Toward a Real-Time Automated Measure of Empathy and Dominance

This paper presents a number of algorithms and related software modules for the automated analysis in real-time of nonverbal cues related to expressive gesture in social interaction. Analysis is based on phase synchronisation (PS) and recurrence quantification analysis. We start from the hypothesis that PS is one of the low level social signals explaining empathy and dominance in a small group of users. Real-time implementation of the algorithms is available in the EyesWeb XMI social signal processing library. Applications of the approach is in user centric media and in particular in active music listening (EUICT FP7 Project SAME).