Enhanced Driver’s Drowsiness Detection System using CNN model

Many accidents are frequently caused because of driver’s drowsiness. While driving at night, people may fall asleep unknowingly due to fatigue, making it possible for the driver to lose control of the vehicle at times. The system then uses machine learning methods and methodologies to identify the indicators of the driver’s tiredness. Even while some older algorithms and techniques can detect a person’s drowsiness, our system uses Deep Learning algorithms and techniques to do so with more accuracy and across a wider range of environmental situations. There are many benefits by using Convolutional Neural Network (CNN) for sleepiness detection, including real-time processing, tolerance to changes in lighting, and the capacity to capture intricate spatial characteristics. When compared to earlier image processing methods, utilizing CNN as a detection algorithm for processing the video is the greatest approach in achieving reliable results. In this model, Haar cascade classifier is used for facial features extraction, Shi-Tomasi corner detection and Lucas-Kanade optical flow algorithm are used to detect the head movements of the driver. The model gave good results with 97% percent accuracy.