Vessel extraction from breast MRI

We present an extension of the previous work, where a multi-seed region growing (MSRG) algorithm was shown, that extracts segments from breast MRI. The algorithm of our extended work filters elongated segments from the segments derived by the MSRG algorithm to obtain vessel-like structures. This filter is a skeletonization-like algorithm that collects useful information about the segments’ thickness, length, etc. A model is shown that scans through the solution space of the MSRG algorithm by adjusting its parameters and by providing shape information for the filter. We further elaborate on the usefulness of the algorithm to assist medical experts in their diagnosis of diseases relevant to angiography.