基于改进局部信息的模糊C均值MR图像分割
Fuzzy C-mean Clustering Based on Improved Local Information for MR Image Segmentation
李征 1郑志帅2
作者信息
- 1. 河南思维轨道交通技术研究院有限公司,河南郑州 450000
- 2. 许继集团有限公司综合能源服务分公司,河南许昌 461000
- 折叠
摘要
为了确保模糊C均值(FCM)在分割磁共振图像时能取得更好的分割结果,在基于局部信息的模糊C均值算法(FLICM)框架下,对邻域项进行改进,提出一种新的FLICM算法.首先,使用非局部均值滤波对原图像进行处理生成附加图像,根据附加图像信息和原始图像信息定义像素一致性系数,再对像素点的含噪情况进行衡量,避免含噪像素对邻域项产生过大影响.然后,构造一个中心像素与邻域像素差异性系数,用于衡量邻域像素的灰度差,并与像素一致性系数相结合建立像素相关系数,从而更有效地计算邻域像素的相关性.最后,使用像素相关系数构造新的模糊因子并与FLICM算法相结合,得到改进的FLICM算法.通过与 3 种算法在不同图像上的比较表明,该算法能得到更准确的分割结果.
Abstract
In order to ensure that the fuzzy C-means(FCM)can achieve better segmentation results when segmenting MRI images,the neighborhood item is improved under the framework of the fuzzy C-means algorithm based on local information(FLICM),and a new FLICM algorithm is proposed.First,the additional images is generated by non-local mean filtering on original image.The pixel consistency coefficient is defined according to the additional image information and the original image information,and the noise level of pixel points is measured to avoid the excessive influence of the noisy pixel on the field items.Then,a difference coefficient between the center pixel and the neighborhood pixel is constructed to measure the gray difference of the neighborhood pixel,and the pixel correlation coefficient is established in combination with the pixel consistency coefficient,which can calculate the correlation of the neighborhood pixels more effectively.Finally,a new fuzzy factor is constructed with pixel correlation coefficient and combined with FLICM algorithm to obtain an improved FLICM algorithm.Compared with the three algorithms on different images,the proposed algorithm can get more accurate segmentation results.
关键词
图像处理/图像分割/模糊C均值聚类/模糊因子/像素一致性/像素差异性Key words
image processing/image segmentation/fuzzy C-means clustering/fuzzy factor/pixel consistency/pixel difference引用本文复制引用
出版年
2024
NETL
NSTL
万方数据