Sleepiness Detection System Based on Facial Expressions

Drowsy driving will lead to a serious accident because the operation of the car becomes impossible. In order to prevent such accidents, Development of a drowsy driving prevention system is necessary. Therefore, this research uses eye movements that are correlated with high sleepiness and mouth movements that are said to be correlated with waking effort. In addition, the authors use Web camera and Dlib to measure these parameters in real time. Furthermore, in order to measure blinks, the authors implemented a blink measurement method corresponding to noise caused by body movement. In addition, the authors have proposed a sleepiness detection method using the acquired data. The control engineering method was used in this method. As a result, the accuracies are 90% for the blink measurement method and 91.7% for the drowsiness detection method.