【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。更多还原

【Abstract】 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.更多还原