摘要
自动的人脑核磁共振(MR)图像分割是许多医学图像应用的关键问题。该文提出了一种有效的自动脑核磁共振图像的分割方法框架体系,脑MR分割框架体系由3个处理步骤构成。首先,采用基于水平集的方法将MR 图像中非脑组织剔除,从脑图像中提取大脑组织结构。然后,对MR脑结构图像进行灰度不均匀性校正。最后,该算法采用最大后验分类器可以将人脑组织分为脑白质、脑灰质、脑髓液。在实验中对大量的MR脑图像数据应用该分割算法。实验结果充分证明该方法的有效性。这种分割算法适用于人脑核磁图像分析的各种实际临床应用。
Automatic segmentation of brain magnetic resonance images is a critical problem in many medical imaging applications. In this paper, a robust automated segmentation algorithm is presented for the brain magnetic resonance images. The segmentation framework is composed of three stages. First, it uses level set method to perform the brain stripping operation. In the second stage, it compensates for nonuniformity in the brain image based on computing estimates of tissue intensity variation. Finally, a maximum aposteriori classifier is used to partition the brain into gray matter, white matter, and cerebrospinal fluid. The proposed method has been tested using magnetic resonance dada. This algorithm may be applied to various research and clinical investigations in which brain segmentation and volume measurement involving Magnetic resonance images dada are needed.
出处
《电子与信息学报》
EI
CSCD
北大核心
2005年第9期1420-1424,共5页
Journal of Electronics & Information Technology
基金
国家自然科学基金(30000224
30000056)资助课题
关键词
核磁共振成像
偏场校正
水平集
马尔可夫随机场
分割
Magnetic resonance imaging, Bias correction, Level set method, Markov random field, Segmentation
