inSensed: Wearable Collection of Multimedia based on Interest
Manuel Blum, Alex Pentland, Zurich Switzerland, Cambridge Ma · 2005
We present a wearable data collection system that allows users to collect their experiences into a continually growing and adapting multimedia diary. The system, called iSensed, uses the patterns in sensor readings from a camera, microphone, and accelerometers, to classify the user's activities and automatically collect multimedia clips when the user is in an `interesting’ situation. `Interest’ is estimated by a user-trained rule relating the sequence of user activities/situations to their concept of `interest’. This allows the multimedia information to be structured in a manner similar to the user’s episodic memory, e.g., unusual and potentially important events are recorded chronologically, allowing rapid browsing and automatic structuring of the multimedia diary. The system is based on commercially available mobile devices, either a PDA or a camera phones, together with two small wireless accelerometers worn on the wrist and belt.