How to generate smoother refinable functions from given ones
In the univariate case, one can obtain a smoother function from a given refinable function \(\phi\) by convolving with the characteristic function \(\chi\) on the unit interval, \(\psi = \chi \ast \phi\). For the corresponding mask of \(\psi\) that means \(a^\ast_\psi(z) =\displaystyle {z+1\over 2} \; a^\ast_\phi(z)\). This factorization property of the masks can also be written in terms of subdivision operators \[ \Delta S_{a_\phi} = S_{2a_\psi} \Delta, \] where \(S_{a_\phi}c := \sum_{k \in {\mathbb Z}} a(\cdot -2k) c(k)\) for sequences \(c :{\mathbb Z}^s \to {\mathbb R}\) and where \(\Delta\) is the backward difference operator. In the multivariate case, this concept can be extended. The algebraic interpretation of the identity \(\Delta S_a = S_b \Delta\) with matrix valued masks \(a,\, b\) and a discrete backward gradient \(\Delta\) now involves the notion of quotient ideals of multivariate algebraic polynomials. For the entire collection see [Zbl 1027.00021].
- Smoothness of multivariate refinable functions with infinitely supported masks
- Characterization of Smoothness of Multivariate Refinable Functions in Sobolev Spaces
- Multivariate refinable functions, differences and ideals -- a simple tutorial
- Smooth refinable functions and wavelets obtained by convolution products
- Polynomial interpolation, ideals and approximation order of multivariate refinable functions
This page was built for publication: How to generate smoother refinable functions from given ones
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4440846)
