基于形态学的MR图像左心室自动定位算法

The Auto-location of MRI Left Ventricular base on Mathematical Morphology

左心室重建为3D空间模型能快速评价各种心脏病,左室扭转揭示了重要的生理及病理发现。但由于左心室MRI图像具有弱边缘、与周围的组织之间存在低对比度区域,利用Snake模型分割左心室图像时,一般需要采用半自动的方法获取初始曲线。为解决此问题,本文提出一种MR图像的左心室自动定位算法,用于的左心室MRI图像分割算法。该算法利用形态学理论自动获取左心室的重心以及边界形状变化允许空间,勾画出相应初始轮廓曲线,以便进行重建。文中以小香猪左室加标记MRI作为应用研究对象,进行了全图像序列的研究。实验表明改算法能够精确的获取左心室区域信息,实现其自动定位。

3D space model of left ventricular reconstruction can estimate several heart diseases fleetly, and reverse movement of left ventricular reveals the important physiological and pathological phenomenon. Owing to weak edges and low contrast area, the initial curve has to obtain by manual when the image was segmented with snake model. Therefore, an auto-location of left ventricular base on mathematical morphology is introduced to solve this problem. This algorithm acquires the barycenter and boundary shape with the morphology theory automatically, and then draw initial curve of left ventricular area for reconstruction. In this paper, the auto-location of small pigs＇ left ventricle from Tagged MRI shows that this algorithm can obtain precise left ventricular regional.