Eye-Gaze and Augmented Reality Framework for Driver Assistance
Muhammad Shahid, Tabassam Nawaz, Hafiz Adnan Habib · 2013
Driver inattention or cognitive overload is among the leading contributor to road accidents. Driver behavioral cues could be employed in advanced driver assistance systems (ADASs) to alleviate such accidents. Facial and eye gaze are among the most import behavioral cues to reflect driver active state within the vehicle. With the intent to improve the human error related accidental controls in mind, a simple and smart ADAS is proposed that could assist the driver based on continues monitoring of facial and eye gaze information. The system measures driver eye gaze within near frontal facial positions and project future position of vehicle on windscreen based on the vehicle parameters and driver's active eye gaze estimates, assuming the movement of vehicle in a straight line. The projections adapt to shift in driver's perspective. The projection over the wind screen give visualization as if the lines are physically drawn over road according to the width of the vehicle. The system also warns the driver when there is constant shift in head or eye gaze from normal forward facing positions, beyond some threshold period. The system is based on live video input from low-end webcam. The system will reinforce driver ability in effective estimation of future positions of vehicle and leverage better control while driving. In addition it can also be employed to assist novice drivers to keep track of the width of vehicle during the training sessions.