RECOGNITION OF ONLINE CURSIVE KOREAN CHARACTERS COMBINING STATISTICALAND STRUCTURAL METHODS

In this paper, we propose a hybrid recognition method, and show its us efulness in recognizing online cursive Korean characters. A finite sta te network is constructed to represent the rules of character composit ion from graphemes. In the network, each are and node expands into sta tistical and structural recognizers, respectively. The statistical rec ognizer produces intermediate recognition results in traditional hidde n Markov modeling, then the structural recognizer analyzes them. The r esults from two recognizers are combined in a probabilistic framework complementing the Markov assumption of hidden Markov modeling. The exp erimental results showed significant performance improvements in error reduction and computation time as compared to the statistical approac h alone. (C) 1997 Pattern Recognition Society. Published by Elsevier S cience Ltd.