A printed Chinese character recognition method

Chinese character recognition is integrated technology that is related to pattern recognition, artificial intelligence, fuzzy mathematics, information theory, computer science and so on. Currently a lot of researches focused on the Chinese character recognition have been done, however, the results are still not very satisfactory. In this paper, A new recognition method is put forward based on the previous researches in this field and the combination of the statistical classification and neural networks. Using neural network to implement vector conversion and thus achieve the recognition of text, the method not only avoids the interference characteristics of Chinese structures, but also much improved the Chinese character recognition rate. Experiments with a large number of training samples show that the method of Chinese character recognition rate with the proposed method is more than 93%.