摘要
图象分割和对象提取是从图象处理到图象分析的关键步骤。本文将经典的模糊C-均值聚类算法和模糊测度和模糊积分结合起来,并将这两种算法应用于医学病理图象的分割。经典的模糊C-均值聚类算法采用欧式距离计算像素之间的相似度,本文中采用模糊测度和模糊积分计算像素之间的相似度,基于模糊测度和模糊积分的特点,这种计算方法可以提高计算的准确度。最后对两种算法的处理结果进行了比较,结果表明改进的模糊C-均值算法对医学病理图像的分割效果比原算法有所改进。
Image segmentation and object extraction are the key steps in image process. In this article we combine the fuzzy C-means algorithm with fuzzy measures and fuzzy integrals and apply the two algorithms to the medicinal pathological image segmentation. Classical fuzzy C-means algorithm calculates the similarity between pixels by Eulidean distance, now we use fuzzy measures and fuzzy integrals to express the distance between pixels, we can improve the veracity of calculation by this way because of the character of the fuzzy measures and fuzzy integrals. Results from both algorithms are compared and they show that the modified fuzzy C-means algorithm fits the pathological image segmentation better.
出处
《计算机与应用化学》
CAS
CSCD
北大核心
2005年第11期1070-1072,共3页
Computers and Applied Chemistry
关键词
模糊C-均值聚类
模糊测度
模糊积分
病理图像
fuzzy C-means clustering algorithm, fuzzy measures, fuzzy integrals, medicinal pathological image
