摘要
本系统基于视感神经网络(Perception Neural Network)的算法结构,着眼于神经网络的实用化,采用计算软件计算关联权重,通过光电计算机结合,实现了判断并学习1—9九个字符图形的实验系统,得到了实验结果。本系统是对神经网络应用于图像识别的新的尝试。
The present system,based on the algorithm strueture of the per- ception neural network,utilizes computing software to comput cognate weight, with the aim of implementing practical applications of neural network.An ex- perimental system is realized for deciding and learning graph of nine charac- ters from l to 9 by means of the integration of optical computer with its elec- tronic counterparts.It isa new attempt for the system to use neural network in image recognition.
出处
《半导体光电》
CAS
CSCD
北大核心
1991年第3期294-297,共4页
Semiconductor Optoelectronics
