EZW algorithm using depth-first representation of the wavelet zerotree

In this paper, we present a scheme for the implementation of the embedded zerotree wavelet (EZW) algorithm. The approach is based on using a depth-first search (DFS) to traverse the ancestor-descendant tree hierarchies of the wavelet coefficients. The EZW algorithm uses a raster scan to encode the coefficients. Two other tree-searching strategies to encode the coefficients would be by using a breath-first or depth-first traversal of the wavelet coefficients in the tree hierarchies. The DFS simplifies and diversifies the implementation of the EZW algorithm at the expense of a slight decrease in coding efficiency. The different tree-searching strategies are compared and simulation results using standard images are presented