A Modification of the Level Set Speed Function to Bridge Gaps in Data
Level set methods have become very popular means for image segmentation in recent years. But due to the data-driven nature of this methods it is difficult to segment objects that appear unconnected within the data. We propose a modification of the level set speed function to add a “bridging force” that allows the level set to leap over gaps in the data and segment an object despite artifacts or partial occlusions. We propose two methods to define such a force, one model-based and one image-based. Both versions have been applied to a series of test images, as well as medical data and photographic images to show their adequacy for image segmentation.
KeywordsImage Segmentation Test Image Curvature Term Speed Function Geodesic Active Contour
Unable to display preview. Download preview PDF.
- 1.Abu-Gharbieh, R., Gustavsson, T., Kaminski, C., Hamarneh, G.: Flame Front Matching and Tracking in PLIF Images using Geodesic Paths and Level Sets. In: IEEE Int. Conf. on Comp. Vis., Workshop on Variational and Level Set Methods in Comp. Vis., pp. 112–118 (2001)Google Scholar
- 3.Deschamps, T.: Curve and Shape Extraction with Minimal Path and Level-Set techniques. Applications to 3D Medical Imaging. Ph.D. thesis. L’Univerité de Paris-Dauphine (2001)Google Scholar
- 7.Malladi, R., Sethian, J.A., Vemuri, B.: Shape Modelling with Front Propagation: A Level Set Approach. IEEE Trans. on PAMI 17(2), 158–175 (1995)Google Scholar
- 11.Paragios, N., Deriche, R.: Geodesic Active Contours and Level Sets for the Detection and Tracking of Moving Objects. IEEE Trans. on PAMI 22(3), 266–280 (2000)Google Scholar
- 12.Rink, K., Törsel, A.M., Tönnies, K.: Segmentation of the Vascular Tree in CT Data using Implicit Active Contours. In: Bildverarbeitung für die Medizin, pp. 136–140 (2006)Google Scholar
- 16.Whitaker, R.T.: Isosurfaces and Level-Set Surface Models. Technical Report UUCS-02-010. School of Computing, Univ. of Utah (2002)Google Scholar