基于金字塔模型和Mean Shift算法的门静脉和主动脉分割方法

摘要

目的 基于目前临床在绘制肝门静脉和主动脉的时间-密度曲线方面存在的问题,提出一种准确分割肝脏CT灌注成像(CTPI)图中肝脏门静脉和主动脉的方法.方法 采用金字塔模型,结合Mean Shift分割算法对肝脏CTPI图像中的门静脉和主动脉进行分割,并在此基础上计算时间一密度曲线.结果 此方法能实现对肝脏CTPI图像中门静脉和主动脉的有效分割,绘制出准确、平滑而无毛刺的门静脉和主动脉的时间-密度曲线.结论 此方法有助于临床客观、准确地评估肝功能和诊断病变.%Objective To develop a method for accurate segmentation of the portal vein and aorta parts in CT perfusion imaging (CTPI) of liver, and to solve problems existing in drawing of clinic time-density curve of the portal vein and aorta.Methods Combining the pyramid model and the Mean Shift algorithm to segment the portal vein and aorta from the liver CT PI, the time-density curve (TDC) of portal vein and aorta were drawn based on the obtained segmentation results.Results The experiment results showed that the portal vein and aorta were segmented correctly from liver CT perfusion images with the presented method.TDC of the portal vein and aorta were drawn accurately and smoothly without burrs.Conclusion The presented method is of great sense and application value for accurate evaluation on liver function and diagnosis of liver pathology.