Title :
3D segmentation and characterization of lower limb vessels in multi-slice computed tomography
Author :
Boldak, C. ; Toumoulin, C. ; Coatrieux, J.-L.
Author_Institution :
Fac. of Comput. Sci., Bialystok Tech. Univ., Poland
Abstract :
This paper describes a fast 3D extraction method for the lower limb vessels in multi-slice computed tomography. It represents an extension of our previous work and makes uses of a two-step process: the first one performs a vessel central axis tracking. It makes use of a semi-automatic 3D geometrical moment-based method to localize the center of the vessel and its local direction. It allows at the same time a first roughly delineation of the calcification areas and a diameter estimation at any point of the vessel. A refinement is then performed based on a level set approach to improve the delineation accuracy of the contour and calcifications. The vascular lumen is then calculated in the calcified area to evaluate the severity of the vessel narrowing.
Keywords :
blood vessels; computerised tomography; image segmentation; medical image processing; patient diagnosis; 3D extraction method; 3D segmentation; calcification areas; delineation; diameter estimation; lower limb vessels; multislice computed tomography; semiautomatic 3D geometrical moment-based method; vascular lumen; vessel central axis tracking; vessel narrowing; Aneurysm; Arteries; Atherosclerosis; Computed tomography; Computer science; Data analysis; Data mining; Diseases; Level set; Pathology;
Conference_Titel :
Engineering in Medicine and Biology Society, 2003. Proceedings of the 25th Annual International Conference of the IEEE
Print_ISBN :
0-7803-7789-3
DOI :
10.1109/IEMBS.2003.1279811