Interactive AI Poetry: Poetic Visual Worlds in Real Time
All body movements, facial expressions and gestures are filmed by a camera, digitized via motion capture and fed into a local image AI.
The system is programmed to create the illusion of continuous, cinematic movement with 10 generated images per second. During the session, a creative dialogue takes place between the person in front of the camera and the AV artist controlling the real-time system – ping-pong with zeros and ones. The prompts for image style and aesthetics are formulated and mixed by the AV artist during the performance and the relationship between the parameters is constantly weighted.
Each generated frame is unique and cannot be repeated; the system delivers deliberately provoked unpredictability in the visual output.
