【摘要】 神经网络对手写字符有较强的识别能力 ,关键在于特征的提取。本文提出按字符字段和非均匀网络结构浓度提取手写文字特征的改进方法。更多还原

【Abstract】 Conpared with the adaptability of neural networks on dealing with handwriting recognition ,we showed the shortcomings of the traditional recognizing method,and presented two effective approaches,which extract word’s features by character and by the chroma of heterogeneous networks structure.更多还原