Shapelets在低剂量CT影像中肺结节检测的应用研究

Shapelets application research of lung nodule detection in low-dose CT image

针对医学影像诊断领域中肺结节检测比较困难的问题,提出了一种新的线性检测肺结节的方法。该方法以Shapelets理论为基础,利用多个二维加权基函数来表示图像中的目标物体。本文通过计算出每个基函数的系数,选取适当的特征尺度参数,使用带权基函数叠加的方法检测出肺结节的准确位置。实验结果表明,该方法能够快速准确的检测出低剂量肺部CT图像中肺结节的具体位置,对医学临床诊断提供了有力的支持。

Lung nodule detection is a difficult and complex problem in the field of medical diagnostic ima- ging, and this paper presents a new method to detect lung nodule. This method is based on Shapelets theo- ry, and the aimed object is represented with a series of weighted basic function. In this paper, through cal- culating the Shapelets coefficient and selecting appropriate characteristic scale parameter, we detect the exact location with superposition of weighted basic function. Experiment result shows that this method can detect the exact location of the lung nodules in low-dose CT lung image, contributing to clinical diagnosis in the medical field.