基于连续型Hopfield神经网络的噪声字符识别  

Identification of Noised Words Using Continuous Hopfield Neural Network Method

作　　者：王韬[1] 

机构地区：[1]空军雷达学院研究生队,武汉430010

出　　处：《系统仿真学报》2003年第9期1288-1290,共3页Journal of System Simulation

摘　　要：采用连续型Hopfield神经网络(CHNN)对噪声字符进行识别。此模型比离散型Hopfield神经网络(DHNN)和布尔神经网络(BNN)更易于硬件实现,且在信息处理的并行性和实时性等方面更接近实际生物神经网络的工作原理。仿真结果表明该方法可以有效地对噪声字符进行识别。In this paper,the method based on Continuous Hopfield Neural Network(CHNN)to identify noised words is presented. This model is easier to be implemented by hardware, and closer to the principle of practical neural network at the parallelity and real-time of signal processing than DHNN and BNN models.The results of simulation show this method can identify the noised words efficaciously.

关 键 词：连续型Hopfield神经网络(CHNN) 布尔神经网络(BNN) 监督分类器 吸引子 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]