30YearOldBodyGan

30YearOldBodyGan is a project where personal intimacy is translated into digital abstraction. For a month, I took daily photographs of my stomach, an intimate and personal area, to create a database for a cGAN (conditioned Generative Adversarial Network). The result is no longer a photograph but a compressed, anonymous archetype of the stomach, where perspective and identity dissolve. The core of the project is the trained model itself, which serves as a generative archive for various outcomes. This allows me to present site-specific results that respond to specific contexts. For example, the work can take the form of an endlessly growing grid of algorithmically generated body hair, a multi-meter-long installation on a building, or a series of printed editions. By conditioning the GAN with line drawings, the model remains open to dialogue and manipulation: text can be integrated into the skin, and location-specific applications are possible.

model .pkl, adjustable to situation