Driver Fatigue Detection Using OpenCV and Dlib Library

Driver fatigue and distraction are major contributors to fatal road accidents worldwide. To enhance vehicle safety, this article proposes a fatigue detection system using EAR and MAR algorithms to identify signs of fatigue, such as eye closure and yawning. When prolonged eye closure or yawning is detected, the system issues an alert to the driver. It leverages OpenCV and Dlib libraries for implementation. Additionally, the system detects driver distraction by monitoring head swivels, counting left and right movements. It's worth noting that Dlib may yield facial landmark deviations for non-frontal faces. To mitigate this, the study selected 69 videos from the YawDD database, focusing on near-frontal face views, achieving a 91.3% accuracy rate.