应用迭代四叉树的肝脏MR图像分割  

Liver MR image segmentation with iterative quadtree decomposition

作　　者：迟冬祥[1] 程伟中[2] 

机构地区：[1]上海电机学院电子信息学院,上海200240 [2]复旦大学附属中山医院放射科,上海200032

出　　处：《计算机工程与应用》2010年第9期142-145,共4页Computer Engineering and Applications

基　　金：上海市教育委员会科研创新项目资助(No.08YZ192);上海电机学院科研启动项目资助(No.07C402)

摘　　要：磁共振(Magnetic Resonance,MR)图像的诊断是公认的确认肝脏有无肿瘤等器质性病变的金标准方法,因此肝脏的正确分割对计算机辅助诊断有非常重要的意义。由于脏器组织浸润和个体差异,在肝脏分割实现方法方面有一定难度,目前尚没有通用的医学分割方法。在既有研究的基础上,提出了基于四叉树的迭代分割算法,得到MR图像中肝脏的自动分割结果。实验分割结果表明这种方法的可行性和优势,并为后续的肿瘤提取奠定基础。Magnetic Resonance（MR）images are generally acknowledged as a golden standard for diagnosis of liver tumors.Therefore a good liver segmentation result is very important to computer-aided-diagnosis.Because of infiltrations among the internal organs and their individual differences,it is difficult for a liver segmentation and there is no universal medical method of segmentation.In this paper,on the basis of formal research,an iterative quadtree decomposition algorithm is put forward and liver image is segmented automatically.Experimental results show its feasibility and advantages and lay a foundation for the tumor extraction.

关 键 词：磁共振图像 迭代四叉树 肝脏分割 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]