The Research For An Artificial Affective Modelbased On Kinetic

Affective computing currently refers to three different affective capabilities: emotion perception, emotion recognition and emotion expression. It had not been studied widely until Pro.Picard published a book named "Affective computing" at MIT lab in 1997. Generally speaking, affective computing technology is to capture facial and physiological change caused by the emotions through all kinds of sensors, and use emotion model to recognize these signals in order to understand the human feelings and react intellectually.A new model of emotional state was proposed in this paper, then introduced the technical realization in detail. Finally, the model was applied to drive a facial synthesis unit. This affective model basically contained two main sub models, one was emotional state model, the other was neural state model. The kernel thought came from a assumption that emotion is excited in a very short time by a certain stimuli, then returned to the calmness smoothly, just like that in flexible material, the process was short, strong and fluctuant in the whole time. Considering the essences of flexible material and the traits of emotion fluctuating, a second-order kinetic model was employed to describe the course of emotion. Meanwhile by the analogy between them, some kinetic conceptions were brought in the emotion model for naming and computing relative parameters. In addition, the view of energy was used to analyze the emotion accumulation and the mutual effects among the emotion states. Another model was named as neural state model which reflects the physiologically arousal level. As we know, emotion is based on physiology, affected by physiology. physiological change obeys to its owned rules, two characters mainly considered, saturation and hysteresis. On the other hand, people think there existed mood which can reflect permanence of emotion and affects the internal experience and action. Then the paper illuminated the building of expression synthesis unit and introduced the facial synthesis method. At the end, applying the model to facial synthesis unit to get a dynamic facial expression.The main creative points in this paper:1.Respectively considered the cognition and the experience process of emotion, the research was focused on the inherent, physiological, and experienced process of emotion.2.According to the affective psychological mechanism and drawing an analogy to kinetics, a engineering model was built to simulate the dynamic process of emotion. all the efforts make the emotional process dynamic expression. Various formulas were based on kinemics, control and psychology theories.3.Considered the interactions between the neural arousing and emotions.4.The input stimulus was labeled as four dimension, that is type, time, strength, which make it meaningful in psychology.