A Novel Multi-feature Multi-classifier Scheme for Unconstrained Handwritten Devanagari Character Recognition

Abstract

This paper presents a novel approach for recognition of unconstrained handwritten Marathi characters. The recognition is carried out using multistage feature extraction and classification scheme. The initial stages of feature extraction are based upon the structural features and the classification of the characters is done according to their parameters. The final stage of feature extraction employs Radon transform and Euclidean distance transform and applied to two separate feed forward back propagation neural networks. The hybrid classifier at the final stage takes the input from two neural network classifiers and template matching classifier and decides the final output based on maximum voting rule. This multistage feature extraction and classification scheme improves the recognition accuracy over individual classifiers considerably. The recognition rate achieved from the proposed method is 95.40%.