边缘检测算法在医学超声液性病变图像中的应用  

Application of edge detection algorithms in medical ultrasonography images of discharge diseases

作　　者：高海娟[1] 平子良[1] 周素华[1] 侯迎宾 

机构地区：[1]北京邮电大学世纪学院,北京102613 [2]北京汇影互联科技有限公司,北京100193

出　　处：《现代电子技术》2012年第4期98-101,104,共5页Modern Electronics Technique

基　　金：国家自然科学基金资助项目(61065004

摘　　要：医学超声液性病变图像多见数个无回声区,呈"蜂窝状",边界不清晰,为了清晰地提取医学超声液性病变图像的边缘,进一步为临床诊断提供可靠依据,在此将几种不同的边缘检测算法应用于医学超声液态病变图像中,经实验结果得出,经典的边缘检测算法不能很好地提取图像的边缘,而基于Snake模型的边缘检测算法,人为设定边缘控制点,智能动态调整曲线,获得了很好的边缘提取效果,具有很高的临床应用价值。The medical ultrasound image with lesion containing liquid always show several hypoecho zones with unclear edge,like 'Honeycomb'.In order to extract clear edge of such images and generate further data for clinic diagnosis,several different edge detection algorithms are applied to medical ultrasound images with lesion containing liquid.Experiments shows,classic edge detection algorithms can't achieve desired image edge,whereas edge detection algorithm based on Snake model manually set edge control points,intellectually and dynamically adjusts curve shape,and finally get an excellent result,which shows high clinic application value.

关 键 词：医学超声图像 液性病变 边缘提取 SNAKE模型 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]