Color image segmentation by a watershed algorithm and region adjacency graph processing

Abstract

A new method for gray-scale, color and multispectral image segmentation is proposed. The method is based on a morphological split-and-merge fast watershed algorithm and region adjacency graph processing. It features partly iterative processing with relatively low computational load. Two system alternatives are presented and their application to color image segmentation is discussed. First computer simulations show correctness of operation and encourage for further research.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>