A Zoning based Feature Extraction method for Recognition of Handwritten Assamese Characters

This paper introduces a novel feature extraction approach for handwritten Assamese character recognition. The performance of an optical character recognition system highly depends on the extracted feature set. Hence, feature extraction plays a significant role in achieving high recognition accuracy. Also, not all the features of an image are useful for classification and therefore feature extraction helps in yielding only the significant features for feeding into a classifier. We have tested the proposed method on offline handwritten Assamese characters and have obtained appropriate result.