Parallel Polygonization of Implicit Surfaces

In this paper, we present an effective parallel algorithm for polygonizing implicit surfaces. The algorithm can produces triangulation approximation of implicit surface with adaptive size of triangles according to the local curvature of the surface. Our algorithm first divides an given implicit surface into eight parts by subdividing its bounding box. Then, the algorithm finds a seed point for each part on the surface and builds a seed mesh centered at the point. The seed mesh is the expanded by using the wave-front method and the surface-tracking approach. The local approximations of eight parts can be constructed in parallel way. Our algorithm finally merges the local meshes to a global approximation of the implicit surface. Resulted triangulations contain triangles with various sizes, i.e. smaller triangles for high curvature areas and larger triangles for flat areas, on the surface. Compared with the existing polygonization methods, our method has more fast speed and can guarantee quality of resulted meshes very well. Our method can be used in the real-time environment for rapid visualization of implicit surfaces.