"Comparison of finite volume schemes for the mean curvature flow level set equation" . "3"^^ . "B35" . . . "RIMS Kokyuroku" . "0" . . "Handlovi\u010Dov\u00E1, A." . "21340" . "http://www.kurims.kyoto-u.ac.jp/~kenkyubu/bessatsu.html" . "14"^^ . "Comparison of finite volume schemes for the mean curvature flow level set equation"@en . . "JP - Japonsko" . . . "1880-2818" . "Oberhuber, Tom\u00E1\u0161" . "127925" . . "Comparison of finite volume schemes for the mean curvature flow level set equation" . . "RIV/68407700:21340/12:00202876!RIV13-MSM-21340___" . . "Mikula, K." . . . "RIV/68407700:21340/12:00202876" . "[F01C17785F5D]" . "S" . . "We discuss two different semi-implicit numerical schemes based on the finite volume method for approximation of the regularised mean curvature flow level set equation. The first, CVS scheme, is based on co-volume strategy and nonlinear terms, given by absolute value of gradient, are evaluated on pixel sides using splitted diamond-cell approach]. In the second, EHM scheme, the absolute values of gradients are evlauated inside the pixels by the Stokes formula and the scheme is obtained by imposing the continuity of fluxes on pixel sides. Results concerning numerical analysis of the schemes are presented and a comparison of these numerical approximations on several representative examples are discussed including performance in image filtering. On testing examples with exact solutions the schemes behave similarly in solution error, but the EHM scheme has higher precision in gradient error. Finite volume numerical schemes also perform better in the filtering of a strong salt & pepper noise as the results obtained using finite difference method." . "Comparison of finite volume schemes for the mean curvature flow level set equation"@en . "1"^^ . . . . "We discuss two different semi-implicit numerical schemes based on the finite volume method for approximation of the regularised mean curvature flow level set equation. The first, CVS scheme, is based on co-volume strategy and nonlinear terms, given by absolute value of gradient, are evaluated on pixel sides using splitted diamond-cell approach]. In the second, EHM scheme, the absolute values of gradients are evlauated inside the pixels by the Stokes formula and the scheme is obtained by imposing the continuity of fluxes on pixel sides. Results concerning numerical analysis of the schemes are presented and a comparison of these numerical approximations on several representative examples are discussed including performance in image filtering. On testing examples with exact solutions the schemes behave similarly in solution error, but the EHM scheme has higher precision in gradient error. Finite volume numerical schemes also perform better in the filtering of a strong salt & pepper noise as the results obtained using finite difference method."@en . . "Regularised mean curvature flow level set equation; stability and convergence of numerical solution; image filtering; finite volume method"@en .