Vision-based estimation of driving inattention fused via Dempster-Shafer theory

Driving inattention is one of the most serious traffic accidents cause whereas a monitoring system of the driver's vigilance could be used as a prevention. In this paper, we propose an intelligent system that estimates the driver fatigue using combination of lane tracking and face detection fused via Dempster-Shafer theory. The approach is aimed to estimate the fatigue based on the expert knowledge for driving vigilance. The method can compensate for the impact of poor detection and improve the accuracy of fatigue alerting.