Deep learning based driver distraction: A review of the literature

Detecting driver inattentive is one of interested research area to reduce vehicle accidents. Driver safety can be improved by detecting driver drowsiness and distraction. The increase of in-vehicle information systems induces the biomechanical and cognitive of driver distraction which affects driving performance qualitatively. The impact of digital technology on driving performance is presented in this paper based on the relative influence of the state-of-the-art techniques. The paper presents the finding of Convolution Neural Networks (CNNs) literature review covering the driver distraction detection systems. The results observe that convolutional neural network (CNN) is a powerful deep learning method that fits the needs of researchers in detecting of driver distraction. It can extract the useful attributes from trained data and train the weights by feeding the data on each level then tune the CNN based on specific driver distraction measurements. Also, it can provide a high accuracy results in real-time environment.