Tree and Pyramid Structures for Coding Hexagonally Sampled Binary Images

摘要

Tree and pyramid type data structures are described which may be used for storing and processing binary images that have been sampled on a hexagonal grid. These are analogous to the quadtrees and pyramids which have recently been developed for images sampled on a rectangular grid. Trees may be formed with 3, 4, 7, 9 ... branches per node. Of these the 'septtree', formed with 7 branches per node, promises to be particularly interesting since each node 'covers' a roughly hexagonal region of the image. (Author)