Open Access

An application of nonstationary wavelets

Journal of Inequalities and Applications20142014:153

https://doi.org/10.1186/1029-242X-2014-153

Received: 25 December 2013

Accepted: 16 April 2014

Published: 2 May 2014

Abstract

Using nonstationary wavelets, we investigate the wavelet expansion in the standard Besov spaces. Especially, the nonstationary wavelets’ characterization for Besov spaces is given.

MSC:42C40, 35Q30, 41A15.

Keywords

Besov spacesnonstationary waveletcharacterization

1 Introduction

In this paper, we shall study the nonstationary wavelet [1] expansion in the standard Besov spaces. Nonstationary wavelet systems are generally obtained from a sequence of nonstationary refinable functions.

Definition 1.1 A sequence of functions { ϕ j 1 } j N in L 2 ( R ) is said to consist of nonstationary refinable functions if, for all j N ,
ϕ ˆ j 1 ( ξ ) = a ˆ j ( ξ 2 ) ϕ ˆ j ( ξ 2 ) , a.e.  ξ R ,

where a ˆ j are 2π-periodic measurable functions, called refinement masks, or simply masks.

The classical Fourier transform is defined by f ˆ ( ξ ) : = R n f ( x ) e i x ξ d x for f L 1 . The standard extension can be made to L 2 functions. Wavelet functions ψ j 1 l , j N and l = 1 , 2 , , Z j (quite often Z j = 3 ), are generally obtained from nonstationary refinable functions by
ψ ˆ j 1 l ( ξ ) = b ˆ j l ( ξ 2 ) ϕ ˆ j ( ξ 2 ) , j N , l = 1 , 2 , , Z j ,

where b ˆ j l ( ξ ) are 2π-periodic measurable functions called wavelet masks. The masks a ˆ j and b ˆ j l satisfy | a ˆ j ( ξ ) | 2 + l = 1 Z j | b ˆ j l ( ξ ) | 2 = 1 . When Z j = 3 , let b ˆ j 1 ( ξ ) : = e i ξ a ˆ j ( ξ + π ) ¯ , b ˆ j 2 ( ξ ) : = 2 1 A j ( ξ ) + e i ξ A j ( ξ ) ¯ and b ˆ j 3 ( ξ ) : = 2 1 A j ( ξ ) e i ξ A j ( ξ ) ¯ , where A j ( ξ ) : = 1 | a ˆ j ( ξ ) | 2 | a ˆ j ( ξ + π ) | 2 . Define X ( ϕ 0 ; { ψ j l } j N 0 , l = 1 , 2 , , Z j ) : = { ϕ 0 ( k ) : k Z } { ψ j ; j , k l : j N 0 , l = 1 , 2 , , Z j } . Then the following theorem holds.

Theorem 1.1 (i) (Theorem  1.3, [1]). Let a ˆ j ( ξ ) : = a ˆ m j , l j I (mask for the pseudo-spline of type I with order ( m j , l j ) ). Then X ( ϕ 0 ; { ψ j l } j N 0 , l = 1 , 2 , 3 ) is a compactly supported and C wavelet frame in H 2 s ( R ) for arbitrary s > 0 ;

(ii) (Theorem  1.4, [2]). When a ˆ j : = a ˆ m j , m j I , then X ( ϕ 0 ; { ψ j } j N 0 ) ( Z j = 1 ) is a compactly supported orthonormal basis in L 2 ( R ) , and ψ j ( Z j = 1 ) has m j + 1 vanishing moments. More precisely, Supp ϕ j [ L , L ] and Supp ψ j [ L , L ] with uniform constant L > 0 .

In this paper, we use H 2 s ( R ) to denote the classical L 2 -Sobolev spaces with the smoothness parameter s. It is well known that Besov spaces contain a large number of fundamental spaces, such as Sobolev spaces, Hölder spaces, Lipschitz spaces etc. [3, 4]. They are frequently used in certain PDEs as the solution spaces. To extend the result of (i) in Theorem 1.1, we shall characterize Besov spaces by using nonstationary wavelets in this paper. It should be pointed out that Bittner and Urban [5] study the following standard Besov spaces: Let 0 < p , q , s > 0 , and let [ s ] stand for the largest integer less than or equal to s,
B p , q s ( R n ) : = { f L p ( R n ) : | f | B p , q s ( R n ) < } .

Here, | f | B p , q s ( R n ) : = ( 2 j s ω p M ( f , 2 j ) ) j Z q with M [ s ] + 1 and ω p M ( f , 2 j ) denotes the M th order smooth modulus of a function f, defined by sup | h | 2 j Δ h M f ( ) L p ( R n ) as usual. The classical difference operator Δ h is defined by Δ h f ( ) : = f ( + h ) f ( ) , as well as Δ h M f = Δ h ( Δ h M 1 f ) for a positive integer M > 1 . The Besov (quasi-)norm is given by f B p , q s ( R n ) : = f L p ( R n ) + | f | B p , q s ( R n ) and the two integers M , M > s yield equivalent norms ([6], Remark 3.2.2).

Based on Hermite multiwavelets, Bittner and Urban characterize B p , q s by using sequence norms,
a p : = ( a j 0 , k ) k Z n p , b p , q s : = ( 2 j ( s + n 2 n p ) ( b j , k ) k Z n p ) j j 0 q

for a = ( a k ) k Z n p , b = ( b j , k ) j j 0 , k Z n p , q s . However, due to the regularity restrictions of the Hermite splines, their characterization requires 1 p < s < min { 3 , 1 + 1 p } in the quadratic case and 1 + 1 p < s < min { 4 , 2 + 1 p } in the cubic one (e.g. [5, 7, 8]). In [9], we remove that restriction of s by using the B-spline wavelets with weak duals as introduced in [10], but the supports of the wavelets become larger as s increases. So, the main result of this paper is to characterize Besov spaces via nonstationary wavelets because of their arbitrary smoothness and uniform support.

Let N, Z, and R be the set of positive integers, the set of integers, and the set of real numbers, respectively, as well as N 0 : = N { 0 } . Throughout this paper, we use A B to abbreviate that A is bounded by a constant multiple of B, A B is defined as B A and A B means A B and B A . Write
f , g = Ω f ( t ) g ( t ) ¯ d t

for f L p ( Ω ) , g L p ( Ω ) with Lebesgue measurable set Ω R n , 1 p + 1 p = 1 , and 1 p . For a Lebesgue measurable function f, the support of f means the set Supp ( f ) : = { x R : f ( x ) 0 } , which is well defined up to a set of measure 0. Define f j , k ( ) : = 2 j 2 f ( 2 j k ) throughout this paper.

Now, we state the Main Theorem of this paper.

Main Theorem Let 1 p 2 , 0 < q ; ϕ j and ψ j are from (ii) of Theorem  1.1. When f = k c 0 , k ϕ 0 ; 0 , k + j = 0 k Z d j , k ψ j ; j , k , then
f B p , q s ( R ) c l p + d l p , q s
for s > 0 . Moreover, when m = inf j m j + 1 , c 0 , k : = f , ϕ 0 ; 0 , k , d j , k : = f , ψ j ; j , k with j N 0 and k Z , then
c l p + d l p , q s f B p , q s ( R )

for f B p , q s ( R ) , 0 < s < m .

2 Proof of Main Theorem

This section is devoted to proving the Main Theorem. We begin with three lemmas for proving upper and lower bounds of the characterization.

Lemma 2.1 Let p ( 2 3 , 2 ] , q ( 0 , ] , s > 0 be arbitrary, then ϕ 0 B p , q s C and ψ j B p , q s C with a uniform constant C > 0 for j N 0 .

Proof First, we will show ϕ j p C with a uniform constant C for all j. By Lemma 2.1, Theorem 2.8 in [2], and Theorem 2.1 in [11], ϕ j are all compactly supported and Supp ϕ j [ L , L ] for a uniform constant L > 0 . Therefore, Supp ψ j [ L , L ] for all j because of ψ ˆ j 1 ( ξ ) : = b ˆ j 1 ( ξ 2 ) ϕ ˆ j ( ξ 2 ) = e i ξ a ˆ j ( ξ + π ) ¯ ϕ ˆ j ( ξ 2 ) . Note that ϕ j L 2 1 (Lemma 2.2 in [2]). This with
[ | ϕ j ( x ) | p d x ] 1 p { [ ( 1 ( 1 + x 2 ) p ) 2 2 p d x ] [ ( ( 1 + x 2 ) p | ϕ j ( x ) | p ) 2 p d x ] p 2 } 1 p C p , 2 ϕ j L 2
(2.1)

for 2 3 < p 2 leads to ϕ j L p C for all j N 0 . ψ j L p C holds similarly.

Second, let [ s ] + 1 = M , by (2.1),
ϕ 0 B p , q s = ϕ 0 p + ( 2 l s ω p M ( ϕ 0 , 2 l ) ) l Z l q ϕ 0 p + 2 l ( s M ) l q ϕ 0 ( M ) p ϕ 0 p + ϕ 0 ( M ) 2 ϕ 0 H 2 M .
Note that
ϕ 0 H 2 M 2 k | ϕ 0 , ϕ 0 ; 0 , k | 2 + j = 0 k 2 2 j s | ϕ 0 , ψ j ; j , k | 2 = 1
because of Corollary 3.3 in [1] and the orthonormality property of { ϕ 0 ; 0 , k } k , { ψ j ; j , k } j , k . This with (2.1) leads to
ϕ 0 B p , q s C ,

as well as ψ j B p , q s C for j N 0 with a uniform constant C > 0 . Thus, the result holds. □

The second lemma comes from [5], Lemma 3.4 and the third one comes from [6], (3.2.26).

Lemma 2.2 Suppose φ B p , σ ( R n ) is compactly supported with 0 < p , q and 0 < s < σ . Then
k Z n c k φ j , k ( ) B p , q s ( R n ) 2 ( σ + n 2 n p ) j c p , j = j 0 k Z n d j , k φ j , k ( ) B p , q s ( R n ) d p , q s .

Note that the constants are uniform because of φ B p , σ ( R n ) C .

Lemma 2.3 If f L p ( R ) , 1 p , define σ as a closed interval of R and let P m 1 be the set of m-order polynomials, then
inf P Π m 1 f P L p ( σ j , k ) ω p m ( f , 2 j , σ j , k ) ,

where σ j , k : = 2 j ( σ + k ) and ω p m ( f , 2 j , σ j , k ) : = sup | h | 2 j Δ h M f ( ) L p ( ( σ j , k ) h , M ) with ( σ j , k ) h , M : = { x σ j , k , x + l h σ , l = 1 , 2 , , M } .

Now, we are in the position to show the Main Theorem.

By Lemma 2.1, ϕ 0 B p , q s C and ψ j B p , q s C . This, with Lemma 2.2, shows that
k c 0 , k ϕ 0 ; 0 , k + j = 0 k Z d j , k ψ j ; j , k B p , q s ( R ) k c 0 , k ϕ 0 ; 0 , k B p , q s ( R ) + j = 0 k Z d j , k ψ j ; j , k B p , q s ( R ) c p + d p , q s .
To prove the lower bound, one finds that | f , ϕ 0 ; 0 , k | f L p ( σ 0 , k ) ,
k | f , ϕ 0 ; 0 , k | p k σ 0 , k | f ( x ) | p d x f p p ,
where σ 0 , k : = Supp ϕ 0 ; 0 , k . Let σ j , k : = Supp ψ j ; j , k . This with m : = inf j m j + 1 shows that
| f , ψ j ; j , k | p = inf P P m 1 | f P , ψ j ; j , k | p inf P P m 1 f P L p ( σ j , k ) p 2 j p 2 j p p 2 j p 2 j p p ω p m ( f , 2 j , σ j , k ) ,
where the equality comes from the m j + 1 vanishing moments of ψ j by Theorem 1.1 and the second inequality holds due to Lemma 2.3. Then ( d j , k ) k l p 2 j 2 j p ω p m ( f , 2 j ) by the same proof as of (3.3) in [8]. Therefore,
d l p , q s ( 2 s + j 2 j p 2 j 2 j p ω p m ( f , 2 j ) ) j Z l q f B p , q s ( R ) .
Remark 2.1 In conclusion, we have a characterization of Besov spaces by
f B p , q s c l p + d l p , q s

with c 0 , k : = f , ϕ 0 ; 0 , k , d j , k : = f , ψ j ; j , k , 1 p 2 , 0 < q , 0 < s < m , and f B p , q s ( R ) .

Remark 2.2 Some questions are left to be considered. Note that we assume 1 p 2 in our Main Theorem. Then a natural question is to study the case for p > 2 . Another one is to discuss whether or not the wavelet frames of (i) in Theorem 1.1 can characterize Besov spaces. The last question is to relax the restriction s ( 0 , m ) .

Declarations

Acknowledgements

The author would like to thank the referees for their helpful comments. This work is supported by the National Natural Science Foundation of China, Tian Yuan Foundation (Grant No. 11226106) and the National Natural Science Foundation of China (Grant No. 11201094).

Authors’ Affiliations

(1)
Department of Mathematics, Tianjin Polytechnic University

References

  1. Han B, Shen Z: Characterization of Sobolev spaces of arbitrary smoothness using nonstationary tight wavelet frames. Isr. J. Math. 2009, 172: 371-398. 10.1007/s11856-009-0079-9MathSciNetView ArticleGoogle Scholar
  2. Han B, Shen Z:Compactly supported symmetric C wavelets with spectral approximation order. SIAM J. Math. Anal. 2009, 40: 905-938.MathSciNetView ArticleGoogle Scholar
  3. Hans T: Theory of Function Spaces II. Birkhäuser, Basel; 1992.Google Scholar
  4. Härdle W, Kerkyacharian G, Picard D, Tsybakov A: Wavelets, Approximation and Statistal Applications. Springer, Berlin; 1998.View ArticleGoogle Scholar
  5. Bittner K, Urban K: On interpolatory divergence-free wavelets. Math. Comput. 2007,76(258):903-929. 10.1090/S0025-5718-06-01949-1MathSciNetView ArticleGoogle Scholar
  6. Cohen A: Wavelet methods in numerical analysis. Handb. Numer. Anal. VII. In Handbook of Numerical Analysis. North-Holland, Amsterdam; 2000:417-711.Google Scholar
  7. Liu Y, Zhao J: Convergence of Hermite interpolatory operators. Sci. China Math. 2010,53(8):2115-2126. 10.1007/s11425-010-4047-yMathSciNetView ArticleGoogle Scholar
  8. Zhao J: Interpolatory Hermite splines on rectangular domains. Appl. Math. Comput. 2010, 216: 2799-2813. 10.1016/j.amc.2010.03.130MathSciNetView ArticleGoogle Scholar
  9. Liu Y, Zhao J: An extension of Bittner and Urban’s theorem. Math. Comput. 2013,82(281):401-411.View ArticleGoogle Scholar
  10. Jia RQ, Wang JZ, Zhou DX: Compactly supported wavelets bases for Sobolev spaces. Appl. Comput. Harmon. Anal. 2003, 15: 224-241. 10.1016/j.acha.2003.08.003MathSciNetView ArticleGoogle Scholar
  11. Cohen A, Dyn N: Nonstationary subdivision schemes and multiresolution analysis. SIAM J. Math. Anal. 1996, 27: 1745-1769. 10.1137/S003614109427429XMathSciNetView ArticleGoogle Scholar

Copyright

© Zhao; licensee Springer. 2014

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.