A spectral characterization of the Delaunay triangulation

  • Authors:
  • Renjie Chen;Yin Xu;Craig Gotsman;Ligang Liu

  • Affiliations:
  • Dept. of Mathematics, Zhejiang University, Hangzhou 310027, China;Dept. of Mathematics, Zhejiang University, Hangzhou 310027, China;Computer Science Faculty, Technion -- Israel Institute of Technology, Haifa 32000, Israel;Dept. of Mathematics, Zhejiang University, Hangzhou 310027, China

  • Venue:
  • Computer Aided Geometric Design
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

The Delaunay triangulation of a planar point set is a fundamental construct in computational geometry. A simple algorithm to generate it is based on flips of diagonal edges in convex quads. We characterize the effect of a single edge flip in a triangulation on the geometric Laplacian of the triangulation, which leads to a simpler and shorter proof of a theorem of Rippa that the Dirichlet energy of any piecewise-linear scalar function on a triangulation obtains its minimum on the Delaunay triangulation. Using Rippa's theorem, we provide a spectral characterization of the Delaunay triangulation, namely that the spectrum of the geometric Laplacian is minimized on this triangulation. This spectral theorem then leads to a simpler proof of a theorem of Musin that the harmonic index also obtains its minimum on the Delaunay triangulation.