BOUNDEDNESS OF THE WAVELET TRANSFORM IN CERTAIN FUNCTION SPACES
R.S. PATHAK AND S.K. SINGH DEPARTMENT OFMATHEMATICS,
BANARASHINDUUNIVERSITY, VARANASI- 221 005, INDIA
ramshankarpathak@yahoo.co.in
Received 06 October, 2005; accepted 25 January, 2007 Communicated by L. Debnath
ABSTRACT. Using convolution transform theory boundedness results for the wavelet transform are obtained in the Triebel space-LΩ,kp , Hörmander space-Bp,q(Rn)and general function space- L∞,k,wherekdenotes a weight function possessing specific properties in each case.
Key words and phrases: Continuous wavelet transform, Distributions, Sobolev space, Besov space, Lizorkin-Triebel space.
2000 Mathematics Subject Classification. 42C40, 46F12.
1. INTRODUCTION
The wavelet transformW of a functionf with respect to the waveletψis defined by (1.1) f(a, b) = (W˜ ψf)(a, b) =
Z
Rn
f(t)ψa,b(t)dt= (f∗ha,0)(b),
whereψa,b = a−n2ψ(x−ba ), h(x) = ψ(−x), b ∈ Rn anda > 0, provided the integral exists. In view of (1.1) the wavelet transform (Wψf)(a, b)can be regarded as the convolution of f and ha,0. The existence of convolutionf∗ghas been investigated by many authors. For this purpose Triebel [6] defined the spaceLΩ,kp and showed that for certain weight functionsk, f ∗g ∈LΩ,kp , wheref, g ∈ LΩ,kp , 0< p ≤1. Convolution theory has also been developed by Hörmander in the generalized Sobolev spaceBp,q(Rn), 1≤p≤ ∞.
In Section 2 of the paper, a definition and properties of the spaceLΩ,kp are given and a bound- edness result for the wavelet transform Wψf is obtained. In Section 3 we recall the definition and properties of the generalized Sobolev space Bp,q(Rn)due to Hörmander [1] and obtain a certain boundedness result for Wψf. Finally, using Young’s inequality a third boundedness result is also obtained.
303-05
2. BOUNDEDNESS OF W INLΩ,kp Let us recall the definition of the spaceLΩ,kp by Triebel [6].
Definition 2.1. LetΩbe a boundedC∞-domain inRn. Ifk(x)is a non-negative weight function inRnand0< p ≤ ∞, then
(2.1) LΩ,kp =
f|f ∈S0,suppF f ⊂Ω;
kf kLΩ,k
p =kkf kLp= Z
Rn
kp(x)|f(x)|pdx 1p
<∞
.
Ifk(x) = 1thenLΩ,kp =LΩp.
We need the following theorem [6, p. 369] in the proof of our boundedness result.
Theorem 2.1 (Hans Triebel). Ifkis one of the following weight functions:
(2.2) k(x) = |x|α, α ≥0
(2.3) k(x) =
n
Y
j=1
|xj|αj, αj ≥0
(2.4) k(x) =kβ,γ(x) =eβ|x|γ, β ≥0,0≤γ ≤1 and0< p ≤1, then
(2.5) LΩ,kp ∗LΩ,kp ⊂LΩ,kp
and there exists a positive numberCsuch that for allf, g∈LΩ,kp ,
(2.6) kf ∗gkLk
p ≤CkfkLk pkgkLk
p.
Using the above theorem we obtain the following boundedness result for the wavelet trans- formWψf.
Theorem 2.2. Letf ∈LΩ,kp andψ ∈LΩ,kp ,0< p≤1,then for the wavelet transformWψf we have the estimates:
(2.7) k(Wψf)(a, b)kLkp≤Caα+n2 kf kLkpkψ kLkp for (2.2);
(2.8) k(Wψf)(a, b)kLk
p≤Ca|α|+n2 kf kLk
pkψ kLk
p for (2.3);
(2.9) k(Wψf)(a, b)k
Lkβ,γp
≤Can2e12βa2γ kf k
Lkβ,γp
kψ k
Lkβ,2γp
for (2.4), whereb∈Rnanda >0.
Proof. Fork(x) = |x|α, α >0,we havek(az) =aαk(z)and kha,0kLk
p = Z
Rn
kp(x)(a−n2|h(x a)|)pdx
1p
=an2 Z
Rn
kp(az)|h(z)|pdz 1p
=an2 Z
Rn
apαkp(z)|h(z)|pdz 1p
=an2+α Z
Rn
kp(z)|h(z)|pdz 1p
=an2+α khkLk p
=an2+α kψ kLk p . Fork(x) =Qn
j=1|xj|αj, αj ≥0, we havek(az) = a|α|k(z)and kha,0kLk
p = Z
Rn
kp(x)(a−n2|h(x a)|)pdx
1p
=an2 Z
Rn
kp(az)|h(z)|pdz 1p
=an2 Z
Rn
ap|α|kp(z)|h(z)|pdz 1p
=an2+|α|
Z
Rn
kp(z)|h(z)|pdz p1
=an2+|α|khkLk
p
=an2+|α|kψ kLk
p .
Next, fork(x) = kβ,γ(x) = eβ|x|γ, β ≥0,0≤γ ≤1, we have kβ,γ(az) = eβ|az|γ =eβaγ|z|γ ≤eβa
2γ+|z|2γ
2 =e12βa2γe12β|z|2γ =e12βa2γkβ,2γ(z), and
kha,0k
Lkβ,γp
= Z
Rn
kβ,γp (x) a−n2
hx
a
p
dz 1p
=an2 Z
Rn
kpβ,γ(az)|h(z)|pdz 1p
≤an2 Z
Rn
e12pβa2γkβ,2γp (z)|h(z)|pdz 1p
=an2e12βa2γ Z
Rn
kpβ,2γ(z)|h(z)|pdz 1p
=an2e12βa2γ khk
Lkβ,2γp
=an2e12βa2γ kψ k
Lkβ,2γp
.
The proofs of (2.7), (2.8) and (2.9) follow from (2.6).
3. BOUNDEDNESS OFW INBp,k
The space Bp,k(Rn) was introduced by Hörmander [1], as a generalization of the Sobolev spaceHs(Rn), in his study of the theory of partial differential equations. We recall its definition.
Definition 3.1. A positive functionkdefined inRnwill be called a temperate weight function if there exist positive constantsCandN such that
(3.1) k(ξ+η)≤(1 +C|ξ|)Nk(η); ξ, η∈Rn,
the set of all such functionsk will be denoted byK . Certain properties of the weight function kare contained in the following theorem whose proof can be found in [1].
Theorem 3.1. If k1 andk2 belong to K ,thenk1 +k2, k1k2, sup(k1, k2), inf(k1, k2), are also inK . Ifk ∈K we haveks ∈K for every reals,and ifµis a positive measure we have either µ∗k≡ ∞or elseµ∗k ∈K.
Definition 3.2. Ifk ∈K and1≤p≤ ∞, we denote byBp,kthe set of all distributionsu∈S0 such thatuˆis a function and
(3.2) kukp,k= (2π)−n Z
|k(ξ)ˆu|pdξ 1p
<∞, 1≤p < ∞;
(3.3) kuk∞,k=esssup|k(ξ)ˆu(ξ)|.
We need the following theorem [1, p.10] in the proof of our boundedness result.
Theorem 3.2 (Lars Hörmander). If u1 ∈ Bp,k1TE0andu2 ∈ B∞,k2 then u1 ∗u2 ∈ B∞,k1k2, and we have the estimate
(3.4) ku1∗u2 kp,k1k2≤ku1 kp,k1ku2 k∞,k2, 1≤p < ∞.
Using the above theorem we obtain the following boundedness result.
Theorem 3.3. Letk1 and k2 belong to K . Assume thatf ∈ Bp,k1
TE0 and ψ ∈ B∞,k2 then the wavelet transform(Wψf)(a, b) = (f∗ha,0)(b),defined by (1.1) is inBp,k1k2, and
(3.5) kWψf(a, b)kp,k1k2≤an2k2 1
2a2
kf kp,k1
1 + C
2t2 N
ψ(t)ˆ ∞
.
Proof. Since
kha,0k∞,k
2 =esssup
k2(ξ)ˆha,0(ξ)
=esssup
k2(ξ)an2ψ(aξ)ˆ
≤an2esssup
k2(at) ˆψ(t)
≤an2k2 1
2a2
esssup
1 + C
2t2 N
ψ(t)ˆ on using (3.1). Hence by Theorem 3.2 we have
kWψf(a, b)kp,k
1k2 =k(f∗ha,0(b)kp,k1k2
≤an2k2
1 2a2
kf kp,k1
1 + C
2t2 N
ψ(t)ˆ ∞
.
This proves the theorem.
4. A GENERAL BOUNDEDNESSRESULT
Using Young’s inequality for convolution we obtained a general boundedness result for the wavelet transform. In the proof of our result the following theorem will be used [3, p. 90].
Theorem 4.1. Letp, q, r≥1and1p+1q+1r = 2. Letk ∈Lp(Rn), f ∈Lq(Rn)andg ∈Lr(Rn), then
kf ∗gk∞,k = Z
Rn
k(x)(f ∗g)(x)dx
= Z
Rn
Z
Rn
k(x)f(x−y)g(y)dxdy
≤Cp,q,r;n kkkpkf kqkg kr. The sharp constantCp,q,r;n = (CpCqCr)n, whereCp2 = p
1p
p0
1
p0 with (1p + p10 = 1). Using The- orem 4.1 and following the same method of proof as for Theorem 3.3 we obtain the following boundedness result.
Theorem 4.2. Letp, q, r ≥1, 1p+1q+1r = 2andk ∈Lp(Rn). Letf ∈Lq(Rn)andψ ∈Lr(Rn), then
kWψf k∞,k ≤ Cp,q,r;nanr−n2 kk kpkf kqkψ kr whereCp,q,r;n = (CpCqCr)n,Cp2 = p
1p
p0
1
p0 with 1p +p10 = 1.
REFERENCES
[1] L. HÖRMANDER, The Analysis of Linear Partial Differential Operators II, Springer-Verlag, Berlin Heiedelberg New York, Tokyo, 1983.
[2] T.H. KOORNWINDER, Wavelets, World Scientific Publishing Co. Pty. Ltd., Singapore , 1993.
[3] E.H. LIEBANDM. LOSS, Analysis, Narosa Publishing House, 1997. ISBN: 978-81-7319-201-2.
[4] R.S. PATHAK, Integral Transforms of Generalized Functions and Their Applications, Gordon and Breach Science Publishers, Amsterdam, 1997.
[5] R.S. PATHAK, The continuous wavelet transform of distributions, Tohoku Math. J., 56 (2004), 411–421.
[6] H. TRIEBEL, A note on quasi-normed convolution algebras of entire analytic functions of expo- nential type, J. Approximation Theory, 22(4) (1978), 368–373.
[7] H. TRIEBEL, Multipliers in Bessov-spaces and inLΩp- spaces (The cases0< p≤1andp=∞), Math. Nachr., 75 (1976), 229–245.
[8] H.J. SCHMEISSER AND H. TRIEBEL, Topics in Fourier Analysis and Function Spaces, John Wiley and Sons, Chichester, 1987.