Combining quadratic classifier and pair discriminators by pairwise coupling for handwritten Chinese character recognition

The accuracy of handwritten Chinese character recognition can be improved by pair discrimination of similar characters. In this paper, we propose a new method for combining the baseline classifier with incomplete pair discriminators to better exploit their complementariness. The outputs of the baseline classifier and pair discriminators are transformed to two-class probabilities, which are then fused by pairwise coupling (PWC) for final decision. In our experiments using the modified quadratic discriminant function (MQDF) as baseline classifier and LDA-based pair discriminators, the PWC method outperforms the filter method. At best, the error rate of MQDF was reduced by factors over 28%.