Snake Extraction Based On Skin Color And Improved Expression Of E-learning Feature

The research in this paper is funded by the National Natural Science Foundation Project "The study of emotion and cognition analysis which is based on academic expression in E-learning system"(No.60970052),and Beijing Natural Science Foundation Project "The key technology study of personalized E-learning community education is based on emotion and psychology "(No.4112014).The main purpose of this paper is to extract the expression characteristics parameters of learners in E-learning situations, and provide the basis for subsequent modeling and analysis of academic emotions, in order to establish the E-learning teaching system of the emotional cognitive interactive. Namely, in the E-learning teaching environment, firstly, through the camera to capture images of the learners and to detect the position of the face, then to locate the expression feature parts such as eye and mouth, ultimately, to extract the parameters of the expression characteristics-eye spacing, pupil spacing and mouth radians.The main research work and results of this paper are as follow:(1)According to the characteristics of application environment in E-learning system, and taking the lower detection rate of traditional color detection method into account, this paper, this paper improve skin color detection method by changing the sizes of the major and minor axes of the ellipse color model. The experiment shows that the improvement algorithm is more in line with the requirements of the E-learning system, and gets better detection rate for different lighting changes.(2) Based on the principle of psychological aversion degree, in this paper, the pupil spacing is selected as one of the parameters of the expression characteristics, in order to study the aversion degree of emotional modeling. The study found that when learners interested in learning content, body will lean forward and the pupil spacing will become larger; on the contrary, body will throw back and the pupil spacing become smaller, when learners are not interested in learning content. The experiment shows that, after the introduction of the pupil spacing, can effectively reduce the calculation error of the aversion degree.(3)Considering the condition that the positioning accuracy of expression feature directly do have an effect on the characteristic parameter extraction, this paper improves the method of Snake by changing the initial curve and external energy. The improved method is more accurate in positioning the contours of the eyes and the mouth, hence, there is a high accuracy rate in whi ch extracted from eye spacing, pupil spacing and mouth radian.(4) Combining with the theory and technology in these fields such as emotional calculation, image processing, artificial psychology and computer science, this paper designed and developed the E-learning facial expression feature extraction system. Experiments showed that the characteristic parameters of learners expression, the system extracted, can provide the basis for modeling and analysis of academic emotions.