Drivers Drowsiness Detection Using Machine Learning (KEY IJP************858)

• Gattu Srinath

Abstract

This project involves detecting and alerting the driver co-passengers if the driver is drowsy. 41% of road accidents in India are caused because the driver feels drowsy when heshe is driving the vehicle, to avoid this problem our system takes the drivers image or video (frame by frame) as input and using the appropriate model classifies the drivers eye position as sleepy or not sleepy, if the system detects that the driver is drowsy an alarm is set off notifying the driver the co-passengers. The system records the videos and detects the drivers face in every frame by employing image processing techniques. The system can detect facial landmarks, calculates Eye Aspect Ratio (EAR) and Eye Closure Ratio (ECR) to detect drivers drowsiness based on adaptive thresholding. The proposed approach focuses on building a drowsiness detection mechanism to alert the driver to avoid the catastrophe. In this work, the detection system can identify whether the drivers eyes were closed or open even in low light or dim light and how much time the eyes were in closed state. Based on the time the system will generate an alert