Gradient Vector Flow Snakes

Studien-/Diplomarbeit

In medical image processing and analysis, the automatic segmentation of regions of interest plays an important role. For the detection of object boundaries, traditionally snakes, or active contours are used. However, using traditional snake models introduces problems such as curve initialization and poor convergence to boundary concavities. Hence, a more sophisticated approach is desired to achieve high quality object contours.
Gradient Vector Flow(GVF) is a new external force for snake models introduced by Xu et. al.[[1]iacl.ece.jhu.edu/projects/gvf/ http://iacl.ece.jhu.edu/projects/gvf/ ]. The GVF differs fundamentally from traditional external snake forces, showing a very large capture range and the ability to move into boundary concavities.
In this thesis, a GVF snake model shall be implemented for contrast enhanced cardiac CT data for the automatic analysis of coronary arteries. Based on a vessel centerline tree, GVF snakes are supposed to find high quality coronary artery contours upon which measurements, such as lumen area and diameter, can be performed. These measurements then can be used to find vascular anomalies. Finally, the quality of the automatically generated segmentation has to be evaluated.

Advisor(s)