Simultaneous skeletonization and graph description of airway trees in 3 D CT images

The goal of this work is to generate a symbolic description of the airway trees from 3D CT images acquired at different ventilation conditions, in order to compare these trees to each other. The method applies to the binary result of a segmentation algorithm. Its originality resides in a comprehensive use of the Dijkstra’s algorithm, namely in a novel technique used to detect the end-points. This technique enables us to extract the skeleton and to detect the bifurcations in a single pass during backtracking from these points. Preliminary results show that the method proposed is well-suited for the intended application.