Online sinhala handwritten character recognition for desktops

Abstract

Online Sinhala handwritten character recognition is a fairly under-developed, albeit well researched area in the local natural language initiative. Recent attention towards this area produced an application named PanhinDA, a Sinhala Handwritten Character Recognition Tool for PDAs from the University of Moratuwa, Dept. of Computer Science and Engineering. However, due to the obvious lack of high processing capability in a PDA, development of this tool was stunted, using just one method of recognition. This resulted in poor recognition capabilities. Nevertheless, the development of such a tool for recognising Sinhala handwritten characters is essential, playing a vital role when the nation moves towards e-governance. Such a tool could also be used for academic purposes. This has prompted further research in the area, with the aim of incorporating multiple (alternative) techniques simultaneously to ensure higher recognition. Sinhala consists of 18 vowels, 41 consonants and 17 character associative symbols, and unlike English, poses a unique array of difficulties, since the letters are of a curved form, as opposed to straight lines.