Isolated Hand Printed Character Recognition for Malayalam Using Kohonen Neural Network

The research in the field of Document analysis and Character recognition has been intensive for the last few decades. In this paper an attempt is made to recognize isolated hand printed Malayalam characters. The normalized and thinned character image is used for feature extraction. For recognition, a Kohonen neural network technique is adopted. The proposed approach is tested and implemented on Isolated Hand Printed Malayalam characters and an overall recognition rate of 90 - 95% is achieved.