Drowsiness Detection System: A Comparative Study Based on SVM and AdaBoost Classifiers

Abstract

Road traffic accidents have been a matter of interest for many researchers, due to its irreversible results. Driver drowsiness is one of the major causes of these accidents, thus endangering road safety. For a decade, teams have gathered in order to have a better understanding of the reasons behind these crashes. Driver Behavior Analysis connected technologies and the adaptation of big data have been used in this matter. Actually, Driver Behavior represents an emerging important topic, it examines different conducts of the driver during his trajectory. In this article, we will recall and discuss the solutions proposed by researchers to try to solve and better control and ameliorate road safety, and the role of smart cities and the integration of Artificial Intelligence in facilitating this analyze, then we will present the design and evaluation of a driver drowsiness detection system that employs computer vision and machine learning techniques. The system analyzes facial images so as to classify the state of the driver as “Drowsy” or “Non-drowsy” and provides immediate visual and audio alerts that could prevent potential accidents. This system was trained and tested on a publicly available dataset, comparing Support Vector Machine (SVM) and AdaBoost classifiers, hence emphasizing the capability of AI based solutions in controlling driver behavior and enhancing road safety.