Character Feature Extraction Method based on Integrated Neural Network

The research of pattern recognition shows: For a complex identification and classification system, only one method is always difficult to identify and separate objects well, while among the different classification methods, the high complementary exist, the integration of multiple classifiers can clearly improve the recognition rate. Therefore, integration methods of multiple classifiers have become a hot spot which the researchers are focusing on. This paper sets up an integrated neural network handwriting recognition system, comprehensively use a variety of pattern recognition, fully reflect various aspects features of handwritten digital characters, this method has improved the accuracy of handwritten numeral recognition system.