Monitoring driver awareness through eye tracking

Abstract

Driving is a task that requires a high level of focus for the security of not just the person driving but also other drivers and passengers out on the road. Hence, it is a difficult task, as a human being might not be able to maintain this focus continuously for a lot of reasons such as drowsiness. Drowsiness is a reason for a lot of accidents that occur around the world, as a result car manufacturers and researchers are looking for solutions to decrease these accidents as much as possible. One key way of decreasing these types of accidents is by monitoring a driver’s eye to track whether a driver is drowsy or not. If a person is drowsy or is falling asleep, he/she will be alerted by means of sound. To solve this we approached this problem by setting up a camera on the driver’s dashboard and then conduct image processing by using OpenCV library to identify whether a driver is fully focused or not and whether it is safe enough for a driver to continue driving. Fortunately, a pre-trained Viola-Jones classifier comes out-of-the-box with OpenCV library. Where, The ViolaJones algorithm is a system for object recognition that enables real-time detection of image features. By going through four different phases this algorithm detects the face and tracks the eye really conveniently with an efficient accuracy rate. In ViolaJones algorithm two pre-processing techniques, along with the three functions, were evaluated and compared.