IMAGE REQUANTIZATION BY CARDINALITY DISTRIBUTION

Abstract

This article presents a new approach to image requantization, called cardinality distribution. The new method introduces dithering patterns as ordered dither does; however, it provides a smoothing parameter p between 0 and 1. When p is chosen to be near 0.5, the new method can capture very sharp details of images. Therefore, the major deficiency of ordered dither can be avoided. Moreover, the reordering technique and its modified version are used to code bilevel and higher gray-level requantized images, respectively.