A Generic Structure of Object Classification Using Genetic Programming

In This paper a method for classification of two types of objects using genetic programming (GP) has been presented. These two objects are coins of different sizes, and different textures. The basic algorithm of genetic programming was presented and explained. The features used for training and testing are mean, standard deviation, skewness and kurtosis. Precision and recall were used as performance measures and they were the main building blocks in building the fitness function. They replaced the false alarm and detection rate that was used in previous works. The result figures as well as values of precision, recall, fitness values, time elapsed, and number of generations used in training was presented. The very basic structure of a GP system was implemented and proved that it can work well as a standalone computational algorithm.