A machine emotion transfer model for intelligent human-machine interaction based on group division

Abstract

One of most important objectives of human–machine interaction is how to make the behavior of machine more like human’s, especially in the field of machine emotion expression. To achieve this goal, this paper solves the problems in the human–machine interaction in three aspects: differentiated emotion, emotion difference resolution, and covert emotion recognition. In the proposed method, the EEG (Electroencephalogram) signal is used to present the expression medium of individual emotion and personality, and a similarity calculation method is designed to measure the difference between EEG signals. In addition, group division approach is developed to achieve the differential output in emotion transfer model. Thus, the aim of this paper is to develop a personalized system of machine emotion expression, and to build the dynamic emotional interaction model between human and machine in an intelligent interactive environment. The results on the DEAP dataset verify feasibility of the group division method and the differentiated emotion transfer model.