基于GPU的高质量隐式曲面四边形化

GPU-Based High Quality Quadrangulation for Implicit Surfaces

为了从隐式曲面快速抽取高质量的四边形网格,提出一种基于GPU的高质量隐式曲面四边形化方法.该方法分为初始网格抽取和网格优化2个阶段.利用GPU的并行性能,首先快速抽取一个粗糙的四边形网格,然后对该网格的几何(顶点位置、法向)和规整性(顶点分布、邻接关系)两方面进行迭代优化.实验结果表明,文中方法极大地提高了隐式曲面四边形化的效率,并且能得到高质量的四边形网格.

In order to efficiently generate high-quality quad meshes from implicit surfaces, this paper introduces a GPU-based high-quality quadrangulation method for implicit surfaces. Our method contains two stages: an ex-traction stage and an optimization stage. We first extract a coarse quad mesh from an input implicit surface, and then optimize the geometry (vertex position and normal) and regularity (vertex distribution and connection) of the quad mesh. Experiments show that our method can greatly improve the performance of quadrangulation while achieving a quad mesh with low errors.