A nonlocal mean curvature flow for multi-dimensional image processing
The author studies a problem which is formulated in the paper as follows: ``Suppose that a complicated, closed level surface is somehow extracted from given noisy image data at high resolution. Then we generally wish to smooth the complicated surface while preserving its geometric shape as much as possible. In the paper this problem is studied for \(n\)-dimensional hypersurfaces in \(\mathbb R^{n+1}\) with \(n \geq 1\) (i.e., the case of curves in the plane is also considered). Some energy functional is proposed for which the Euler-Lagrange equations lead to an anisotropic mean curvature equation with nonlocal terms. For this equation the author proves short-time existence and uniqueness of solutions assuming that the initial surface is sufficiently smooth. Some numerical experiments on this flow applying to curves are exposed.
- A nonlocal mean curvature flow and its semi-implicit time-discrete approximation
- Solution and application of anisotropic curvature driven evolution of curves (and surfaces)
- On mean curvature diffusion in nonlinear image filtering
- Geodesic Evolution Laws—A Level-Set Approach
- Geometric partial differential equations and image analysis
This page was built for publication: A nonlocal mean curvature flow for multi-dimensional image processing
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5697660)
