Traffic Violation Invigilation Using Transfer Learning

The modern era of evolution and urbanization in this populated world makes the abundance of vehicles on the roads. The plenty number of road accidents are increased as a consequence of violation of traffic rules. Even though many techniques are applied to invigilate the traffic violation, still it is difficult to supervise the traffic violation. An automated system for traffic invigilation is required to handle and track the people who are infringing the rules. By using Convolutional Neural Network (CNN), the right model is built to predict the class of violation. In this work, the classification of the vehicle, detection of helmet violation, wearing mask is also done using CNN model. Transfer Learning is applied here to use the learned weightage from the existing model to improve the detection.