Robust Vessel Path Tracking in Imperfect Volumetric Data
Zabih R, Wang Y, Codella N
In this work we present a fast and efficient algorithm that can determine the path of an artery of any length, visualized in 3D angiography data with low SNR and loss of contrast along one side of the artery from neighboring structures, while relying on minimal user input. Arterial dimensions can be simultaneously recorded during program execution. The required user input consists of two points that define the initial arterial axial vector, and a tracking length limit. A fixed number of other inputs controlling edge detection and scan resolution also exist, but are optional.