Publication

Abstract : Driving a car is a complex task, and it requireres complete attention of the driver. Distracted driving is an activity that takes away the driver's focus and attention from the road. Worldwide number of road accidents are being recorded in the last few years. This paper attempts to establish an alert system which warns the driver in case of distracted activity. This project presents a CNN based model that gives the feedback of the driver to the owner of the car by sending a text message. The model has been trained with different classes of distraction that has been observed among the drivers. Proposed model consists of vgg16 which is fine tuned pre trained model to perform the classification with the help of an Adam Optimizer. Twillio SMS API is used for sending SMS to the owner of the vehicle or corresponding authority. This model achieved an accuracy of 93.75%.