MULTIPROCESSOR PYRAMID ARCHITECTURES FOR BOTTOM-UP IMAGE ANALYSIS.

Computing properties of digitized images takes considerable computer power because of the large data volume involved. In certain cases, however, arranging multiple processors hierarchically to exploit both parallel and pipelined processing methods permits faster image analysis. Various pyramid approaches as well as the quadtree image representation are considered. To obtain its quadtree, an image is overlaid with a sequence of increasingly fine tesselations that define a recursive embedding of quadrants and thus a hierarchy over image windows. The hierarchy is described by a tree whose root node is associated with the entire image. Each node in the tree represents a square window containing four quadrants. Each child node is associated with a quadrant of the parent window. Leaf nodes correspond to windows of the smallest size, usually single pixels; thus they do not contain quadrants.