A sufficient and necessary condition for construction of 3-band dual tree complex wavelet (Q2846503)
From MaRDI portal
| This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use this page instead for the normal view: A sufficient and necessary condition for construction of 3-band dual tree complex wavelet |
scientific article; zbMATH DE number 6206142
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | A sufficient and necessary condition for construction of 3-band dual tree complex wavelet |
scientific article; zbMATH DE number 6206142 |
Statements
5 September 2013
0 references
shift-invariance
0 references
3-band wavelet
0 references
dual-tree complex wavelet
0 references
wavelet
0 references
0 references
A sufficient and necessary condition for construction of 3-band dual tree complex wavelet (English)
0 references
The complex wavelet transform, introduced by Kingsbury and further elaborated by Selesnick, offers an approximately shift-invariant multiscale approach with some advantages with respect to the conventional wavelet approach, such as less redundancy than undecimated wavelets and improved directional sensitivity. This transform can be implemented using a dual-tree structure that adapts the main ideas of the classical wavelet approach and relies on the explicit relation between the scaling and wavelet filter for the perfect reconstruction for orthogonal or biorthogonal wavelets. However, for \(M\)-band wavelets where \(M\neq 2\) the relationship between the scaling and the wavelet filter is not explicit and a different approach is needed for their construction. This paper examines the construction of 3-band dual-tree complex wavelets and establishes necessary and sufficient conditions for the analyticity of such systems. In particular, it derives a constructive approach where, if scaling filters satisfy one-delay condition, then the corresponding 3-band dual-tree complex wavelets are analytic. Wavelet filters impulse responses are explicitly computed.
0 references
0.7287402153015137
0 references
0.7226701378822327
0 references
0.7218189239501953
0 references
0.7206032276153564
0 references