than 10 000 images incorporating an architect’s personal style characteristics into residential house model, effectiveness of base AI models. Furthermore, various styles were integrated enhance process. method involved additional training for with low similarity rates, which required extensive data preparation their integration model. Demonstrated be effective across multiple scenarios, this technique markedly enhances efficiency speed production images. Highlighting vast potential in visualization, our emphasizes technology’s shift toward facilitating user-centered personalized applications.