摘要
本文提出一种通过竞争 Hopfield神经网络 (CHNN)对二维灰度向量聚类和进行图象分割的方法。该方法兼顾了图象的邻域相关信息及图象的边缘特性 ,因而分割准确、抗噪能力强。由于引入竞争学习机制 。
A method of image segmentation is proposed in this paper.It conducts image segmentation based on clustering a two dimensional gray vector by a competitive Hopfield neural network(CHNN).This new method has the virtues of more accurate segmentation and strong noise suppression ability for it takes account of both neighboring information and edge specialty of the image,and converges fast by means of a competitive learning mechanism.
出处
《计算机工程与科学》
CSCD
2000年第2期70-72,共3页
Computer Engineering & Science
基金
航空科学基金资助项目!( 95 D5 3 0 45 )
博士学科点基金资助项目!( 97-3 0 )
关键词
二维特征向量
图象分割
神经网络
图象处理
two dimensional characteristic vector
image segmentation
competitive Hopfield neural network
