Let ${\displaystyle V}$  be the Voronoi diagram for a set of sites ${\displaystyle P}$ , and let ${\displaystyle V_{p}}$  be the Voronoi cell of ${\displaystyle V}$  corresponding to a site ${\displaystyle p\in P}$ . If ${\displaystyle V_{p}}$  is bounded, then its positive pole is the vertex of the boundary of ${\displaystyle V_{p}}$  that has maximal distance to the point ${\displaystyle p}$ . If the cell is unbounded, then a positive pole is not defined.

Furthermore, let ${\displaystyle {\bar {u}}}$  be the vector from ${\displaystyle p}$  to the positive pole, or, if the cell is unbounded, let ${\displaystyle {\bar {u}}}$  be a vector in the average direction of all unbounded Voronoi edges of the cell. The negative pole is then the Voronoi vertex ${\displaystyle v}$  in ${\displaystyle V_{p}}$  with the largest distance to ${\displaystyle p}$  such that the vector ${\displaystyle {\bar {u}}}$  and the vector from ${\displaystyle p}$  to ${\displaystyle v}$  make an angle larger than ${\displaystyle {\tfrac {\pi }{2}}}$ .

• Boissonnat, Jean-Daniel (2007). Effective Computational Geometry for Curves and Surfaces. Berlin: Springer. ISBN 978-3-540-33258-9.