An Efficiency of Block-based Data Array Representations

摘要

The efficiency performance of block-based representations of n-dimensional data arrays is evaluated when the arrays are given by independent and identically distributed m-ary random variables. Two repre-sentations are investigated, which correspond to arrays that decompose into the uniformly valued cubic blocks to be induced by the scheme of n-dimensional runs and 2n-tree, respectively. The evaluated performance is the information theoretic representation rate, which is defined by the ratio of average number of representing blocks within an array to number of array elements.