一种B超序列心脏图像的腔体分割方法及其在心腔体积定量计算中的应用  

Heart Chamber Segmentation of Serial Ultrasound Heart Images Based on Snake Model and its Application in Volume Measurement in Single Heart Chamber

作　　者：罗渝兰[1] 王景熙[1] 郑昌琼[2] 尹鹏[1] 

机构地区：[1]四川大学计算机学院,四川成都610065 [2]四川大学生物医学工程中心,四川成都610065

出　　处：《四川大学学报（工程科学版）》2002年第3期114-117,共4页Journal of Sichuan University (Engineering Science Edition)

基　　金：国家自然科学基金资助项目 (6963 10 2 0 )

摘　　要：采用一种Snake模型 ,提出了针对高噪声B超心脏图像的腔体分割算法 ,并将这一方法运用到序列图像的分割中 :对于序列中的第一幅图像 ,通过人工取点获得初始闭合轮廓 ,再运用Snake模型对整个序列图像进行自动的轮廓逼近。最后将所有图像的分割结果运用于单个心腔体积的定量计算中。实验结果证明了该方法的可行性和有效性。Based on a new Snake model, a segmentation algorithm is proposed to segment heart chambers (ventricle or atria) of ultrasound heart image with high noise, which is applied to serial images. In the first image of the serial, a raw contour is gotten by marking some points manually. Then closed contour in each image is drawn by Snake tracking automatically. In addition, a volume measurement of single heart chamber is presented with the segmentation algorithm. Our experiments show that the new algorithm has the advantages of feasibility and efficiency.

关 键 词：B超序列心脏图像 腔体分割 定量计算 SNAKE模型 梯度矢量流 图像分割 边缘检测 体积测量 心腔 心脏疾病 诊断 

分 类 号：R541[医药卫生—心血管疾病] R540.4[医药卫生—内科学]