Topics in multiresolution and filter banks

Abstract

This thesis studies a number of topics relevant to signal and image representation and coding, in the context of multiresolution analysis by filter banks and wavelets. It discusses signal approximation through multiresolution analysis, and applications to the study of the symptotic approximation power of radial basis function neural networks; it proposes a genuine finite-dimensional framework for multiresolution analysis in finite dimensional spaces; it discusses an improved image coding technique which can be made to work according, to the characteristics of the human visual system at high compression ratios. (...)