Developed by Ali Özgün Akyüz as part of Sónar+D, this exhibition presents a real-time generative installation that explores the space between embodied presence and computational perception. Visitors encounter a system that captures the human body through a camera, segments it with a neural network, and transforms movement into abstract visual traces via generative AI (StreamDiffusion). Each frame, the body is detected, decomposed into masks and vectors, and reconstructed as evolving color trails on a digital canvas.

As attendees move through the space, their gestures leave ephemeral yet persistent marks — the body, translated into code, does not disappear but is reconstituted through GPU circuits, memory registers, and projection. The algorithmic traces are not static residues; each frame recalculates segmentation thresholds, mask coefficients, and fade parameters, continuously defining what becomes visible and how.

The installation operates as an entangled system: camera, neural network, projection surface, body, and environment all interact without a single element dominating, following Deleuze’s notion of assemblage. The work occupies the threshold between organic movement and computational logic, where space emerges through their interaction rather than existing beforehand — a living, generative environment shaped by human presence and machine perception.