Maintaining 2D Delaunay triangulations on the GPU for proximity queries of moving points
Abstract
In this paper we explore the problem of maintaining the Delaunay triangulation of moving 2D points on the GPU with discrete time steps, using only local transformations. We show that our Delaunay triangulation structure is efficient at answering proximity queries, such as closest neighbor or fixed-radius nearest neighbors problems. We also characterize the difficulties of updating the triangulation, and also the cases where it is possible to do it only through local operations.
| Origin | Files produced by the author(s) |
|---|