Chinese Journal of Computers   Full Text
Title	3D Model Remeshing Using Circular Parameter Domain and Importance Sampling
Authors	FU Yan ZHU Xiao-Ming ZHOU Bing-Feng
Address	(Institute of Computer Science and Technology, Peking University, Beijing 100871)
Year	2007
Issue	No.12(2124—2131)
Abstract &   Background	Abstract A 3D model remeshing algorithm based on circular parameter domain and importance sampling is proposed in this paper. The algorithm automatically searches a cut path to cut the model into a disk-like surface and then parameterizes it into a circular domain. Based on the curvature of the mesh and the stretch introduced by the parameterization, a control map is defined to control the sampling distribution. The control map is then sampled by a hierarchical blue noise important sampling. The generated sample points are triangulated and mapped to 3D space. As a result, we obtain a high-quality adaptive mesh whose vertex distribution is determined by the density of the control map. keywords remeshing; importance sampling; mesh parameterization background This research is supported by National Natural Science Foundation of China under grant No.60573149. Triangle meshes generated by current 3D acquisition techniques are usually dense and irregular, which brings a lot of difficulty to further process such as storage, transmission and display. Due to the restriction of capacity and capability of the computer, triangular meshes are expected to be represented by fewer vertices without the loss of accuracy. This can be achieved by mesh simplification followed by remeshing. This work focuses on the process of remeshing. Most of the existing techniques remesh the surface by regular sampling, which is inflexible to control the distribution of the vertices. Others use local mesh optimization techniques to improve the mesh quality, which is relatively slow. By adjusting the distribution and location of the vertices, we can remesh the original mesh into a high-quality one. This paper combines the techniques of mesh parameterization and 2D importance sampling to remesh 3D triangular models. By revising the boundary of parameterization domain, the stretch introduced by the parameterization is reduced. The technique of importance sampling is adopted to control the distribution of the sampling vertices. As shown by the experimental results, the algorithm effectively improves the quality of 3D meshes, and highly-regular meshes can be obtained after remeshing process.