A new zero-tree image compression algorithm

In this note we describe a new class of zero-tree image compression algorithms which are under evaluation for international image compression standards. A wavelet or multiwavelet image decomposition can be viewed as a tree structured set of coefficients giving a hierarchical data structure for representing the transformed image. Some of these tree structures may have all the coefficients zero, such a tree we call a zerotree. From the point of view of compressing the image, all their coefficients in a zerotree can be coded by just a 1 bit token. The wavelet transform is very efficient at removing the linear correlations in an image, but by its nature it cannot remove the nonlinear inter-band correlations in the transformed image. However, the zero-tree algorithm uses the nonlinear spatial correlation inherent in a transformed image to compress it at below its entropy.