A characterization of dimension functions of wavelets

DOI10.1006/ACHA.2000.0327MaRDI QIDQ5929153zbMATH OpenOpenAlexFDO

Authors Marcin Bownik, Ziemowit Rzeszotnik, Darrin Speegle

Publication date 17 September 2001

Published in Applied and Computational Harmonic Analysis (Search for Journal in Brave)

Full work available at URL https://semanticscholar.org/paper/589d1d38440aed03431f4b71f282cd833657cf68

zbMATH Keywords

dimension function MSF wavelet multiwavelet orthonormal wavelet

Mathematics Subject Classification ID

Nontrigonometric harmonic analysis involving wavelets and other special systems (42C40)

Review text

A finite set \(\Psi = \{ \psi^1, \ldots, \psi^L \} \subset L^2(R^d)\) is called an orthonormal multiwavelet, if the collection NEWLINE\[NEWLINE\{ \psi^l_{j,k} (x) = |\det A|^{j/2} \psi^l (A^j x-k) : j\in Z,\;k \in Z^d,\;l=1, \ldots, L \}NEWLINE\]NEWLINE forms an orthonormal basis for \(L^2(R^d)\). Here, \(A\) is an expansive matrix which preserves \(Z^d\). The dimension function of such a multiwavelet is defined as NEWLINE\[NEWLINED_\Psi (\xi) = \sum_{l=1}^L \sum_{j=1}^\infty \sum_{k\in Z^d} |\widehat \psi^l (B^j(\xi+k))|^2,NEWLINE\]NEWLINE where \(B=A^T\). NEWLINENEWLINENEWLINEIt is well known that dimension functions are measurable, nonnegative, \(Z^d\)-periodic and integer-valued. In this paper the authors provide a full characterization of all dimension functions, by adding to the above properties the following conditions: NEWLINE\[NEWLINE\int_{(-1/2,1/2]^d} D(\xi) d\xi = L/(q-1), \quad \text{where }q = |\det A|,\tag{1}NEWLINE\]NEWLINE NEWLINE\[NEWLINE\liminf_{n\to \infty} D(B^{-n}\xi) \geq 1,\tag{2}NEWLINE\]NEWLINE NEWLINE\[NEWLINE\sum_{k \in Z^d} \chi_\Delta(\xi+k) \geq D(\xi), \text{ a.e. for }\Delta = \{ \xi \in R^d: D(B^{-j}\xi) \geq 1;\;j\in N\cup\{0\} \},\tag{3}NEWLINE\]NEWLINE NEWLINE\[NEWLINE\sum_{d\in {\mathcal D}} D(\xi+B^{-1}d) = D(B\xi)+L, \quad \text{a.e.} .\tag{4}NEWLINE\]NEWLINE Here, \({\mathcal D}= \{d_1, \dots, d_q \}\) denotes the set of representatives of \(Z^d/BZ^d\), with \(d_1=0\). NEWLINENEWLINENEWLINEThe authors further study the properties of the dimension function using these characterizing conditions, and provide several examples of wavelets for given dimension functions. In particular, they show that for any dimension function \(D\), there exists an MSF wavelet whose dimension function is \(D\).

Recommendations

Cites work

Cited in

(59)

This page was built for publication: A characterization of dimension functions of wavelets

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5929153)