Undecimated wavelet transform from orthogonal spline wavelets
The undecimated wavelet transform is the discrete wavelet transform without subsampling. It is known to be translation invariant. For some applications certain local phenomena are characterized by computing for a continuous wavelet transform, certain indicators like zero crossings, relative extrema, or vertical edges. It is argumented in this paper that a viable alternative for the continuous transform is the undecimated transform. This transform is redundant and does not in general correspond to a multiresolution analysis. However, in this paper, it is shown that when orthogonal spline wavelets are used, then fast algorithms exist and the analysis can be placed into the context of some extended notion of a multiresolution analysis. For the entire collection see [Zbl 0868.00052].
This page was built for publication: Undecimated wavelet transform from orthogonal spline wavelets
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4352716)
