Automatic driver distraction detection using deep convolutional neural networks

Recently, the number of road accidents has been increased worldwide due to distraction drivers. This rapid crush often leads injuries, loss properties, even deaths people. Therefore, it is essential monitor and analyze driver's behavior during driving time detect mitigate accident. To various kinds like- using cell phone, talking others, eating, sleeping or lack concentration driving; machine learning/deep learning can play significant role. However, this process may need high computational capacity train model by huge training dataset. In paper, we made an effort develop CNN based method distracted driver identify cause distractions like talking, eating means face hand localization. Four architectures namely CNN, VGG-16, ResNet50 MobileNetV2 have adopted for transfer learning. verify effectiveness, proposed trained with thousands images from a publicly available dataset containing ten different postures conditions analyzed results performance metrics. The showed that pre-trained best classification efficiency.