Detecting Drivers' Drowsiness using Haar Cascade Classifier

Abstract

Every year the number of fatal injuries and deaths is increasing worldwide. Drowsiness and driver's fatigue are two of the most common causes of car accidents. The driver sleepiness detection system is essential to avoid accidents. In this work we propose an approach to identify drivers' drowsiness using facial and eye expressions. Machine learning and deep learning techniques have been utilized in this work to anticipate a driver's emotion, which would improve road safety. The authors have used the Haar-based cascade classifier to detect eye movement and OpenCV to identify the driver's state effectively. In addition to detecting driver drowsiness, our proposed system can work in adverse conditions, such as varying lighting conditions, use of spectacles, and the presence of a beard on the person's face. Our work is able to classify drowsiness in such extreme conditions as well accurately. Keywords- Drowsiness Detection, CNN, Haar Cascade