一种自适应阈值的模糊连接算法及其在肝脏血管分割中的应用  

Vascular Segmentation in Hepatic CT Images Using Adaptive Threshold Fuzzy Connectedness Method

作　　者：符晓珠 黄绍辉[1] 王博亮[1] 黄晓阳[1] 

机构地区：[1]厦门大学信息科学与技术学院,福建厦门361005

出　　处：《厦门大学学报（自然科学版）》2015年第4期546-550,共5页Journal of Xiamen University：Natural Science

基　　金：国家自然科学基金(61001144;61102137;61327001)

摘　　要：近年来模糊连接算法已经在医学图像分割中得到了应用.然而将此算法直接应用于肝脏血管分割时,由于其过高的计算成本以及需要手动选取阈值,分割效率并不是非常理想.为此,提出一种基于查找表计算模糊场景的方法,能够使模糊连接算法的运算速度提高1.5~2倍;同时提出一种类分水岭的自适应阈值搜索算法,能够对肝脏血管分割的阈值进行自动选取,从而实现了整个模糊连接分割流程的自动化处理.并且以3套肝脏CT数据为实验对象进行了验证,实验结果表明基于查找表的计算方法具有高效性,并且根据自适应阈值进行肝脏血管分割能产生正确的结果.In recent years,fuzzy connectedness method （FCM） was used to extract fuzzy objects from medical images and show its effectiveness. However,when FCM was applied to hepatic vessel segmentation task, two problems may occur. One is the expensive computational cost^the other is the difficulty of choosing a proper threshold value. In this paper, an accelerated method which is based on a lookup table is presented first. This method can reduce the connectivity scene calculation time and achieve a speed-up factor of a- bout 1.5-2. When this step finished,FCM needs a threshold to generate the final result. Currently this threshold can only be guessed by users. Since different thresholds may generate different results,a proper threshold is usually needed. By analyzing the hepatic ves- sel structure,a watershed-like method can then be used to find the threshold. Experiments based on three different data sets demon- strate the efficiency of the lookup table method. These experiments also show that the threshold found by this method can usually generate correct segmentation results.

关 键 词：肝脏血管分割 模糊连接算法 模糊亲和力 自适应阈值 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]