HANDWRITTEN DIGIT RECOGNITION USING COMBINED ID3-DERIVED FUZZY RULES AND MARKOV-CHAINS

In this paper, we present an approach for handwritten digit recognitio n based on combined ID3-derived fuzzy rules and Markov chains. Both te chniques use statistical models on the structural representation of di git images. Skeleton images are adopted to produce fuzzy rules based o n the ID3 approach, and contour images are used for the Markov chain b ased approach. Decision trees produced from the ID3 algorithm are conv erted to a set of simplified rules which are then fuzzified into a set of fuzzy rules. To retain the classification performance, a two-layer perceptron is applied to optimize defuzzification parameters. On the other hand, a digit contour is traversed in a well-defined order and t he Markov chain is used to perform sequential analysis and to match wi th models. The two classifiers are then combined to complement each ot her using a three-layer perceptron. The combined classifier achieves a good classification performance and at the same time it overcomes the difficulties in a conventional syntactic approach for handwritten cha racter recognition, such as the scaling problem and lack of machine le arning ability. Experimental results on NIST Special Database 3 show t hat the combined classifier has a significantly improved performance i n terms of substitution versus rejection rates. After about 15% of dig its that cannot be classified with high confidence by the combined cla ssifier are re-classified by a nearest neighbor classifier using optim ized prototypes, the overall classification rate can be as high as 98. 6% without rejection. Copyright (C) 1996 Pattern Recognition Society.