This work presents a real-time clustering technique that takes advantage of the unique properties of event-based vision sensors. Our approach redefines the well-known mean-shift clustering method using asynchronous events instead of conventional frames.