Driver Gaze Tracking And Eyes Off The Road Detection System

Abstract

Existing System: Driver monitoring has been a long standing research problem in computer vision and automotive driver monitoring system applications. Broadly speaking, there are two approaches to estimate gaze direction: one is Techniques that only use the head pose and the second is those that use the driver’s head pose and gaze. For systems that rely only on head pose estimation the processing complexicity increases in regard to the software requirement and for those that relay on head pose and gaze estimation are grouped into hardware and software based approaches Active Appearance Models (AAMs) for facial feature tracking and head pose estimation and 3D eye-model was used for accurate gaze estimation, both of theserelay on near-infrared (IR) illuminators to generate the bright pupil effect. This system is not person-independent and must be calibrated for every system configuration and driver. Besides their complexicity of integration and implementation these systems are too expensive to be installed in mass-produced cars and they require periodic re-calibration because vibrations cause the system calibration to drift over time.