Published Paper Details:

This research paper presents the "User-Driven Generative AI Architectural Design" project, a pioneering endeavor aimed at reshaping architectural design methodologies that makes use of artificial intelligence (AI), notably Text-to-Image Generative Adversarial Networks (T2GAN), Stable Diffusion techniques, and Google Gemini for caption production. By combining user-driven inputs with AI-generated layouts, the "User-Driven Generative AI Architectural Design" initiative seeks to democratise architectural design. The collection of curated architectural layouts from real estate companies, architects, and the internet is used as the training dataset for the model. To increase diversity, rule-based techniques are added to the collection. This extensive dataset makes it easier to create a strong generative model that can generate architectural designs that are visually appealing. Through the integration of spoken language and the production of visual materials, this innovative method signifies a revolutionary development in the domains of architecture and interior design. The framework develops to create customised and contextually relevant architectural compositions through iterative refinement and user feedback mechanisms, representing a major step towards a more imaginative and user-centered architectural design approach.