Fully automated segmentation of carotid and vertebral arteries from CTA

摘要: We propose a method for segmenting and labeling the main head neck vessels (common, internal, external carotid vertebral arteries) from contrast enhanced computed tomography angiography (CTA) volume. First, an initial centerline of each vessel is extracted. Next, are segmented using 3D active objects initialized first step. Finally, true identified by smoothly deforming it away mask edges spline-snake. focus particularly on extraction technique. It uses locally adaptive front propagation algorithm that attempts to find optimal path connecting ends vessel, typically lowest image scan Circle Willis in brain. patient adapted anatomical model different both initialize constrain this fast marching, thus eliminating need manual selection seed points. The evaluated applying datasets MICCAI 2009 bifurcation challenge, which paper only fully automatic submitted.