医疗特征图像边缘曲率分析实例  

A practical example of edge curvature rate analysis on medical feature image

作　　者：马旭[1] 杨文美[2] 

机构地区：[1]辽宁大学计算中心,辽宁沈阳110036 [2]辽宁大学校医院口腔科,辽宁沈阳110036

出　　处：《医学影像学杂志》2010年第11期1713-1715,共3页Journal of Medical Imaging

基　　金：沈阳市科学计划项目(061833)

摘　　要：目的:以X线口腔龋齿图像为例,提出其边缘曲率定义及计算的实用方法。参数形式的离散曲率定义及有限差分法的计算虽然受图像噪声影响较大,但针对已去噪处理的医疗特征图像,其简单、实用,是可行的。依据实验数据,采用MATLAB编写了相应的分析函数,返回龋病分析中所需的曲率特征值向量X。对实验数据图示并列表,发现图2两峰差值很大,且间距又小,说明曲线变化幅度大且剧烈,是龋病龋蚀严重的牙体曲线;图3两峰差值较大,间距较小,说明曲线变化幅度较大且不较剧烈,是龋病龋蚀不很严重的牙体曲线;图4两峰差值很小,且间距又大,说明曲线变化幅度很小,且缓慢,是正常牙体曲线。实验判断与医疗诊断相符,证明参数形式的离散曲率定义及有限差分法的计算是可实用的。Objective：Analysis of the feature segmentation＇s edge is the key point of the medical image analysis.Taken the tooth X-ray caries image for example,this paper proposes a calculation method of edge curvature.The definition and the calculation of the parametric function is influenced easily by noise,however.It＇s simple and practical to the denoised medical image.On the basis of experimental data,with the MATLAB program,I got the vectors which are necessary to the caries analysis.After the graphical representation and list,I found that the graph 2 is deep caries,the graph 3 is mediate caries,and the graph 4 is normal tooth image.The experimental judgment is consistent with the medical diagnosis.

关 键 词：医疗图像特征边缘 边缘曲率 龋病 病症分析 MATLAB 

分 类 号：R811[医药卫生—放射医学] R814.3[医药卫生—临床医学]