A hybrid approach for vessel enhancement and fast level set segmenatation based 3d blood vessel extraction using MR brain image

Abstract

In this research we present a robust prototyping method for segmentation of brain MR images to extract the 3d evolution model of the blood vessel present in the complex regions under the surface. We proposed a hybrid technology based on two levels, the smoothing and segmentation process for extraction of blood vessels. For this approach we compared robust automated algorithms for filtering the MR images. Furthermore, in second stage fast level set segmentation process was implemented to complete the extraction of blood vessels process with in a magnetic resonance (MR) image. Vessel extraction process was implemented in a virtual environment and used to convert complex vascular geometry of the selected MR region into a replica with large anatomical coverage and high spatial resolution. Experiments were conducted to evaluate the performance of the VED filters enhancing vessels in brain region and further used with fast level set segmentation to extract the vessel models.