OPTICAL CHARACTER RECOGNITION FOR MULTI-FONT ENGLISH LANGUAGE FROM PRINTED DOCUMENTS

Sathya Narayanan, Saranya
International Journal of Exploring Emerging Trends in Engineering (IJEETE)   unpublished
In this paper, we propose a new method for recognizing English character in different fonts. The proposed method is based on radial basis function of neural network instead of SOM neural network is resist to font variant. The proposed method can be used for both upper and lower case alphabet and also for digits. Here the size of the character is not a big problem as we are using 10x10 matrix resizing for all the characters. The correspondence measure neural network is applied to identify
more » ... to identify character but in radial basis function the extracted character are compared to database and target is defined. The accuracy of the existing method is rapidly decreases as the samples to the database but in proposed method the accuracy is almost constant and does not decreases rapidly. We achieved more than 92% of accuracy for each character that includes 20 different font.
fatcat:zomaeg4eb5fztaaojw2rhihnvu