Application of genetic algorithm for handwritten Kannada characters

Automation of character recognition system is an open research problem. Many researchers have worked on recognition of printed characters, numerals but a very little amount of work is available on handwritten character recognition. Our work is just a step towards checking the applicability of genetic algorithm for the recognition of handwritten Kannada characters. Unfortunately, results obtained are very poor. First we have used Euclidian distance method as a fitness function for our genetic algorithm. Then we have used Mahalanobis distance method as a fitness function. Results of the two are compared to check their effects on recognition rate of genetic algorithm.