Adaptive Model-Based Segmentation of Human Vessels from 3D MRA and CTA Data

We introduce an adaptive model fitting approach for the segmentation of human vessels from 3D images. The shape and size of the region-of-interest (ROI) used for model fitting are automatically adapted to the local width, curvature, and orientation of the vessel to increase the robustness and accuracy of model fitting. In conjunction with our previously proposed cylindrical model, the new adaptive approach has been successfully applied to segment vessels from 3D MRA and CTA images. Our experiments show that the adaptive approach yields superior segmentation results compared to approaches based on a fixed size ROI.