Learning to Draw, Again translates hand drawing into architectural form through machine vision, point clouds, and physical modelling. The project asks how drawing changes when it is made for a visual system trained on photographs, where line, tone, density, erasure, and blank paper are read as spatial evidence.

Depth-estimation software converts marks on paper into unstable digital terrains. Through repeated drawing, scanning, projection, extraction, and printing, these misreadings are tested until they become a usable design process. Drawing becomes a negotiation between hand, paper, software, and spatial output.

The resulting point clouds are treated as sites to be occupied, cut, projected through, and translated into architectural fragments. Developed across scales, from landscape to handrail, the paper surface begins to acquire thickness, programme, structure, and use.

Situated within Bologna’s industrial and ecological edge, the project proposes an architecture formed through translation. It asks how a drawing must be made to become legible as space, and what forms emerge when the architect learns to draw according to the limits, habits, and errors of machine vision.