The Use of Convolutional Neural Networks for the Monitoring of Driver Drowsiness

Abstract

Automated self-driving automobiles, etc., are a result of computer vision advancements that aid drivers. About 20%
of accidents occur because drivers are too sleepy or exhausted to react properly. A number of solutions have been
suggested in response to this major issue. On the other hand, real-time processing is not something they excel at.
These approaches suffer from a lack of robustness when it comes to dealing with lighting circumstances and
variations in human faces. Our goal is to install a smart processing system that will significantly cut down on traffic
accidents. The method allows us to detect the driver's facial features, such as the frequency of blinking, the angle of
the eyes relative to the lips, the amount of yawning, the amount of head movement, and the proportion of closed
eyes. A camera is used in this system to continually observe the driver. When a motorist looks into the camera, haar
cascade classifiers can identify their face and eyes. To determine whether the eyes are closed or open, we use a
custom-designed convolutional neural network to categorize photos of the eyes. It is the categorization that
determines the ocular closure score. There will be an alert that goes off if the driver is determined to be too sleepy.