The automated extraction of the airway tree from 3-D chest CT data can greatly reduce the workload of physicians during diagnosis (e.g. quantification of airway morphology) and treatment (computer-aided bronchoscopy) of lung disease. This paper presents a method to automatically extract the airways driven by a sharpening filter, which enhances the branch edges in the input image based on the Laplacian of Gaussian, and adaptive cuboidal volumes of interest that an adaptive region growing algorithm uses to trace the airway tree. The method was trained on 20 data sets and evaluated on another 20 data sets from various scanners, using a wide range of acquisition and reconstruction parameters, including low dose scans. Compared to other state-of-the-art methods, our algorithm features the highest detection and extraction rates of bronchial branches. Future research needs to focus on the development of a method for automatic leakage detection.     «

The automated extraction of the airway tree from 3-D chest CT data can greatly reduce the workload of physicians during diagnosis (e.g. quantification of airway morphology) and treatment (computer-aided bronchoscopy) of lung disease. This paper presents a method to automatically extract the airways driven by a sharpening filter, which enhances the branch edges in the input image based on the Laplacian of Gaussian, and adaptive cuboidal volumes of interest that an adaptive region growing algorith...     »