Kannada Confusing Character Recognition and Classification Using Random Forest and SVM

Abstract

Dravidian scripts have a lot of confusing characters because of the complexity of the characters and curve nature so recognizing those confusing characters are a tedious process. Kannada have many confusing characters which cause difficulties in extraction from a kannada document. The proposed work deals with recognition and classification of confusing characters. In method uses Random Forest and SVM as the classifiers to classify the confusing characters. The proposed system achieved a classifier accuracy of 78%. Finally the system will recognize the confusing character using template matching and feature value outcome based out of the classifiers. In proposed work deals with ten classes that are used to classify and recognize the confusing characters.