| Distance education is a new form of evolution in the field of contemporary education,and this form of education has made a flourish driven by the Internet education platform.The human eye localization algorithm in this paper provides the basis of human eye location for monitoring module in distance education system.The main purpose is to determine the state of study according to the time of positioning the human eye in a certain period of time.In this paper,ASEF eye location algorithm,LBP integral projection eye location algorithm and Adaboost eye location algorithm,are mainly discussed and simulated by MATLAB.the main work is as follows:(1)Through labeling human eye position of a large number of pictures manually,ASEF algorithm makes an average value of all filters corresponding to all samples,which is named as average synthetic exact filter,and then,take human eye positioning operation.The advantage of the ASEF algorithm is that the speed of operation is fast,but the training workload is too heavy,which brings more trouble to update the sample database.(2)LBP integral projection method is to use LBP operator to extract features,and then to calculate horizontal and vertical integral projections to determine the location of the human eye.LBP method easily influenced by environment may led to the inaccurate localization.(3)The Adaboost algorithm uses the Haar-like feature to calculate the eigenvalue and generate the weak classifier.And the weak classifier is trained into a strong classifier,then to cascaded.Finally,it locates the human eyes by constructing two-layer classifiers.The Adaboost algorithm can effectively improve the speed and robustness of locating human eyes.(4)According to the comparison of above three methods,the best-result eye-localization Adaboost algorithm is realized on OpenCV platform using C language in MITCBCL database and self-build database,the experimental results verifies the effectiveness and real-time performance of this algorithm in this paper. |
PDF Full Text Request