SUST-DDD: A Real-Drive Dataset for Driver Drowsiness Detection

Driver drowsiness is one of the most important factors in traffic accidents. For this reason, systems should be developed to detect drowsiness early and to warn the driver by examining the driver or driving situations. This kind of systems play an important role to prevent traffic accidents. Three techniques are used to detect drowsiness: (1) based on vehicle parameters, (2) based on physiological parameters and (3) based on behavioral parameters. In this study, a new dataset for drowsiness has been created and some kind of deep learning methods such as AlexNet, LSTM, VGG16, VGG19, VGGFaceNet and hybrid deep networks have been applied on this dataset to predict drowsiness of the drivers. The experimental results show that the created dataset and implemented hybrid deep networks are successful to predict drowsiness with more than 90,53% for accuracy, 91,74% for precision, 91,28% for recall and 91,46% for f1score.