The paper is devoted to solving actual scientific and technical problem of increasing interactivity of user’s interaction with sensor human-computer systems.
A new emotion state model of user of human-computer systems with natural user interface Kinect is proposed in this paper. The emotion assessmentis done in two phases. In the first phase, the data from the SDK Kinect is interpreted as main features of user’s emotion –head pose angle, expression of face and position of twenty joints of the user's body. As for the second phase, the selected features of emotion are determined as fuzzy variables in accordance of cognitive psychology. Problems of features’uncertainty are solved with the help of developed discrete interval type-2 fuzzy logic system.
The proposed model allows us to organize effective feedback, timely and appropriate response of human-computer system to changesof emotional state of user.
The main areas of use of the model areindustry of videogamesand entertainments, robotic systems, operator stations, social networks, etc.