Online Character Recognition

Character Recognition has become an intensive research areas during the last few decades because of its potential applications. However, most existing classifiers used in recognizing handwritten online characters suffer from poor feature selection and slow convergence which affect training time and recognition accuracy. This paper proposed a methodology that is based on extraction of structural features (invariant moment, stroke number and projection) and a statistical feature (zoning) from the characters. A genetic algorithm was modified through its fitness function and genetic operators to minimize the character recognition errors. The Modified Genetic Algorithm (MGA) was used to select optimized feature subset of the character to reduce the number of insignificant and redundant features. A genetic based neural network model was developed by integrating the MGA into an existing Modified Optical Backpropagation (MOBP) learning algorithm to train the network. Three classifiers (C1, C2 and C3) were then formulated from MGA-MOBP such that C1 classified without using MGA at classification level, C2 classified using MGA at classification level while C3 employed MGA at feature selection level and classified at classification level The developed C3 achieves a better performance of recognition accuracy and recognition time