Region-Oriented Segmentation of Vascular Structures from DSA Images Using Mathematical Morphology and Binary Region Growing

In this paper we describe a region-oriented segmentation algorithm suited for the detection and extraction of large blood vessels in angiographic images. The algorithm is based on the mathematical morphology top-hat operator followed by a binary region growing process. The combination of these two steps generates an image containing only the largest vessels regions, allowing a fully automatic segmentation without any manual interaction. We demonstrate the advantages of our approach comparing its results with those of other classical and morphological algorithms.