Research of eye tracking method based on Kanade-Lucas optical algorithm

The sudden change of the lighting or view angle will lead to the decreasing of the eye recognition rate very much.So advanced a new algorithm,which combined the Kanade-Lucas optical algorithm (K-L) with the AdaBoost cascade classifier.This algorithm recognized and saved the Harris corner using the AdaBoost algorithm.Those saved corner features would be tracked using K-L algorithm,if the AdaBoost algorithm did not recognize them again.The method does not change the number of cascade classifier and the field of image pixels,but the recognition rate is higher,the misrecognition rate is lower,and reduces the iterate computation of identification efficiently.