Topology Noise Removal for Curve and Surface Evolution
- 1.3k Downloads
In cortex surface segmentation, the extracted surface is required to have a particular topology, namely, a two-sphere. We present a new method for removing topology noise of a curve or surface within the level set framework, and thus produce a cortical surface with correct topology. We define a new energy term which quantifies topology noise. We then show how to minimize this term by computing its functional derivative with respect to the level set function. This method differs from existing methods in that it is inherently continuous and not digital; and in the way that our energy directly relates to the topology of the underlying curve or surface, versus existing knot-based measures which are related in a more indirect fashion. The proposed flow is validated empirically.
KeywordsActive Contour Homology Class Active Contour Model Topology Control Signed Distance Function
Unable to display preview. Download preview PDF.
- 7.Edelsbrunner, H., Morozov, D., Patel, A.: Quantifying transversality by measuring the robustness of intersections (2009) (preprint)Google Scholar
- 8.Freesurfer, http://surfer.nmr.mgh.harvard.edu/
- 12.Kichenassamy, S., Kumar, A., Olver, P.J., Tannenbaum, A., Yezzi, A.J.: Gradient flows and geometric active contour models. In: ICCV, pp. 810–815 (1995)Google Scholar
- 18.Zhang, Y.: The matlab toolbox for 2d/3d image segmentation using level-set based active contour/surface with aos scheme, http://ecson.org/resources/active_contour_segmentation.html