The relationship between information prioritization and visual distinctness in two progressive image transmission schemes

摘要

The goal of this paper is to analyze the relation between the visual distinctness and the priority assigned to the information in two different progressive transmission schemes: the Set Partitioning in Hierarchical Trees (SPIHT) and the Rational Embedded Wavelet Image Coding (REWIC). SPIHT assigns priority based on the mean squared-error (MSE) distortion measure. REWIC quantifies the visual distinctness through the expected increase in utility per coding bit, using a simple function of the Kullback–Leibler's information gain, called distREWIC. We will show that in progressive transmission, the information selected through the increase in utility per coding bit corresponds to information with the largest visual relevance to a higher degree than the information selected on the basis of the MSE metric.