Online Devanagari isolated character recognition for the iPhone using Hidden Markov Models

Abstract

In this paper, we present a novel scheme, which is to be implemented on the iPhone, for the recognition of online handwritten basic isolated characters of the Devanagari script. Devanagari is an Indian script that is used for several major languages such as Hindi, Sanskrit, Marathi & Nepali and is spoken as well as written by more than 500 million people. Unconstrained Devanagari writing is more complex than English cursive due to the possible variations in the order number, direction and shape of constituent strokes. The Devanagari alphabet is split into 13 vowels & 36 consonants. A manual study of various characters was done and 42 stroke classes were created. A stroke based recognition approach has been designed where strokes are recognized using Hidden Markov Models (HMM). One HMM is constructed for each stroke class. A second stage of classification has been designed and is used for recognition of characters using stroke classification results along with look up tables. The distinguishing feature of our implementation of online handwriting recognition of isolated Devanagari characters is that it is being designed and implemented for the new iPhone platform and it takes care of various constraints this platform presents us with.