Aorta Segmentation and Blood Flow Obtained from Phased Contrast MRI Based on an Active Contours Approach

Abstract

In this paper, a segmentation method based on an active contours algorithm capable to detect not well-defined edges in 2D images, based on Mumford–Shah functional for segmentation and level sets, is used for the detection of the aortic arch anatomy in a Phase-Contrast Magnetic Resonance Imaging (PCMRI) data. The segmentation strategy followed requires an initial contour extracted from the central image of the aortic arch, which is obtained from merging two initial structures; then, the initial contour is propagated to the adjacent images in order to segment all the images with aorta information. The results found suggest that the proposed approach represents a robust alternative for detection of the aorta anatomy in PCMRI data based on the 3D reconstruction from 2D segmentations.