Morphing Based on Mended Sparse-Field Algorithm of Level-Set Method

Abstract

Level-Set method is a good way to do metamorphosis. Narrowband and sparse-field algorithms improve its performance. Results of narrowband morphing are smoother when the sparse-field algorithm is faster. Sparse-field algorithm is mended to fit Euclidean distance model and the narrowband algorithm is used to make up the error of the sparse-field morphing. Topological relationship replaces distance band to define layer sets, and a single side active set is proposed to improve efficiency and robustness. To make up the error of the sparse-field algorithm, which causes obvious alias at the last half stage, two remedy methods are proposed. Averaging and translation method is simpler and more efficient. Narrowband evolution and back method is better for reserving the sharp shape.