Real-time stealth intervention for motor learning using player flow-state

Abstract

We present a novel approach to real-time adaptation in serious games for at-home motor learning. Our approach assesses and responds to the “flow-state” of players by tracking and classifying facial emotions in real-time using the Kinect camera. Three different approaches for stealth assessment and adaptation using performance and flow-state data are defined, along with a case-study evaluation of these approaches based on their effectiveness at maintaining positive affective interaction in a subject.