Ml-Based Speculative Queries of Anthropometric Spatial Transformations: The Case of Mythoskeleton

Over the past decade, advancements in artificial intelligence models like Generative Adversarial Networks (GAN) have accelerated exploration of spatial transformations in architecture, design, art, and science. The triad of human, body, and space is prominent with speculative approaches imagining anatomical changes through AI. This study examines body perception through anthropometric analysis of a Machine Learning (ML) model to be trained. Inspired by E.B. Hudspeth's sci-fi anatomy book The Resurrectionist: The Lost Work of Dr. Spencer Black, the Mythoskeleton case study explores the question, 'What if mythological beasts were evolutionary ancestors of humans? How would architectural space transform?' The study involves (1) gathering skeleton images of mythological beasts, (2) training the StyleGAN2 model, (3) generating predicted skeleton models, and (4) analyzing spatial transformations through projection mapping. Findings show anthropometric differences between human and predicted skeletons could lead to organic and dynamic spatial transformations, significantly impacting spatial design in the digital era.