Title :
Segmenting Dot Patterns by Voronoi Diagram Concavity
Author_Institution :
Department of Mathematics and Computer Science, James Madison University, Harrisonburg, VA 22807.
Abstract :
This correspondence defines a signed distance, called ``internal concavity,´´ on paths of the Voronoi diagram of a dot pattern. An algorithm using internal concavity to segment dot patterns is described. The segmentation algorithm produces subsets of the Dirichlet tessellation (Delaunay triangulation) of the dot pattern.
Keywords :
Approximation algorithms; Clustering algorithms; Computer graphics; Computer science; Image processing; Image segmentation; Joining processes; Mathematics; Nearest neighbor searches; Pattern recognition; Blum transform; Delaunay; Dirichlet; Voronoi diagram; clustering; dot patterns; internal concavity; segmentation; spread angle;
Journal_Title :
Pattern Analysis and Machine Intelligence, IEEE Transactions on
DOI :
10.1109/TPAMI.1983.4767353
