A New Le Gall Wavelet-Based Approach to Progressive Encoding and Transmission of Image Blocks

In this paper, a modified version of the discrete reversible (integer-to-integer) Le Gall wavelet transform (DLGT), distinguishing itself by apparently improved space localization properties and visibly extended potential capabilities, is proposed. The key point of the proposal – ensuring full decorrelation of Le Gall wavelet coefficients across the lower scales. Based on the latter circumstance, a novel exceptionally fast procedure for computing the integer DLGT spectra of the selected image blocks (regions of interest ROI) is presented. It is shown that the new developments can be efficiently applied to progressive encoding and transmission of image blocks. Progressive encoding and transmission of image blocks is achieved by first transmitting a “rough” estimate of the original image, then sending further details related to one or another image block (ROI). To translate the idea into action, the zero-treebased encoder SPIHT (Set Partitioning in Hierarchical Trees) with an improved quad-tree analysis scheme is employed.