Extension of Deformable Models: Hybrid Approaches for Analysis of Medical Images
Sébastien Ourselin, Rongxin Li · 2005
Despite deformable models' wide applicability in medical image segmentation, they are sometimes not adequate by themselves to achieve a desired degree of automation, a guaranteed level of accuracy, or an outcome beyond segmentation. In such situations, other methods may be combined with a deformable model to extend its abilities, overcome its shortcomings or increase the assurance of the segmentation accuracy. As examples of such combinations, we discuss the use of deformable models in hybrid approaches, to obtain curvilinear models of tubular structures, patient specific models of the abdominal aortic vessel system, and tissue models for radiation simulation.