Affective Computing Modeling System Based On Facial Expression

Affective computing is an important branch of artificial intelligence. The agent lack of emotion is incomplete. Studying Emotion is trying to understand the essence of things, or even to contribute to the exploration and study of the causes and the development of the intelligent.Based on pervasive computing and proactive computing with psychology, cognitive science, anthropology, physiology, mathematics, computer science and others affective meta-modeling is proposed. It is a general theory for the emotions describing common characteristics in the process of modeling. Emotional signals are received through the sensory systems; recognition is judged by cognitive system and thus repository is established; emotional expression is given through experience system and expression intent is given through emotional arbitration mechanism; finally feedback system adjusts the mode. Affective personalized model is built on the basis of meta-modeling according to differences among tests'expressions which can study through collaborative mechanism system.Based on multiple statistics, emotional recognition arithmetic supported by facial expression is given. The arithmetic is the core in meta-modeling that uses principal component analysis to reduce the dimension of signals, sets up repository for the classification by means of dynamic clustering and gives the result through Fisher discriminate. On the basis of meta-modeling and emotion recognition arithmetic affective recognition system is established through camera to capture facial expression and build personalized repository and recognize emotions.Personalized models are based on meta-modeling; recognition arithmetic is based on facial expression through multiple statistics; affective recognition system is established. Based on the experiment's data the system gives higher accuracy of pleased or displeased emotional state and achieves satisfactory results.