基于混合瑞利分布的三维医学图像分割算法  

THE ALGORITHM OF 3D SEGMENTATION ON MEDICAL IMAGE BASED ON MIXED RAYLEIGH DISTRIBUTION

作　　者：孟令旗[1] 林财兴[1] 崔芳芳[2] 骆建珍[1] 

机构地区：[1]上海大学CIMS＆机器人中心,上海200072 [2]河北工业大学机械工程学院,天津300130

出　　处：《计算机应用与软件》2009年第1期236-239,共4页Computer Applications and Software

摘　　要：水平集算法是当前比较重要的分割算法,研究该算法的三维分割技术。提出高斯分布和瑞利分布相结合的方法进行概率密度函数的确定与分析,通过随机极大似然算法(SEM)把瑞利分布引入到水平集三维分割演化模型中。该方法成功应用于多种医学组织的三维分割,提高了分割速度和分割效果,减少了参数的设置。The 3D segmentation technique in Level Set algorithm is studied,the Level Set is a rather popular segmentation algorithm currently. A new way to evaluate and analysis the probability density function of medical images is proposed .with the integration of Gaussian dis- tribution and Rayleigh distribution, and the Rayleigh distribution is introduced to the evolution model of 3D segmentation of Level Set through Stochastic Expectation-Maximisation algorithm. The method has been successfully applied to quite a few cases of 3D segmentation of medical images of tissues. It improves the segmentation speed and effect,and reduces the parameters established.

关 键 词：瑞利分布 三维分割 概率密度函数 水平集 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TN911.73[自动化与计算机技术—计算机科学与技术]