核素肺通气/灌注显像的伪影识别与分割  

ARTIFACT IDENTIFICATION AND SEGMENTATION FOR RADIONUCLIDE VENTILATION AND PERFUSION LUNG IMAGING

作　　者：王静[1] 李素晓[1] 董洁[1] 姜恩宇[1] 赵书俊[1] 

机构地区：[1]郑州大学物理工程学院,河南郑州450052

出　　处：《计算机应用与软件》2010年第4期249-252,共4页Computer Applications and Software

摘　　要：SPECT核素肺通气/灌注显像的定量分析是诊断肺动脉栓塞的一种有效手段。对通气/灌注图像中肺部区域能否准确定位及有效分割是整个定量分析的前期工作与基础。提出一种均值检测法实现核素肺部成像的伪影识别及感兴趣区ROI提取,并基于二维最大熵理论确定阈值分割出双肺影像。通过对16例正常人和10例肺栓塞病患的SPECT通气/灌注图像进行处理,该方法提取效率高、运算速度快,较好地实现了肺部感兴趣区定位与分割。Quantitative analysis of the SPECT radionuclide ventilation and perfusion lung imaging is one of the effective ways for diagnosing pulmonary embolism（PE）.It is the preparatory work and base of the whole quantitative analysis that whether or not the pulmonary region in ventilation and perfusion image can be accurately located and effectively segmented.This paper proposes a method of mean detection to realise artifact identification and region of interest extraction of radionuclide pulmonary imaging,then based on fast two-dimensional maximizing entropic theory the thresholding is worked out to have segmented the image with both lungs.By processing the SPECT ventilation and perfusion images of 16 patients with normal scans and 10 with PE,the method is proved to have high extraction efficiency and fast calculating speed,and realises good location and segmentation of the pulmonary region of interest.

关 键 词：SPECT核素肺通气/灌注显像 伪影识别 二维最大熵 阈值分割 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]