Parallel Shortest Path Algorithm for Voronoi Diagrams with Generalized Distance Functions

Voronoi diagrams are fundamental data structures in computational geometry with applications on different areas. Recent soft object simulation algorithms for real time physics engines require the computation of Voronoi diagrams over 3D images with non-Euclidean distances. In this case, the computation must be performed over a graph, where the edges encode the required distance information. But excessive computation time of Voronoi diagrams prevent more sophisticated deformations that require interactive topological changes, such as cutting or stitching used in virtual surgery simulations. The major bottleneck in the Voronoi computation in this case is a shortest-path algorithm that must be computed multiple times during the deformation. In this paper, we tackle this problem by proposing a GPU algorithm of the shortest-path algorithm from multiple sources using generalized distance functions. Our algorithm was designed to leverage the grid-based nature of the underlying graph used in the simulation. Experimental results report speed-ups up to 65× over a current reference sequential method.