Combining Discrete Orthogonal Moments and DHMMs for Off-Line Handwritten Chinese Character Recognition

Discrete orthogonal moment set is one of the novel feature moment-based descriptors for image analysis. The Tchebichef moments and Krawtchouk moments are the two representatives in this class. This paper studies the performance of the two discrete orthogonal moments in the recognition of off-line handwritten Chinese amount in words under discrete-time hidden Markov models (DHMMs) framework. The lower order moments are employed as features. A serial of experiments are carried out to compare their performance with that of the continuous orthogonal movements such as Zernike and Legendre. Experimental results suggest that the recognition performance of two discrete orthogonal moments is higher than that of the continuous discrete moments. In additional, different values of the number of zones, observation symbols and states are also used to find the better model structure for the new approach