Real Time Driver Drowsiness Detecion using Transfer learning

Abstract

According to statistics, drowsy driving is the leading cause of accidents worldwide that result in the loss of precious lives and worsen public health. When a driver is fatigued, cameras can be employed to detect their drowsiness and inform them well before which can help in decreasing accidents. This work employes a transfer Learning model DenseNet to identify the driver drowsiness in real time. The MRL eye dataset of 84923 images has been used and the model works well with 91.56% accuracy.