Multi-lingual character recognition using artificial neural networks

Abstract

Character recognition is one famous problem that artificial neural networks have been applied to deal with. But, recognizing characters from multiple languages is relatively new and needs to be investigated in depth. This paper describes one such implementation. An artificial neural network was designed to recognize a bilingual character set of Tamil and English, even in a noisy environment. The system was designed, implemented, trained and tested and was found to exhibit an accuracy of 93% on noisy characters.