一种医学图像增强算法  

A medical image enhancement algorithm based on indiscernibility relation

作　　者：王丽丽[1] 宋余庆[1] 

机构地区：[1]江苏大学计算机科学与通信工程学院,江苏镇江212013

出　　处：《辽宁工程技术大学学报（自然科学版）》2013年第9期1274-1277,共4页Journal of Liaoning Technical University (Natural Science)

基　　金：江苏省自然科学基金资助项目(BK2012209)

摘　　要：为实现医学图像中感兴趣区域辨识度的增强.针对医学图像中CT图的特点提出一种增强算法(Rough Computed Tomography Algorithm,RCTA).算法以粗糙集(Rough Sets,RST)中不可分辨关系理论为基础,根据医学图像中不同的人体组织对应不同的CT值的特点,来定义等价关系,从而将医学图像划分为不同区域,然后对感兴趣区域灰度值保持不变,对其他区域的灰度值作最大化/最小化处理.用RCTA对临床300余张肺部医学图像进行了实验,最后,使用DSM(Distribution Separation Measurement)对RCTA与其他3种常用的增强算法进行量化对比.结果表明:RCTA对医学图像感兴趣区域辨识度的增强,有较好的效果.In order to enhance identification of interested region in medical image,an enhanced algorithm(Rough Computed Tomography Algorithm,RCTA)according to the characteristics of CT graph in medical images was proposed.This algorithm is based on indiscernibility relation theory in Rough Set,and the equivalence relation was defined according to the characteristic of different human tissue in medical image corresponding to different CT value,then the medical image was divided into different regions.The gray value of interested region remains unchanged,and the gray value of other region is for maximization/minimization.Experiments with RCTA algorithm were done on clinical more than 300 pulmonary medical images,finally,the RCTA and other three kinds of commonly used enhancement algorithms were quantitatively contrasted by means of DSM.The results show that RCTA has a good effect on identification enhancement of interested region in medical image.

关 键 词：RCTA 粗糙集 不可分辨关系 医学图像 图像增强 像素 CT值 DSM 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]