摘要
以往的免疫遗传聚类算法都要事先设置聚类数及聚类中心,采取的是有教师学习的方式,对环境的适应性不太。结合免疫网络算法和免疫遗传分类,提出了事先通过一种无教师学习,确定聚类数及聚类中心的免疫遗传分类算法,同时在聚类分类的基础上运用粗糙集对图像进行分割。通过对人脑MR图片的聚类和分割实验,验证了该方法的有效性。
The former immune genetic algorithms have beforehand to establish gathering number and gathering center,adopt the way which the teacher studies,and aren＇t good to the environment compatibility.This paper unifies the immunity network algorithm and the immunity heredity classification,proposes immunity heredity classification algorithm through one kind of non-teacher study,determines gathering number and gathering center,simultaneously utilizes the rough collection to carry on the division to the picture.Through gathered kind and the division experiment to the human brain MR picture,the result has confirmed this method validity.
出处
《计算机工程与应用》
CSCD
北大核心
2009年第16期180-181,218,共3页
Computer Engineering and Applications
