For this project, students explored the intersection of artificial intelligence and architectural visualization by leveraging neural networks to generate, manipulate, and refine architectural renderings. The exercise began with a simple massing model in Rhino, serving as a foundational geometry for further AI-assisted transformations.

Using tools such as Stable Diffusion, LoRA models, ControlNet, and Pix2Pix, I experimented with image synthesis techniques to produce detailed renders, ranging from exterior facade studies to immersive interior perspectives. The goal was not only to achieve photorealistic results but also to critically examine the nature of "realism" in architectural representation.