１引言人工神经网络对字符的学习和识别，如果按通常方法，直接把字符的位置数据输灵敏度神经网络，由于字符书写大小、位置有差异，在输入网络前必须进行前处理，即对字符正规化．这个过程较复杂，识别所需的网络很大，同时神经网络对学习样本的依赖也很大，因此对未学习字符 1 Introduction Artificial neural network learning and recognition of characters, according to the usual method, the character’s position data input directly to the sensitivity of neural networks, due to the character writing size, location differences in the input network must be pre-processed, that is, the formal characters The This process is more complex, recognition of the network required for large, while the neural network is also dependent on the learning sample is very large, so the un learned characters