序列颅脑CT图像的颅腔内结构自动化分割  

A morphology and knowledge-based automatic segmentation of intracranial structures on cerebral computerized tomography

作　　者：李传富[1] 周康源[1] 陈曾胜[1] 黄丹[1] 何力[1] 王庆临[1] 

机构地区：[1]中国科学技术大学电子工程与信息科学系

出　　处：《中国科学技术大学学报》2006年第2期148-152,共5页JUSTC

基　　金：安徽省教育厅自然科学基金(2003kj238)资助

摘　　要：为实现颅腔内结构的自动化分割,利用颅脑CT的解剖学和影像学等先验知识,提出了一种基于图像形态学和颅腔内结构连续性等先验知识的序列颅脑CT颅腔内结构的自动化分割算法.通过对连续100个颅脑CT检查病例的自动化分割,结果表明,该算法能够实现所有具有完整颅骨环的颅腔内结构的计算机自动化分割(97/100),分割结果准确.In order to segment intracranial structures on cerebral computed tomography （CCT）. With the prior knowledge of anatomy and diagnostic imaging, a method of automatic segmentation of intracranial structures on CCT was put forward, based on the morphology and the connectedness of intracranial structures. The segmentation results of 100 cases of consecutive CCT demonstrated that satisfactory results had been achieved in all the cases with closed skull （97/100）. However, further research is needed on those cases without closed skull.

关 键 词：先验知识 图像形态学 图像分割 颅脑CT 

分 类 号：R318[医药卫生—生物医学工程]