Computer-assisted identification and modelling of internal carotid arteries from 3D CE-MR angiograms for stenosis quantification

3D Geometry-Based Tracking and Segmentation of Carotid Arteries from CE-MR Angiograms Using Locally Adaptive Thresholdin

“…The data volume can be reduced by removing all information outside the neck, above the ear level and below the bottom of the neck, near the shoulder level. This is a novel technique previously developed by the authors [6]. The reduced data is then further reduced by dividing it into right and left halves and removing the sides of the head.…”

“…For the purpose of this study, the isosurface model is used to aid the selection of a global threshold, which will enable an initial coarse segmentation for the pre-processing stage of the analysis. The value of this threshold is chosen low (in relation to the mean intensity in the dataset), so that the entire vessel tree is visualised, without discontinuities in the stenotic parts of the artery, (which are known to have lower intensities compared to the non-stenotic parts [6]). The chosen threshold is then applied to the entire data, resulting in a binary volume.…”

“…A novel technique has been suggested for automatic identification of the ICA [6]. This technique is based on the physiology of the carotids, as seen in Figure 1.…”

“…In each image (x-z plane) of the reformatted axial scan, the ICA is present as an approximately circular object, from which the centroid (centre of gravity) point can be calculated. The arterial axis can be determined as the line connecting each ICA centroid point between successive images [6]. Having identified this central axis, image slices are taken perpendicular to it (i.e.…”

“…The ICA is identified automatically, by tracking the carotid bifurcation and selecting the artery branch with no further arterial branches [6]. An isosurface model is used to visualise the vasculature.…”