基于双重竞争共振机制的模糊ART神经网络  

Fuzzy ART neural network based on dual competition and resonance technique

作　　者：汪国有[1] 张磊[1] 王文涛[1] 

机构地区：[1]华中科技大学图像识别与人工智能研究所,湖北武汉430074

出　　处：《华中科技大学学报（自然科学版）》2006年第10期34-36,共3页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：航天支撑基金资助项目;武器装备预研基金资助项目

摘　　要：提出了一种基于双重竞争共振机制的模糊ART神经网络模型．该模型将输入节点的竞争共振机制引入到输出类别节点，采用输入节点和输出节点双重竞争共振机制，形成了一种新的模糊ART结构，解决了传统模糊ART网络记忆不稳定问题．将该模型应用于图像分割，解决了传统模糊ART网络图像分割结果随警戒参数的升高而出现的过度分割．实验结果表明，和原始模糊ART算法相比，新算法具有更好的分类识别性能，在飞机目标识别中平均识别率比原始算法提高3％～5％．The competition and resonance mechanism of input nodes was introduced into output categorization nodes in the new neural network model. This dual competition and resonance technique for the input and output nodes establishes a new Fuzzy ART neural network structure to improve the memory stability in traditional Fuzzy ART neural networks. In the application of the image segmentation, the new model solves the over-segmentation of the original Fuzzy ART neural network algorithm due＇ to the vigilance＇s increase. Experimental results show that compared with the original Fuzzy ART neural network, this algorithm has better clustering performance and the average improvement is about 3 %～5% in the aeroplane target recognition.

关 键 词：模式识别 双重竞争共振机制 模糊ART神经网络 

分 类 号：TP273.4[自动化与计算机技术—检测技术与自动化装置]