基于3D区域增长法和改进的凸包算法相结合的全肺分割方法

肺实质分割结果的准确性在实际临床应用中具有非常重要的意义。但由于肺结节的位置、大小、形状的不规则性，肺部病变的多样性，以及人体胸部解剖结构的明显差异等，使得各类分割方法不能统一地适用于所有的胸部CT图像，所以对于肺实质分割方法的研究仍具有很大的挑战。该文在国内外研究分析的基础上提出基于3D区域增长法与改进的凸包修补算法相结合的全肺分割方法。在3D区域增长法的粗分割基础上，对分割的结果进行细化工作，通过连通域标记法与形态学方法相结合去除气管和主支气管，得到初步的肺实质掩膜，最后应用改进的凸包算法对肺部轮廓进行修补平滑，最终得到肺部分割结果。通过与凸包算法及滚球法相对比，证明该文所提改进的凸包算法能够有效地修补肺部轮廓凹陷，修补后的结果分割精度较高。

Lung Segmentation Method Based on 3D Region Growing Method and Improved Convex Hull Algorithm

The accuracy of lung segmented results is important in actual clinical application. However, all kinds of segmentation methods can not be uniform for all the chest CT (Computed Tomography) images because of the irregularities and diversity of lung disease, as well as significant differences in the anatomy of the human chest. Lung parenchyma segmentation studies still have a great challenge. Based on the analysis of domestic and international research, a new lung segmentation method is presented by combining with 3D region growing method and improved convex hull patching algorithm. Firstly, the 3D region growing method is adopted for the rough segmentation of lung CT images. Then the refining work is done to the segmented results. The connected domain labeling and morphological methods are used to remove the trachea and main bronchi to get the pulmonary parenchyma mask. The improved convex hull algorithm is presented to repair and smooth the concavities of lung contour. Finally, the segmented results can be gotten. The improved convex hull algorithm can repair the concavities of lung contour effectively in comparison with the convex hull algorithm and the rolling ball method, and the segmentation precision of results is very high after repairing.