Online Handwriting Recognition using HMM

Abstract

Basically handwriting recognition can be divided into two parts as Offline handwriting recognition and Online handwriting recognition. Highly accurate output with predefined constraints can be given by Online handwriting recognition system as it is related to size of vocabulary and writer dependency, printed writing style etc. Hidden markov model increases the success rate of online recognition system. Online handwriting recognition gives additional time information which is not present in Offline system. A Markov process is a random prediction process whose future behavior rely only on its present state, does not depend on the past state. Which means it should satisfy the Markov condition. A Hidden markov model (HMM) is a statistical markov model. In HMM model the system being modeled is assumed to be a markov process with hidden states. Hidden Markov models (HMMs) can be viewed as extensions of discrete-state Markov processes. Human-machine interaction can be drastically getting improved as On-line handwriting recognition technology contains that capability. As instead of using keyboard any person can write anything by hand with the help of digital pen or any similar equipment would be more natural. HMM build a effective mathematical models for characterizing the variance both in time and signal space presented in speech signal.