Clinicians cannot easily visualize 3D intracerebral vascular anatomy in adequate detail. We present methods of creating a detailed, 3D vascular map from an MRA and a pair of x-ray angiograms. The MRA provides a 3D reconstruction base. Core-based methods segment the datasets and provide 3D-2D registration. Angiographic data not included in the MRA are automatically reconstructed into 3D, using methods capable of dealing with viewplane error. We provide algorithms for automatically associating 2D curves and test our methods in phantom data.
Elizabeth Bullitt, Alan Liu, Stephen R. Aylward, S