基于形心同心圆结构的自由手写体数字神经网络分类器

本文提出了一种基于自由手写体数字的形心同心圆结构来提取贯穿特征码的神经网络识别方法。该方法是用自由手写体数字的形心同心圆来抽取其贯穿特征码，将获得的这些模式特征训练改进的ＢＰ神经网络分类器，从而达到快速分类的目的。将其应用于自由手写体数字的信函自动分拣系统，单字的识别率达到９７％以上，整信的识别率也可达到９２％以上，得到了令人满意的结果。

NEURAL NETWORK CLASSIFIER FOR UNCONSTRAINED HANDWRITTEN NUMERALS BASED ON SHAPE-CENTER CONCENTRIC CIRCLES

A new neural network classifier is proposed for unconstrained handwritten numerals based on the features of shape-center concentric circles in this paper.The method is extracting the cross feature codes of concentric circles based on the shape-center of unconstrained handwritten numerals, using the obtained cross feature codes to train the back propagation neural network classifier, and then recognizing these characters. Applying the classifier to the zip code recognition system, it is shown that the average recognition rate of single digit is above 97%,and the average recognition rate of total letter is about 92%. The experimental results are very satisfactory.