Video-Based Abnormal Driving Detection

Abnormal driving includes the distractions like talking/ using mobile while driving, operating radio, eating, doing certain action like reaching back while the vehicle in motion. All these activities are related to endangerment of life and accident are prone. With the help of AWGRD network model using deep learning fusions, the state of driver at a particular frame of time can be determined and warned with a level of accuracy. There are, many models are present which are used to predict the state of driver like WGD, WGRD, AWGRD out of which AWGRD is better because it takes super positions of other layers, so the accuracy level increases. Our model predicts the behavior of driver by taking the help of ten classes which are ordered according to their state of activity.