Systematic development of real-time driver drowsiness detection system using deep learning

Advancements in globalization have significantly seen a rise road travel. This has also led to increased car accidents and fatalities, which become global cause of concern. Driver's behavior, including drowsiness, contributes many the deaths. The main objective this study is develop system diminish mishaps caused by driver's drowsiness. Recently deep convolutional neural networks been used multiple applications, identifying anticipate driver However, limited studies investigated systematic optimization (CNNs) hyperparameters, could lead better anticipation To bridge gap, holistic approach based on learning method proposed paper drivers' drowsiness provide an alerting mechanism prevent related accidents. ensure optimal performance achievement system, database real-time images preprocessed via Haar cascade's classifiers systematically optimize CNN model's hyperparameters. Different metrics, accuracy, precision, recall, F1-score, confusion matrix, are evaluate model. training evaluation results model achieved accuracy 99.87%, while testing accurately classify drowsy with 97.98%.