Fatigue state detection from multi-feature of eyes

In recent years, the problem of fatigue driving has attracted more and more attention. Methods of fatigue detection are no longer just by the way of questionnaires. This paper presents an algorithm of fatigue state detection from multi-feature of eyes, which can determine whether a driver is in a state of fatigue by analyzing the behavior of the driver's eyes. Firstly, image preprocessing, face detection based on AdaBoost and active shape model of human eye positioning will be done in sequence to improve the accuracy of identification. After that, according to the P80 standard, the eyes state of opening and closing is identified by the area of eyes, which can obtain the percentage of eye closure time, the average time of eye closure and the frequency of blinking as eye multi-feature parameters. Finally, the support vector machine of fatigue state detection model is built with these three types of eye multi-feature parameters, and then the driver's driving state can be determined. Experiment results show the efficiency of the proposed method.