Eye Detection and Recognition in the Fatigue Warning System

For a driver fatigue warning system, one of the most important problems to solve is eye detection and recognition. The blink information was used to preliminary eye location. Modified Susan operator of eye inner corner extraction to accurate eye location. According to the characteristics of the external environment are less prone to interference from eye inner corner, small template of the eye inner corner was extracted for eye tracking. In the methods of eyes state recognition, the correlation between an online generated template of open eyes in the process of eye tracking and the eye region extracted from the image was used. According to the size of correlation, to further determine the eyes state . The experiment result shows that data quantity disposed is little, realized speed is fast and the accuracy rate of the eye recognition is increased in this method.