Rapid tracking of vascular tree in angiography images based on adaptive sampling

Efficient assessment of vascular structures plays a significant role in many medical procedures. We present a practical approach to segmentation of vascular tree in angiography images. It is implemented in an interactive 3D visualization-assisted system and consists of the following main steps. First, angiography image is filtered to enhance vessels and eliminate irrelevant structures. Second, the centerline and thickness of the vessel are extracted utilizes a novel local tracking algorithm named adaptive sampling. The sampling starts from a single seed point and marches recursively forward along possible vessel continuations to capture whole tree-like structure, in which the branch detection and thickness estimation are developed in a consistent framework. Finally, the result could be proofread in a cooperative environment. We validated and evaluated the approach using synthetic data and real images from clinical. The results showed that the system achieves a reasonable balance between fast speed and high accuracy.