Abstract
This paper explains the sensing method we propose, which allows high-speed and accurate detection of driver’s eye region required for preventive safety technologies for automobiles. This evolutionary eye sensing method deals with problems such as eyeglass frame or reflection of lens surface, a difference in day and night outside lights, and change of lighting environment in tunnels and shadow of buildings, with the information about the surrounding region of eye. To support brightness variations, a genetic algorithm with the objective function that changed according to brightness of target images was adopted. Robustness and effectiveness of the proposed method is verified by experiments.
Keywords: Eye tracking, Evolutionary Video Processing, Driver Monitoring