Automatic Extraction Algorithm for Complex Vascular Skeletons Based on TMOBB

Vascular diseases pose significant challenges in the realm of medical sciences, often giving rise to critical health issues and potential life-threatening conditions. Accurate extraction of vascular skeletons plays a pivotal role in furnishing vital information and support for the diagnosis, treatment, and prognostic management of vascular ailments. Computational image analysis assumes a crucial role in the diagnosis and quantification of vascular anomalies. Given the scale and complexity of data from modern vascular imaging acquisition, achieving rapid, automated, and precise vascular segmentation constitutes a challenging task.In this paper, we introduce an automatic propagation-based approach utilizing directed Oriented Bounding Boxes (OBB). This technique efficiently extracts the skeletal lines of individual vessels without necessitating prior segmentation. Leveraging raw 3D Computed Tomography Angiography (CTA) images as input, the method involves successive stages encompassing Directed Oriented Bounding Box computation, threshold segmentation, centroid coordinate determination, ultimately yielding a fitted skeletal representation of the vessels. Moreover, we propose a complex vascular skeleton extraction method based on Two Moves Oriention Bounding Box (TMOBB), which encompasses skeleton curve sampling, bifurcation point discrimination, and erroneous skeleton line rectification. Experimental outcomes underscore that the approach presented in this paper exhibits notable precision and elevated extraction velocity in the domain of vascular skeleton extraction, thereby furnishing robust support for hepatic vascular structure analysis.