模M多维差分变换和数据压缩

MODULO-M MULTI-DIMENSIONAL DIFFERENTIAL TRANSFORMATION AND DATA COMPRESSION

本文提出了模M多维差分变换的图象表示和压缩方法.变换是在对N维图象数组的2~N元分割所产生的2~N元树上进行,其效果是利用了图象的N维数据相关性,使得图象灰度值分布得到了显著的改善,它既适合于二值图象,也适合于多值图象,对于线图、文本以及具有块状或体状的图象或数组尤为有效,变换算法的复杂度非常简单,并且是递归可执行的。

A method of representing and compiessing binary-valued or multi- valued images or arrays is developed. It is a promising technique of data compression for images such as texts, documents, charts or any other images or arrays composed of blocks or cubes. The method is based on the modulo-M differential transformation of 2N-ary tree of N-dimensional images or arrays. The algorithm is very simple and recursively executable. The experimental results of the method with DF-expression are also described in this paper.