电子学报
電子學報
전자학보
ACTA ELECTRONICA SINICA
2014年
2期
398-404
,共7页
吕泽华%赵盛荣%梁虎%唐赫
呂澤華%趙盛榮%樑虎%唐赫
려택화%조성영%량호%당혁
乳腺肿块分割%Gmac (Global minimum active contour ) 模型%split bregman方法
乳腺腫塊分割%Gmac (Global minimum active contour ) 模型%split bregman方法
유선종괴분할%Gmac (Global minimum active contour ) 모형%split bregman방법
breast lumps segmentation%global optimal active contour model%split bregman method
在乳腺CAD系统中,乳腺肿块分割是一个重要的先前步骤,分割结果的好坏直接影响到肿块的分类和检测.本文将Gmac模型应用到乳腺肿块分割上,并提出了求解Gmac模型的两种改进方法:改进的变分水平集法、改进的split bregman方法.实验选取了483幅医学乳腺肿块图片进行分割,得到了两种改进方法的CM均值分别为64%和76%;AMED均值分别为4.4750和1.4602.结果表明:改进的split bregman方法对乳腺肿块进行了更有效的分割.实验也利用经典的ACWE模型和GAC模型对上述乳腺肿块图片进行了分割实验,与基于改进split bregman方法的Gmac模型相比,结果表明:Gmac模型具有更好的分割性能.
在乳腺CAD繫統中,乳腺腫塊分割是一箇重要的先前步驟,分割結果的好壞直接影響到腫塊的分類和檢測.本文將Gmac模型應用到乳腺腫塊分割上,併提齣瞭求解Gmac模型的兩種改進方法:改進的變分水平集法、改進的split bregman方法.實驗選取瞭483幅醫學乳腺腫塊圖片進行分割,得到瞭兩種改進方法的CM均值分彆為64%和76%;AMED均值分彆為4.4750和1.4602.結果錶明:改進的split bregman方法對乳腺腫塊進行瞭更有效的分割.實驗也利用經典的ACWE模型和GAC模型對上述乳腺腫塊圖片進行瞭分割實驗,與基于改進split bregman方法的Gmac模型相比,結果錶明:Gmac模型具有更好的分割性能.
재유선CAD계통중,유선종괴분할시일개중요적선전보취,분할결과적호배직접영향도종괴적분류화검측.본문장Gmac모형응용도유선종괴분할상,병제출료구해Gmac모형적량충개진방법:개진적변분수평집법、개진적split bregman방법.실험선취료483폭의학유선종괴도편진행분할,득도료량충개진방법적CM균치분별위64%화76%;AMED균치분별위4.4750화1.4602.결과표명:개진적split bregman방법대유선종괴진행료경유효적분할.실험야이용경전적ACWE모형화GAC모형대상술유선종괴도편진행료분할실험,여기우개진split bregman방법적Gmac모형상비,결과표명:Gmac모형구유경호적분할성능.
In breast-CAD system ,the segmentation of breast mass is an important step for pretreatment .In this work ,the Gmac model is applied for breast mass segmentation ,meanwhile ,two improved methods to solve the Gamc model are proposed :the improved Level-set algorithm and the improved split-bregman algorithm .In experiment part ,we select 483 medical images for test . The CM for the two algorithms is 64% and 76% and AMED is 4.4750 and 1.4602 respectively .The experimental results show the improved split bregman method is more effective for breast-mass segmentation .Besides ,the ACWE model and GAC model are used for comparison and the results indicate that the Gmac model is more suitable for breast-mass segmentation .