Segmentation of the Airway Tree from Chest CT using Local Volume of Interest

Lung diseases such as COPD and asthma affect airway morphology. Automated segmentation is an essential first step toward the analysis of airways. We propose a fully-automated algorithm to segment the airway tree from chest CT scans. The proposed algorithm requires no manual intervention and uses a 3D region growing based method and allows for accurate detection of leakage by growing regions within a locally-defined envelope. The algorithm is run within a tree segmentation framework which breaks down the problem into the segmentations of individual branches. The method was evaluated using 20 chest CT scans from EXACT09 challenge. The provided scans were taken from the patients with various health conditions. The results show that the algorithm is able to segment the airway tree while keeping the leakage level low, as only 0.11% of the segmentation was classified as false positive and the average number of leakage was less than 1 per patient.