Image Classification using Machine Learning

Problem Description:
Develop predictive models that can determine, given an image, which one of 11 classes it is.
Traffic congestion seems to be at an all-time high. Machine Learning methods can be developed to help solve traffic problems. In this program, features extracted from traffic images depicting different objects are analysed to determine their type as one of 11 classes, noted by integers 1-11: car, suv, small_truck, medium_truck, large_truck, pedestrian, bus, van, people, bicycle, and motorcycle. The object classes are heavily imbalanced. For example, the training data contains 10375 cars but only 3 bicycles and 0 people.