This project investigates the use of computer vision machine learning (ML) techniques to create a piece of architecture on a site in a city obsessed with the human act of looking. It proposes a paediatric ophthalmology specialist research and treatment centre open to the public beside the 14th Century Torre de Niccolò, an ancient watch tower of the municipality that was previously a part of Florence’s city wall.

The development of this project is inspired by medical conditions affecting stereo vision, such as strabismus. The project proposes a methodology in which ML algorithms are trained using bespoke datasets of stereo image pairs of different subjects at different focal lengths to create hidden spaces of binocular vision. The ML technique is used to inform the design of the building at different scales, rapidly creating variations that can be selected by the designer according to different criteria including aesthetic and environmental properties.