论文部分内容阅读
介绍了用于人脸模式自动识别系统中的ＢＰ网的基本原理和算法及步骤，提出了隐层神经元个数的物理意义和选取原理，讨论了Ｓｉｇｍｏｉｄ函数的改进形式和网络初始权值的设置原则．
This paper introduces the basic principle, algorithm and procedure of BP neural network used in face recognition automatic recognition system. The physical meaning and selection principle of neurons in hidden layer are proposed. The improved form of Sigmoid function and the initial weight of network Set the principle.