基于模糊速度函数的活动轮廓模型的肺结节分割

摘要

Pulmonary nodules are potential manifestation of lung cancer.In order to detect juxta-vascular pulmonary nodules and ground glass opacity pulmonary nodules in computer-aided diagnosis (CAD) system,the above two types of pulmonary nodules need to be accurately segmented.At present,the segmentation algorithm of pulmonary nodules using traditional active contour model may cause boundary leakage.In order to avoid this phenomenon,a new segmentation algorithm of pulmonary nodules using active contour model based on fuzzy speed function is proposed in this paper.First,the fuzzy membership degree in fuzzy speed function is calculated by using the fuzzy clustering algorithm,which uses gray feature and local shape index.Second,a fuzzy speed function is incorporated into the active contour model.At the boundary of pulmonary nodules,tbe fuzzy speed function equals zero and the evolution of the contour curve stops,so that the accurate segmentation of pulmonary nodules is completed.Experimental results show that the proposed algorithm can achieve accurate segmentation of juxta-vascular pulmonary nodules and ground glass opacity pulmonary nodules.%肺结节是肺癌在早期阶段的表现形式.利用计算机辅助诊断(Computer-aided diagnosis,CAD)技术对血管粘连型肺结节和磨玻璃型肺结节进行检测,需要对这两类肺结节进行准确的分割.目前基于传统活动轮廓模型的肺结节分割算法,存在边界泄露现象.对此,本文提出一种基于模糊速度函数的活动轮廓模型的肺结节分割算法.首先,采用结合灰度特征和局部形态特征的模糊聚类算法,计算模糊速度函数中的模糊隶属度；其次,将模糊速度函数引入到活动轮廓模型中,在肺结节的边界处,该速度函数为零,轮廓曲线停止演变,从而完成肺结节的分割.实验结果表明,本文提出的算法可以精确地分割血管粘连肺结节和磨玻璃型肺结节.