Driver Monitoring System using an Embedded Computer Platform

In this paper, the driver monitoring system (DMS), which monitors the inattention and drowsiness of the driver, is proposed. For the face detection and localization of the 468 3D facial points, the MediaPipe FaceMesh algorithm is used. The area over the distance (ADR) is used to detect the eye state, while the mouth aspect ratio (MAR) is used to detect the mouth state, and Euler’s angles are used for head position estimation. The developed DMS performance is verified using the YawDD dataset videos on a personal computer (PC) and validated in the Raspberry Pi 3.