Driver Drowsiness Detection System Using Deep Neural Network

The detection of drowsy drivers has been the approach for avoiding road accidents in recent years. The purpose of this study is to establish a method for building intelligent vehicles that can reduce drowsy driving impairment. Therefore, it is vital to develop a reliable warning system to prevent the incident's cause. In this proposed work, we provide a drowsy driver alert system developed using a way in which the video stream process (VSP) is evaluated by an inborn reflex construct applying an eye aspect ratio (EAR) and geometrical distance of the attention. An algorithmic technique based on facial landmarks is also applied for reliable eye recognition. When driver drowsiness is detected, the module delivers a warning message at the side of the crash impact and site information, notifying the voice surveillance system.