Fast mapping and morphing for genus-zero meshes with cross spherical parameterization

Abstract

We introduce an efficient approach to establish surface correspondences between genus-zero triangle meshes and animate a morph between them. We employ the concept of mesh simplification and refinement to progressively parameterize meshes using a novel kernel sampling algorithm. Our two-step feature alignment method optimizes the distribution of vertices in a parametric domain. A new mesh is generated after merging mesh topologies, and it represents the shapes of all input meshes and their transitions in-between. Our approach merges topological structures rather than mapping surface properties. Complex meshes with tens of thousands of vertices can be parameterized and merged in a few minutes. The advantages of our approach are demonstrated through 3D morphing applications. Comparing to existing approaches, our solution advances in performance while maintaining high-quality parameterization and morphing.