COOPERATIVE CLASSIFIERS FOR HIGH-QUALITY HANDPRINTED CHARACTER-RECOGNITION

In this paper, a high quality handprinted character recognition system is presented. Four classifiers based on simple features work in parallel and their co-operation is used for quality improvement. The four classifiers are based on two different normalization sequences, two different feature extraction methods and two different classification techniques. The results of the classifiers are combined using a multilayer perceptron as a supervisor, which extracts the overall information contained in the output of the classifiers. The results obtained on the NIST Test Data 1 are reported using the uppercase letters in the NIST Special Database 3 as a training set; the error rate is 3.68% when no rejection is allowed.