引入模糊因子的EM图像分割  

Medical Image Segementation Algorithm Based on Fuzzy Factor Constrained EM

作　　者：胡亨伍[1] 李松涛[1] 

机构地区：[1]广东医学院信息工程学院,广东东莞523808

出　　处：《电脑与信息技术》2016年第1期18-19,30,共3页Computer and Information Technology

摘　　要：基于有限高斯混合模型,在采用EM算法对混合模型参数求解的过程中,引入模糊因子,改善图像分割方法。该方法可以降低其对初选值的依赖和像素点值对混合模型分量间参数的干扰。结果表明,对含有高斯噪声影响的图像有较好的分割结果,目标图像的区域形状保持好、相对位置明确。According to the mixture model in an image, this paper presents a new fuzzy factor constrained mixture model for image segementation. In the process for solving the parameters of model by the standard EM algorithm, the fuzzy factor is introduced to correct the indepandent assumption of the pixel and reduce the influence among the parameters of mixture components. The results show that the image segmentation results are better when the Gauss noise influence is dissolved. The target image is kept in shape and the location is good.

关 键 词：混合模型 图像分割 模糊因子 EM算法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]