Morphological Zerotree Compression Coding Based on Integer Wavelet Transform for Iris Image

Abstract

After comparing features of the EZW and the MRWD which are famous wavelet image compression code algorithm, we present an algorithm in view of iris texture characteristic. This algorithm is based on integer wavelet transformation while it has less bit planes, and wavelet coefficients do not need to be quantified, so the image can be completely recovered. Under the condition, we utilize the wavelet coefficient zerotree structure and the important wavelet coefficient clustering with similar statistical property which is based on bit plane decomposing, applying zerotree structure to express non-important wavelet coefficient effectively, using morphology cluster operation to simulate iris texture growth characteristic of important wavelet coefficient, realizing morphological zerotree compression. The experimental results indicate this algorithm has the higher compression rate and the better restoration effect and it can be applied effectively in iris identification.