Detection of Driver's Face Orientation for Safety Driving Assistance

Abstract

We propose face orientation estimation system to detect driver's behavior at the intersection. At first, we detected driver's face and face parts from images shooted by near infrared camera on the dashboard in the car. And then, estimated face orientation using detected face parts positions, 3D face model and particle filter. We show the algorithm of face and face parts detection by SVM(Support Vector Machine) and face parts tracking and orientation estimation by particle filter and also show the result of estimation