generative installation, 2026 / custom application, live self-portrait from a private photo archive, dimensions variable
Obsession is a self-portrait generated live, assembled from tens of thousands of photographs stored in my private digital archive. A custom application continuously processes the collection and builds my current likeness out of its fragments. The portrait never settles: every new photograph taken on any device connected to my network immediately changes its structure. In the gallery, visitors can photograph themselves on a shared device; their image enters the archive and the portrait rebuilds itself in front of them, while a counter ticks up by one.
The work concerns the moment when a private archive stops being a store of memory and starts to behave like a system that produces identity: in the cloud, in automated processing, in the everyday gestures of saving and looking. The archive is not mine alone, it holds the faces of the people closest to me, so the portrait is always partly made of others. Obsession asks whether a portrait can still describe a person when it is built from the remnants of his own images.
The work uses algorithmic image processing, not generative AI.