Eye State Recognition Algorithm GAHMM of Web-based Learning Fatigue

With the increasingly deeper application of multimedia and network technology in education, the proportion of web-based learning presents a trend of continuous growth in people’s daily study, and the webbased learners tend to have learning fatigue during their learning process of distance education. Under the 4 learning states of normal study, fatigue and doubt, the eye opening degrees of web-based learner have certain difference. Through preprocessing of the color images of web-based learners’ eye state, the 2-dimensional kernel function is selected to build 48 optimal filters and obtain 48 eigenvalues, and then, through HMM, training is conducted to identify the eye state and feature classification. The experiment results show that this algorithm has very high identification ability of the web-based learning fatigue.