Online Handwriting Recognition for Tamil

Abstract

A system for online recognition of handwritten Tamil characters is presented. A handwritten character is constructed by executing a sequence of strokes. A structure-or shape-based representation of a stroke is used in which a stroke is represented as a string of shape features. Using this string representation, an unknown stroke is identified by comparing it with a database of strokes using a flexible string matching procedure. A full character is recognized by identifying all the component strokes. Character termination, is determined using a finite state automaton. Development of similar systems for other Indian scripts is outlined.

Design of a Template for Handwriting Based Hindi Text Entry in Handheld Devices

“…Handwriting based text entry solution for mobile devices, has been extensively researched in the recent times, some of the areas of focus being recognition algorithms [6] and interface issues [4] [6].…”

“…In the field of text input in Indian Languages, Aparna, K. et al [4] presents a method for online recognition mechanism in the case of hand-written Tamil text. It describes character recognition as recognizing the constituent strokes and matching them against a pre-defined database.…”

“…Character recognition is done in the following steps [4]: 1) Stroke identification-where the input stroke is compared with the database of existing defined strokes 2) Character recognition-grouping already identified stroke labels and converting them to suitable character codes. For Indian alphabet, the ISCII (Indian Script Code for Information Interchange) is used.…”

“…The ideal input conditions for high recognition accuracy include [4]: 1. Consistent size of the hand written strokes constituting the letters 2.…”

“…Our literature review shows that recognition accuracy for hand-written text improves considerably if the size of the letterform (glyph) remains consistent and the orientation, vertical [4]. In order to constrain the hand-written input text in this manner, we propose a writing template (Fig.3).…”