Expression Intensity Measurement Based On Fuzzy

The facial expression is one of the most important tools in our communication, as the development of the affective computing and intelligent man-machine, we hope the computer can perceive and understanding of our human’s expression just like ourselves. So the first step is to analyze the people’s expression. As an important issue, there are many scholars have been spent lots of time in the study of facial expression analysis and many expression recognition methods have been raised, while the study of the expression measurement are few. As we know our expression is formed by the facial muscles, we should study and track the muscle movements firstly if we want to study the facial expression. Our work in this paper is based on Ekman’s FACS system, by the MHI theory and fuzzy model, the intensity of the facial expressions is classified. The facial expression is usually influenced by the race, gender, age and other factors. By analyzing the intensity of the expression we have achieved some research results.In this paper, we have studied the status of the facial expression recognition, and have an in-depth study for the methods of the expression intensity. Finally and from the video to identify face facial expression. The main works is as follows:Extract the expression features by the MHI motion template.Recognize the expression by studying the facial muscles and BP neural network.Use the fuzzy to measure the intensity of the expression. Introduce the expression intensity level, extract the expression intensity values by BP neural network .Then take the path of the feature points and the corresponding frame of the intensity level as the samples of the fuzzy. Set up the intensity model and classify the expression intensity.From the video to identify face facial expression, and try to classify the expression of“happy”. The experiment results have shown that the facial expression intensity system is valid.