Tifinagh handwritten character recognition using genetic algorithms

摘要

Handwritten character recognition system involves many different process including character images preprocessing, database preparation (features extraction), generation of best features and classification. Building best features is the complex phase during the implementation of a character recognition system. In this paper we have performed feature extraction using gradient direction technique. The novelty of our approach is to generate new features and achieve better accuracy using Genetic Algorithm which outputs new vectors based on the fitness parameter. The classification phase is performed using a feedforward neural network. The experimental results show that the performance of the Optical Character Recognition system is around 89.5%.