Driver fatigue monitoring system based on eye state analysis

Driver fatigue has been one of the major causes of accidents all over the world. This paper presents a real-time fatigue monitoring system which exploits driver's eye to detect fatigue. The approach uses Viola-Jones Face Cascade of classifiers for the detection of Driver's Face. The eye region is estimated heuristically with respect to the width and height of the detected face. The run length of the distribution of the pixel intensities quantised into bins are used as features and are extracted from the eye region on a frame by frame basis. The feature is well able to discriminate the different states of the driver's eye like open, nearly closed and closed. A Support Vector Machine (SVM) is finally integrated within the system to classify the facial appearance as either fatigued or otherwise. The overall system achieved an accuracy of 93.5%.