A divide-and-conquer approach for automatic polycube map construction
Polycube map is a global cross-surface parameterization technique, where the polycube shape can roughly approximate the geometry of modeled objects while retaining the same topology. The large variation of shape geometry and its complex topological type in real-world applications make it difficult to effectively construct a high-quality polycube that can serve as a good global parametric domain for a given object. In practice, existing polycube map construction algorithms typically require a large amount of user interaction for either pre-constructing the polycubes with great care or interactively specifying the geometric constraints to arrive at the user-satisfied maps. Hence, it is tedious and labor intensive to construct polycube maps for surfaces of complicated geometry and topology. This paper aims to develop an effective method to construct polycube maps for surfaces with complicated topology and geometry. Using our method, users can simply specify how close the target polycube mimics a given shape in a quantitative way. Our algorithm can both construct a similar polycube of high geometric fidelity and compute a high-quality polycube map in an automatic fashion. In addition, our method is theoretically guaranteed to output a one-to-one map. To demonstrate the efficacy of our method, we apply the automatically-constructed polycube maps in a number of computer graphics applications, such as seamless texture tiling, T-spline construction, and quadrilateral mesh generation.
Shape modeling Polycube map Discrete Ricci flow Uniform flat metric Computational geometry Object modeling Geometric algorithms
Ying He Hongyu Wang Chi-Wing Fu Hong Qin
School of Computer Engineering, Nanyang Technological University, SO Nanyang Avenue, Blk N4, Singapo Computer Science Department, Stony Brook University, NY, USA
国际会议
IEEE International Conference on Shape Modeling and Applications (SMI)(2009年形状建模国际会议)
北京
英文
369-380
2009-06-26（万方平台首次上网日期，不代表论文的发表时间）