Voronoi diagram and medial axis algorithm for planar domains with curved boundaries I. Theoretical foundations

In this first installment of a two-part paper, the underlying theory for an algorithm that computes the Voronoi diagram and medial axis of a planar domain bounded by free-form (polynomial or rational) curve segments is presented. An incremental approach to computing the Voronoi diagram is used, wherein a single boundary segment is added to an existing boundary-segment set at each step. The introduction of each new segment entails modifying the Voronoi regions of the existing boundary segments, and constructing the Voronoi region of the new segment. We accomplish this by (i) computing the bisector of the new segment with each of the current boundary segments; (ii) updating the Voronoi regions of the current boundary segments by partitioning them with these bisectors; and (iii) constructing the Voronoi region of the new segment as a union of regions obtained from the partitioning in (ii). When all boundary segments are included, and their Voronoi regions have been constructed, the Voronoi diagram of the boundary is obtained as the union of the Voronoi polygons for each boundary segment. To construct the medial axis of a planar domain, we first compute the Voronoi diagram of its boundary. The medial axis is then obtained from the Voronoi diagram by (i) removing certain edges of the Voronoi diagram that do not belong to the medial axis, and (ii) adding certain edges that do belong to the medial axis but are absent from the Voronoi diagram; unambiguous characterizations for edges in both these categories are given. Details of algorithms based on this theory are deferred to the second installment of this two-part paper.