AI-Driven Explorations: Parametric Topologies

This project investigates the use of diffusion-based generative AI for architectural design, focusing on the creation of fluid, parametric topologies. By leveraging advanced noise seeding techniques and reference-driven initialization, the process achieves a consistent morphological language while still embracing the natural variation of generative methods. The resulting images articulate a balance between organic fluidity and structured perforations, offering an experimental perspective on spatial organization and surface articulation.

For the video sequences, careful prompt engineering was employed to control camera movement and reduce abrupt morphing—common artifacts in generative animation pipelines. While some irregularities remain, the approach demonstrates a methodical integration of emerging AI technologies into architectural form exploration.