阈值分割与动态轮廓模型在胃壁磁共振成像中的应用  

MRI Kidney Segmentation Based on Thresholding and Active Contour Model

作　　者：吴晓唯[1] 周传清[1] 任秋实[1] 

机构地区：[1]上海交通大学生物医学工程系,上海200240

出　　处：《中国医疗器械杂志》2011年第3期161-163,共3页Chinese Journal of Medical Instrumentation

基　　金：973计划[2010CB933903];高等教育博士点基金[200802481029];上海市科委重点项目(10JC1407900);教育部高等学校创新引智计划("111"计划)[B08020]

摘　　要：提出了一种基于肾脏核磁共振图像来提取肾脏内部的皮质与髓质区域的图像分割算法,它结合了阈值分割与动态轮廓模型。首先,通过预处理和阈值分割得到初始的二值图像;接着,手动设定起始轮廓,它在简化后的动态轮廓模型作用下,模仿"气球"膨胀变形,最终,得到所需目标组织的初始轮廓;最后,它在传统的动态轮廓模型作用下停靠在由原图的梯度图像定义的组织边缘处。The proposed hybrid image segmentation method,based on MRI renal images,aims at extracting the cortical and medullary regions of the kidney.It incorporates the thresholding segmentation and the active contour model.To begin with,the binary mask was obtained by preprocessing and thresholding.Then,the manually set starting contour deformed like a balloon under the framework of a simplified active contour model,thus getting the initial contour of the medulla or the cortex.Finally,the initial contour deformed according to the traditional active contour model and was attracted by the actual boundary of the tissues,defined by the gradient image.

关 键 词：肾脏 核磁共振图像 阈值分割 动态轮廓模型 

分 类 号：R445.2[医药卫生—影像医学与核医学]