Matthew S. Ryan and Graham R. Nudd, Dynamic Character Recognition Using Hidden Markov Models (May 1, 1993).
Hidden Markov Model theory is an extension of the Markov Model process. It has found uses in such areas as speech recognition, target tracking and word recognition. One area which has received little in the way of research interest, is the use of Hidden Markov Models in character recognition. In this paper the application of Hidden Markov Model theory to dynamic character recognition is investigated. The basic Hidden Markov Model theory is reviewed, and so are the algorithms associated with it. A quick overview of the dynamic character recognition process is considered. Then three types of describing characters are considered, position, inclination angle of small vectors and stroke directional encoding using a Freeman code. A system using each of these descriptions, using Hidden Markov Models in the comparison stage, is described. It is recognised that experiments using the different encoding systems have to be carried out to check the validity of this chosen method.