A Deep Learning Based Approach for Real-time Driver Drowsiness Detection

Proper detection of driver drowsiness plays a key role to reduce the accidental rate in our country. Drowsiness occurs due to sleeping during driving. A Convolutional Neural Network (CNN) model is suggested in this paper to identify the driver’s tiredness in real-time. The CenterFace algorithm is used for face detection in this paper along with Haar-Classifier that is used for eye feature extraction. The suggested model has a high accuracy rate and can detect whether the driver is drowsy or aware in atypical conditions where the driver wears a glass or mask. The training accuracy is 97% and validation accuracy is 98%. The test accuracy for the non-glassed and glassed eye datasets is 97% and 92%. Moreover, our proposed method performs better in a real-time environment.