A fast HMM algorithm for on-line handwritten character recognition

A fast HMM algorithm is proposed for on-line hand written character recognition. After preprocessing input strokes are discretized so that a discrete HMM can be used. This particular discretization naturally leads to a simple procedure for assigning initial state and state transition probabilities. In the training phase, complete marginalization with respect to state is not performed (constrained Viterbi). A simple smoothing/flooring procedure yields fast and robust learning. A criterion based on the normalized maximum likelihood ratio is given for deciding when to create a new model for the same character in the learning phase, in order to cope with stroke order variations and large shape variations. Preliminary experiments are done on the new Kuchibue database from the Tokyo University of Agriculture and Technology. The results seem to be encouraging.