Pattern recognition using genetic algorithm

Genetic algorithms have been proved to be quite effective in solving certain optimization and artificial intelligence (AI) problems. They have been used in many application areas, including pattern recognition. However, the applications of genetic algorithms in pattern recognition have concentrated primarily on training neural networks for pattern recognition (Montana 1989, Whitley 1992, Kitano 1994). The research in this paper is aimed at using a genetic algorithm to perform pattern matching directly. The basic idea is to use a genetic algorithm to find the best match between nodes of the two patterns. An objective function can be defined in terms of the total difference in the magnitudes of angles between the corresponding edges of the two patterns. Experiments designed to evaluate the algorithm have shown very promising results with high accuracy in classifying the input patterns