Driver Drowsiness Detection System using DLib

Fatigue and Drowsy state are the common reasons for road accidents. Every year road accidents occur due to driver's sleepiness and fatigue increases. In this proposed paper, an application for the driver drowsiness detection system is introduced which will be helpful to prevent road mishaps that were mainly caused by driver's fatigue, and hence it will help to improve traffic safety. This detection system deals with Deep learning algorithms as well as the DLib library to detect and predict the results. Also, an alert system is added to this application to alert the person if he/she is feeling drowsy. We have used a highly enhanced face tracking algorithm which can make the prediction more accurate. It uses 68 face landmarks to detect facial features. Detecting the facial feature using 68 landmarks is more accurate and precise than the PERCLOS and Haar cascading algorithm.