一种具有选择注意特性的模板匹配网络模型  

A Neural Network for Selective Attentional Template Matching

作　　者：叶芗芸[1] 戚飞虎[1] 尹虎君[1] 

机构地区：[1]上海交通大学计算机科学与工程系,上海200030

出　　处：《电子学报》1998年第8期22-26,88,共6页Acta Electronica Sinica

基　　金：国家自然科学基金!69772002

摘　　要：本文提出了一种选择性注意相似度量、并构造了一种用于模板匹配分类的网络模型（SelectiveAttentionalTemplateMatchingNetwork）.该网络由匹配子网和竞争子网络成：匹配子网络完成输入模式与样本模式的选择注意性比较；竞争子网络则选择最佳匹配的模式作为分类结果。本文提出的选择性注意相似性度计不便叫用于二值模式问的比较，也可用于连续模式问的比较实验证明该相似性度量较若干种常用相似性度量更能突出反映模式问的差异．A kind of similarity measure with selective attentional property is proposed. The neural network (Selective Attentional Template Matching Network, SATMN in brief)model built with the proposed measure can be applied to template matching and classification. The model is composed of matching subnetwork and competitive subnetwork, in which comparation and classificatiorl are carnal out respectively. The similarity measure is available to both binarized and continous patterns, and the experimental results show thai it is superior to several widely used similarity measures in describing the variance between patterns. The implement method and experimental results for neural network are given.

关 键 词：选择性注意 模板匹配 相似性度量 神经网络 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]