EXPLORING FEATURE EXTRACTION TECHNIQUES FOR CHARACTER RECOGNITION IN HANDWRITTEN MODI SCRIPT DOCUMENTS

Abstract

From the 19th century Devangari scripting is used for documentation in Maharashtra. Before Devanagari, Modi script were used mostly for documentation purpose, but due to cursive writing style, complex nature, and lack of awareness it diminishes over time. Modi script has historical roots and large number of manuscripts available in various Museums,National libraries and at cultural heritages. With time the knowledge within these documents remains unrevealed, due to lack of proper sources. The Government takes steps to preserve these documents and uncover the information within them. Now a days, emerging technologies pave the way to restore cultural heritage by using HOCR.In this study, we focus on recognizing script data from image documents, particularly handwritten Modi script. This paper explores various feature extraction techniques that facilitate character recognition from these historical handwritten documents.