Representing Multidimensional Medical Image in Compact Format

In this paper, we propose a novel approach for effectively expressing multidimensional medical image data in spatial database systems with a compact form, which uses the vector compressed from position vector. Here hyperplane are used to express the position and forms of a various dimensions spatial object. A set of hyperplanes splits a space into convex polytopes, each having a position vector [+...-0-...--], which is a sequence of positions with the set of hyperplanes. Position vectors can be used to evaluate the spatial database operations, intersection, union and difference between two or more objects in any dimension. For easy compression, we firstly propose the cell position vector (CPV) representation of a position vector. The CPV compression algorithm is based on following ideas: only the position numbers of `0´s in a position vector of zero-dimensional cell, and the position numbers of `+´s, `0´s, `-´s in a position vector of a top-cell are saved. We introduce a modified run length algorithm to compress CPVs of top-cells. Experimental evidence shows that our model has only 5% data size of position vector model for a three dimensional heart test data. We also have implemented this storage method on an object database system based on a compact form of CPV