Designing a Driver Face Monitoring System for Driver Drowsiness and Distraction Detection

Annually, many vehicle crashes happen in the world because of driver drowsiness (fatigue) or distraction (inattention). These crashes lead to many fatalities and losses. Driver face monitoring is one of the main approaches for driver drowsiness/distraction detection. A system based on this approach gets an image sequence from camera and process it to extract drowsiness and distraction symptoms from driver's face. In this article, a driver face monitoring system is designed that estimates driver hypo-vigilance based on extracting symptoms from eye regions. In proposed system, two features are extracted: (1) percentage of eye closure (PERCLOS) and (2) eye blink rate. Experimental results on some sequences that captured in both real conditions (in vehicle) and laboratory conditions show that proposed system is accurate in feature extraction and hypo-vigilance estimation.