Auto-segmentation of pathological lung parenchyma based on region growing method

Abstract

Lung parenchyma extraction is a precursor to the diagnosis and analysis of lung diseases. In this study, we propose a fully automated lung segmentation method that is able to extract lung parenchyma from both normal and pathological lung. First, we adapt the threshold algorithm to perform image binary, and then utilize the connected domain labeling method to select seed for region growing segmentation method which will be performed next. Then region growing image segmentation method is adopted and a rudimentary lung volume is established. A further refinement is performed to include the areas that might have been missed during the segmentation by an improved convex hull algorithm. We evaluated the accuracy and efficiency of the proposed method on 10 3D-CT scan sets. The results show that the improved convex hull algorithm can repair the concavities of lung contour effectively and the proposed segmentation method can extract the lung parenchyma precisely.