Filters for anisotropic wavelet decompositions (Q1713113): Difference between revisions

The paper is written on the whole topic of multivariate (mutual or mixed) analysis. Such type of analysis considers simultaneously many objects such as functions, statistical series, measures, \dots and studies their simultaneous behavior with some correlation type measures and thus extracts their characteristics and their mutual effects at the same time. This type of analysis has been proved to be efficient especially in the study of anisotropic data. In the present work, the authors investigate the use of anisotropic wavelets and associated filters characterized by an interplay between filterbanks and anisotropic dilations to study directed features in higher dimensions. Filterbanks presenting refinable functions such as wavelets are characterized. These facts induce two important findings: Firstly, the construction of anisotropic wavelet bases from generalized shearlets have been proved to be possible for higher dimensions with no restriction on the scaling factors. Secondly, the multivariate filterbank system is proved to satisfy the slope resolution property, which is a crucial point in the study and application of directional transforms. Overall, the arguments treated are important as they induce high accurate algorithms. The paper is interesting, with hard techniques and innovative subject. Moreover, it is surely significant for scientists in both theoretical and applied fields.