Traffic Solution for Indus 2019 Aim: To automate the process of reporting traffic violations through the use of machine learning. The system takes in a short clip of a traffic condition at an intersection, then outputs a violation code prediction. The system has been trained to identify two types of violation: Vehicle occupying the yellow box and Vehicle blocking the pedestrian path. The system is a purely software system (i.e. do not need hardware interface to make prediction). Note: This project has been completed (Sept 2019) and is no longer active. No future improvements will be made.