基于EM和Mean-shift的肺结节分割

针对结节与血管相连且两者CT值相近造成分割困难的问题,提出了一种基于期望最大(estimationmaximum,EM)的自适应带宽参数选择的方法,并采用均值漂移(Mean-shift)算法解决结节分割.与基于统计分析规则的带宽选择方法和基于最优化的带宽选择方法相比,该方法能直接求得正确带宽参数,且时间复杂度低.应用血管梯度的法向量方向服从正态分布,而结节梯度的法向量方向服从均匀分布,建立血管粘连型结节模型,并用期望最大估计模型参数,根据均匀分布的权重和带宽选择定理确定带宽参数.该方法对仿真数据和CT数据(19个粘连血管性肺结节)进行评估实验,都取得了正确的分割结果.结果表明,该方法对分割粘连血管型结节是有效的.

Pulmonary Nodule Segmentation Based on EM and Mean-shift

Aiming at solving the segmentation problem caused by the connection of lung nodule and vessel, a new adaptive bandwidth selection method based on EM is proposed and we apply it into nodule segmentation. Compared it to the method of bandwidth chosen based on statistical analysis rule or optimized rule, it has some advantages such as low time complexity and correct bandwidth in accord with a real problem.The vertical orientation vectors of vessels gradient was constructed as to normal distribution and the vertical orientation vectors of nodules gradient as uniform distribution, we modeled the nodule connected vessel, estimated model parameters by EM, and extracted bandwidth values in Mean-shift based on the weight of uniform distribution and bandwidth selection theorem. The proposed method was tested on synthetic data and the clinical chest CT volumes, and all the results were correct. The results revealed that the proposed method is successful in segmentation lung nodules connected vessel.