基于SUV值的智能肺癌识别  

Intelligent Algorithm for Lung Tumor Positron Emission Tomography Images Identification Using Standard Uptake Values

作　　者：赖芳敏[1] 李彬[1] 田联房[1] 陈萍[2] 纪岱山[1] 

机构地区：[1]华南理工大学自动化科学与工程学院,广州510640 [2]广州医学院第一附属医院,广州510120

出　　处：《医疗卫生装备》2008年第11期8-10,共3页Chinese Medical Equipment Journal

基　　金：国家自然科学基金项目(30570458);广东省自然科学基金项目(05006530)

摘　　要：目的:提出一种新的智能识别肺部肿瘤(Positron Emission Tomography,PET)图像的方法,提高分割速度和精度。方法:先对标准摄取值(Standard Uptake Values,SUV)值进行非线性化以增强图像,然后用迭代法二值化图像,最后用连通标记法来识别肿瘤。结果:该方法分割效果好,速度快。结论:和传统分割方法相比较,该方法具有智能性,分割精度高,更适合于肺部肿瘤。Objective To improve the segmentation precision and rate, a novel method for lung tumor PET image segmentation is presented. Method8 Firstly, the images were enhanced by non linear sealing of Standard Uptake Values. Secondly, gray images were converted into the binary images by using iterative. Finally, tumors were identified by using the connected component analysis. Results The segmentation precision and speed were improved by using this method. Conclusion Compared with used method, the proposed method is more intelligent and has a good process result which is fit to segment lung cancer.[Chinese Medical Equipment Journal,2008,29（11 ）：8-10]

关 键 词：肺肿瘤图像分割 智能识别 SUV值 连通标记 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]