CT影像具有空间分辨率高的优点,是肺部疾病影像学诊断的首选方式。肺部病灶的检测和测量、肺功能的定量分析均需要精确的肺组织分割。为解决CT影像由于噪声、伪影、部分容积效应等干扰而导致的肺部各组织之间灰度交叠、边界模糊、难以分离的问题,系统地综述了针对肺部各个分割对象的有效解决方法。从肺实质分割、肺血管分割、肺气道分割、肺叶分割、肺结节分割以及肺部病变组织的分割等方面,详细分析了面临的挑战性问题和当前研究进展,并阐述了肺组织分割方法的发展趋势。
With the advantages of high spatial resolution, CT imaging is the preferred method for imaging diagnosis of pulmonary diseases. The detection and measurement of pulmonary lesion, and the pulmonary functional quantitative analysis need accurate segmentation of lung tissues. To solve the problem of the gray overlap and the blurred boundary, lung tissues were difficult to be separated due to noise, artifact and partial volume effect on the CT image, this paper summarized the effective solution to the various segmentation objects of the lung. From the pulmonary parenchyma segmentation, pulmonary vascular segmentation, pulmonary airways segmentation and pulmonary lobe segmentation, pulmonary nodules segmentation and pulmo- nary lesions segmentation, etc. , this paper analyzed the challenging problems and current research progress, and expounded the development trend of the lung tissues segmentation method.