医学信息
醫學信息
의학신식
MEDICAL INFORMATION
2014年
11期
40-41,42
,共3页
赵成%沈若武%苏华伟%晁凡%管文敏%刘晓彤%卞明心%李顺见%夏玉军
趙成%瀋若武%囌華偉%晁凡%管文敏%劉曉彤%卞明心%李順見%夏玉軍
조성%침약무%소화위%조범%관문민%류효동%변명심%리순견%하옥군
Mimics%三维重建%重建方法%医学影像学%肝脏%血管%男性
Mimics%三維重建%重建方法%醫學影像學%肝髒%血管%男性
Mimics%삼유중건%중건방법%의학영상학%간장%혈관%남성
Mimics%3D-reconstruction%Methods,reconstruction%Medical imaging%Vascular system,liver%Human males
目的对Mimics10.01软件在三维重建中的3种分割方式进行比较,为该软件在基础医学等医学领域的开发应用提供新的思路与方法。方法选取100例无器质性病变的成年男性的肝脏器官,利用64位螺旋CT的薄层扫描技术,基于1.00mm层厚,1026层人体肝脏的连续断层图像,将其直接以Dicom格式的原始图像读入Mimics软件中,采用"直接分割"、"加法分割"和"减法分割"3种模型分割技术,获得肝脏的轮廓及肝内血管系统的蒙板,将蒙板内相邻像素连接并重组成图像后获得三维立体模型。结果3种三维重建方法得到的肝脏模型轮廓清楚,可实现旋转,方便角度、体积测量,并且可以得到任何重建模型的断面。"直接分割"方法能够较轻易地重建肝脏的血管主干及大的分支;"加法分割"方法可以在某个局部主动将几条血管的2级、3级甚至终末分支实现重建;"减法分割"可以将视野内的几乎所有血管的分支分割出来实现精确重建,三种方法重建出肝动脉、肝门静脉和肝静脉的个数之间比较,有显著性差异(<0.05)。结论以上3种方法各有优、缺点,其中,"减法分割"方法重建比其他两种方法信息丢失少,效果更加逼真,是目前可实现的最佳分割方式。
目的對Mimics10.01軟件在三維重建中的3種分割方式進行比較,為該軟件在基礎醫學等醫學領域的開髮應用提供新的思路與方法。方法選取100例無器質性病變的成年男性的肝髒器官,利用64位螺鏇CT的薄層掃描技術,基于1.00mm層厚,1026層人體肝髒的連續斷層圖像,將其直接以Dicom格式的原始圖像讀入Mimics軟件中,採用"直接分割"、"加法分割"和"減法分割"3種模型分割技術,穫得肝髒的輪廓及肝內血管繫統的矇闆,將矇闆內相鄰像素連接併重組成圖像後穫得三維立體模型。結果3種三維重建方法得到的肝髒模型輪廓清楚,可實現鏇轉,方便角度、體積測量,併且可以得到任何重建模型的斷麵。"直接分割"方法能夠較輕易地重建肝髒的血管主榦及大的分支;"加法分割"方法可以在某箇跼部主動將幾條血管的2級、3級甚至終末分支實現重建;"減法分割"可以將視野內的幾乎所有血管的分支分割齣來實現精確重建,三種方法重建齣肝動脈、肝門靜脈和肝靜脈的箇數之間比較,有顯著性差異(<0.05)。結論以上3種方法各有優、缺點,其中,"減法分割"方法重建比其他兩種方法信息丟失少,效果更加逼真,是目前可實現的最佳分割方式。
목적대Mimics10.01연건재삼유중건중적3충분할방식진행비교,위해연건재기출의학등의학영역적개발응용제공신적사로여방법。방법선취100례무기질성병변적성년남성적간장기관,이용64위라선CT적박층소묘기술,기우1.00mm층후,1026층인체간장적련속단층도상,장기직접이Dicom격식적원시도상독입Mimics연건중,채용"직접분할"、"가법분할"화"감법분할"3충모형분할기술,획득간장적륜곽급간내혈관계통적몽판,장몽판내상린상소련접병중조성도상후획득삼유입체모형。결과3충삼유중건방법득도적간장모형륜곽청초,가실현선전,방편각도、체적측량,병차가이득도임하중건모형적단면。"직접분할"방법능구교경역지중건간장적혈관주간급대적분지;"가법분할"방법가이재모개국부주동장궤조혈관적2급、3급심지종말분지실현중건;"감법분할"가이장시야내적궤호소유혈관적분지분할출래실현정학중건,삼충방법중건출간동맥、간문정맥화간정맥적개수지간비교,유현저성차이(<0.05)。결론이상3충방법각유우、결점,기중,"감법분할"방법중건비기타량충방법신식주실소,효과경가핍진,시목전가실현적최가분할방식。
Objective To explain three 3D reconstruction methods of Mimics10.01 software in order to provide the mathematical model for the basic medical teaching and clinical medicine. Methods The livers of the adult male without structural diseases in 100 cases were selected. The thin layer scan technique of 64-slices spiral CT and CT continuous pictures based on 1026 layer (1.00mm) of human body livers were used. Mimics was used to read the original Dicom format pictures directly. With 3 kinds of model-handling methods which were named "direct segmentation"、"additive segmentation"、"subtraction segmentation", the new masks of the surface of livers and the intrahepatic vessels were gotten. The 3D Object model was got after connection of adjacent Pixel in the new mask and reconstruction. Results The liver model that was got by three 3D reconstruction methods was clear. The angle and volume could be calculated via the measurement tool easily. 3D Object model could be cut and any kind of cross-sections in 3D Object model was got en as well. Vascular trunk and large branches in livers were easily reconstructed by the method of "direct segmentation";two, three or even terminal branches of several blood vessels could be reconstructed initiatively in a part of livers by the method of"addition segmentation"; almost al of the blood vessels branches in the field of vision could be reconstructed by the method of"subtraction segmentation". Through the comparison about the number of hepatic arteries, portal veins and hepatic veins which are reconstructed by three kinds of methods, the dif erences of the number of vessels by three measurements are statistical y significant ( <0.05). Conclusion All of the three 3D reconstruction methods have their advantages and disadvantages. But based on the data of measurement, the 3D reconstructions by "subtraction segmentation" are more realistic and lose less datas than the other two methods. So this method is the most effective way to get 3D reconstruction for clinical medicine.