Image coding using overlapping fractal transform in the wavelet domain

Fractal and frequency decompositions lead to comparable image compression results. However, these two approaches are not incompatible and it is possible to express the operations performed by some fractal coders in the frequency domain. The aim of this paper is to propose an image representation technique combining both fractal and frequency decompositions thanks to the use of the wavelet multiresolution representation. This model takes into account a certain kind of self-similarity that can be translated in terms of simple operations on the high frequency coefficients of a subband decomposition of the image. The way to solve analytically the inverse problem is also described in this context