A hybrid method for the segmentation of large vessels in MRI data

Cardiovascular diseases are the leading cause of death in the western industrial nations. The disorder doesn’t only concern the heart itself but the blood vessels as well. The segmentation of blood vessels is an important preprocessing step for various computations and modeling, which play a crucial role in surgery planning and the diagnosis of cardiovascular disorders.


The scope of the research work is the development of adequate segmentation techniques of cardiac vessels, such as the aorta or the pulmonary vessels, and their evaluation based on image data of children with congenital vascular or cardiac disorders. The research is based on so-called tri-directional 4D-flow measurements acquired by use of the phase contrast MRI. The data contains the morphological image including the voxel intensity values as well as the three-dimensional flow vectors representing the velocity and the direction of the voxels. The objective is the development of a mostly automatic segmentation process, that facilitates and accelerates the work of the physician.


The research work proceeds in cooperation with the section of pediatric radiology of the university hospital Heidelberg.