A Novel Tri-Stage Recognition Scheme for Handwritten Malayalam Character Recognition

Abstract

The paper proposes an efficient and novel approach of Malayalam character recognition which acquires higher recognition rate for entire character set of Malayalam script. The proposed scheme follows a three stage feature extraction. In the first stage we are grouping characters into different classes based on the number of corners, bifurcations, loops and endings. In the second phase we are identifying exact character in the class based on the different feature extraction technique specially defined for each class. In the third stage we are checking the probability of occurrence of the current character in the given position based on defined rules for the formation of words.