基于BP神经网络的含噪字符识别系统及MATLAB实现

A BP Neural Network to Recognize Block Letters with Disturbing Signals

运用BP神经网络的原理,在MATLAB的平台上,对含噪与不含噪的印刷体英语字符进行识别和误差统计,经大量样本测试,结果达到设计目标,辨认出错率较低。

A BP neural network was designed to recognize block letters with disturbing signals, The recognition system was tested by a large amount of cases with different level disturbing signals. The test results show that the errors rate of recognition systemcan reach the design aim.