Driver Monitoring System Based on Gaze Classification by Using Eye Center Location Recognition
Juyong Lee, Seungdo Jeong, Jihoon Lee · International Journal of Control and Automation · 2018
Due to the rapid development of IT technology, a variety of technologies are being developed to improve the safety of drivers by using various sensors and cameras.Recently, various studies have been conducted to improve the safety of the drivers by using a relatively low-cost camera instead of the expensive sensors.However, the problem is that the existing camera system cannot accurately predict the driver's sight due to various constraints such as camera angle and illuminance.Therefore, in this paper, we propose an algorithm that can improve the safety of the driver by using predicted gaze and drowsiness prediction using eye center information.The proposed algorithm consists of two steps that predict the gazing region by recognizing the center of the eye and a stage that recognize the drowsiness based on the closing time of the eyes.Through performance analysis, we confirmed that the proposed algorithm accurately predicts the sight of the driver from various angles.In addition, we set up scenarios for driving and conducted additional performance analysis, which confirmed that it could help improving driver safety.