Duct tracking in 3D medical data

Abstract

Implementing duct tracking of bronchi and blood vessels in 3D medical data can significantly help in examining these body structures and reduce time needed for correct diagnosis. Nature of the problem involved may require some operator interaction when selecting duct branches of interest or correcting possible mistakes in duct path segmentation. Noise, badly discernible duct walls, branching or extreme duct curvatures, that can be found in computer tomography data, make the segmentation challenging. There are more ways to display segmented ducts, as overlay over original data or unrolled to planar tube, allowing easier measurements of length, width or observation of tube anomalies.