New region expansion for quadtrees

Abstract

A one-pass algorithm that performs region expansion in images represented by quadtrees is presented. The algorithm changes to black those white pixels within a specified distance of any black mode in the image. The algorithm yields a significant improvement over previous approaches by reducing both the number of black nodes that must be considered for expansion and the number of nodes that must be inserted as a result of the expansion. The reductions are achieved by introducing the concepts of a merging cluster and a vertex set. Empirical tests show that the execution time of this algorithm generally decreases as the radius of expansion increases, whereas in previous approaches the execution time generally increased with the radius of expansion.