摘要
利用前向多层神经网络的反向传播算法 ,即BP算法 ,采用MATLAB软件构建用于特征识别的两层前向神经网络 ,将已知字母经图象处理后对BP网络进行训练 ,此BP神经网络可以成功地识别A ～Z2
In this paper, the back propagation algorithm of a feedforward neural network was used with MATLAB, and a multilayer feedforward neural network was founded for the character distinguishing. After image processing, the known example English letters were made to train the BP neural network. So the BP neural network would be able to distinguish 26 written English letters from A to Z .
出处
《应用科技》
CAS
2001年第10期28-30,共3页
Applied Science and Technology
