Identification of multiscale model for image processing

Uses wavelet representation to provide a new linear scale autoregressive (LSA) model for images. This modeling approach takes advantage of the information contained in the lower resolutions as well as of the information contained in the detailed images. Detailed images are available from information provided by the wavelet decomposition. It is expected that the proposed model will lead to an improved LSA model, which in turn further enhances the quality of multiscale-based image processing applications. Test images are used to illustrate the benefits of the new modeling approach.