(c) Larry Ewing, Simon Budig, Garrett LeSage
1994 .

Department of Computer Science

PetrSU | Software projects | AMICT | Staff | News archive | Contact | Search

Typewritten symbols recognition using Genetic Programming

Prof. Igor L. Brathikov, Alexey A. Popov(Saint-Petersburg state university, St.Peterburg, Russia)

Genetic programming is a new technique in Artificial Intelligence based on the evolutionary algorithms and inspired by biological evolution. As a matter of fact, genetic programming is a special case of genetic algorithms, where each individual is a computer program. Therefore, this technique could be used to optimize a population of computer programs to solve the problem.

This report demonstrates how genetic programming can be used to solve the problem of optical character recognition, specifically typewritten symbols. At present, there are many approaches to solve this problem, but all of them have their own limitations.

A main purpose was to determine and estimate the different applications of genetic programming: independent (single) use of genetic programming or the usage of this methodology in common with the other techniques. The approach given in this report could be successful at learning, maintaining and upgrading rules for typewritten symbols recognition, particularly in disputable situations. Specific fitness functions, terminals and functions that satisfy the requirements of a main problem were considered.

This research presents the successful application of genetic programming to a difficult and topical task.