A practical driver fatigue detection algorithm based on eye state

Abstract

Fatigue driving is an important reason of traffic accidents. This paper proposes a practical algorithm to detect eye closure using an infrared camera to monitor and recognize driver dozing. Face and eye of the driver are first localized and then marked in every frame. After extracting the eye region, the proposed algorithm can detect eye corners, and with the eye corners restrictions, eyelids movement can be detected precisely and whether the eye closed or not can also be classified. An alarm is generated if the eye is closed for more than a specified period of time. The accuracy of the algorithm is demonstrated using real data under day and night with different drivers.