Online Handwritten Gurmukhi Character Recognition Using Elastic Matching

Abstract

This paper presents implementation of elastic matching technique to recognize online handwritten Gurmukhi characters. We have discussed a process that recognizes characters in two stages. First stage recognizes the strokes, in second stage, character is evaluated on the basis of recognized strokes. Feature are computed to strengthen recognition results. Also, we have discussed a simple way to store data for handwritten strokes and characters. The database for strokes stores script number, stroke number and stroke sample number for every point of a stroke. For 60 writer's and a set of 41 Gurmukhi characters, we have obtained recognition rate as 90.08%.