<< Chapter < Page Chapter >> Page >

Approximation of functions with wavelets

In Example from Multiresolution analysis , we saw the Haar wavelet basis. This is the simplest wavelet one can imagine, but its approximation properties are not very good. Indeed, the accuracy of the approximation is somehow related to the regularity of the functions ψ and ϕ . We will show this for different settings: in case the function f is continuous, belongs to a Sobolev or to a Hölder space. But first we introduce the notion of regularity of a MRA.

Regularity of a multiresolution analysis

For wavelet bases (orthonormal or not), there is a link between the regularity of ψ and the number of vanishing moments. More precisely, we have the following:

Let { ψ j , k } be an orthonormal basis (ONB) in L 2 ( I R ) , with ψ C [ 0 . 1 e x ] 0 . 05 e m 1 . 25 e x m , ψ ( l ) bounded for l m and | ψ ( x ) | C ( 1 + | x | ) - α for α > m + 1 . Then we have:
x l ψ ( x ) d x = 0 for l = 0 , 1 , ... , m .

(this describes the “decay in frequency domain”).

This proposition implies the following corollary:

Suppose the { ψ j , k } are orthonormal. Then it is impossible that ψ has exponential decay, and that ψ C [ 0 . 1 e x ] 0 . 05 e m 1 . 25 e x , with all the derivatives bounded, unless ψ 0

This corollary tells us that a trade-off has to be done: we have to choose for exponential (or faster) decay in, either time or frequency domain; we cannot have both. We now come to the definition of a r - regular MRA (see [link] ).

(Meyer, 1990) A MRA is called r-regular ( r I N ) when ϕ C [ 0 . 1 e x ] 0 . 05 e m 1 . 25 e x r and for all m I N , there exists a constant C m > 0 such that:
| α ϕ ( x ) x α | C m ( 1 + | x | ) - m α r

If one has a r - regular MRA, then the corresponding wavelet ψ ( x ) C [ 0 . 1 e x ] 0 . 05 e m 1 . 25 e x r , satisfies [link] and has r vanishing moments:

x k ψ ( x ) d x = 0 , k = 0 , 1 , ... , r .

We now have the tools needed to measure the decay of approximation error when the resolution (or the finest level) increases.

Approximation of a continuous function

Let { ψ j , k } come from a r-regular MRA. Then, we have, for a continuous function f C [ 0 . 1 e x ] 0 . 05 e m 1 . 25 e x s ( 0 < s < r ) the following:
| < f , ψ j , k > | = O ( 2 - j ( s + 1 / 2 ) )
( s = 1 ). As ψ ( x ) ¯ = 0 and | f ( x ) - f ( k / 2 j ) | C | x - k / 2 j | , ( f is Lipschitz continuous), we have:
| < f , ψ j , k > | = | 2 j / 2 [ f ( x ) - f ( k / 2 j ) ] ψ ¯ ( 2 j x - k ) d x | , k Z Z C 2 j / 2 | x - k / 2 j | | ψ ¯ ( 2 j x - k ) | d x = C 2 - j / 2 | 2 j x - k | | ψ ¯ ( 2 j x - k ) | d x ( let u = 2 j x - k ) = 2 - j ( 1 + 1 / 2 ) C | u | | ψ ¯ ( u ) | d u < = O ( 2 - j ( 1 + 1 / 2 ) ) .

For s > 1 , we use proposition [link] and we iterate.

Note that the reverse of [link] is true: [link] entails that f is in C s .

Under the assumptions of [link] , the error of approximation of a function f C [ 0 . 1 e x ] 0 . 05 e m 1 . 25 e x s at scale V J is given by:
| | f - k < f , ϕ J , k > ϕ J , k | | 2 = O ( 2 - J s )
Using the exact decomposition of f given by Equation 30 from Multiresolution Analysis , one has:
| | f - P J f | | 2 = | | j J k < f , ψ j , k > ψ j , k | | 2 j J C 2 - j ( s + 1 / 2 ) | | k ψ j , k | | 2

Now, if we suppose that | ψ ( k ) | C ' ( 1 + | k | ) - m , ( m s + 1 ) , we obtain:

| | k ψ j , k | | 2 C ' 2 j / 2

Putting inequalities [link] and [link] together, we have:

| | f - P J f | | 2 C ' ' j J 2 - j s = C ' ' ( 1 - 2 - s ) - 1 2 - J s = O ( 2 - J s )

Hence, we verify with this last corollary that, as J increases, the approximation of f becomes more accurate.

Approximation of functions in sobolev spaces

Let us first recall the definition of weak differentiability, for this notion intervenes in the definition of a Sobolev space.

Let f be a function defined on the real line which is integrable on every bounded interval. If there exists a function g defined on the real line which is integrable on every bounded interval such that:
x y , x y g ( u ) d u = f ( y ) - f ( x ) ,

then the function f is called weakly differentiable. The function g is defined almost everywhere, is called the weak derivative of f and will be denoted by f ' .

A function f is N -times weakly differentiable if it has derivatives f , f ' , ... , f ( N - 1 ) which are continuous and f ( N ) which is a weak derivative.

We are now able to define Sobolev spaces.

Let 1 p < , m { 0 , 1 , ... } . The function f L p ( I R ) belongs to the Sobolev space W p m ( I R ) if it is m-times weakly differentiable, and if f ( m ) L p ( I R ) . In particular, W p 0 ( I R ) = L p ( I R ) .

The approximation properties of wavelet expansions on Sobolev spaces are given, among other, in Härdle et.al (see [link] ). Suppose we have at our disposal a scaling function ϕ which generates a MRA. The approximation theorem can be stated as follows:

(Approx. in Sobolev space) Let ϕ be a scaling function such that { ϕ ( . - k ) , k Z Z } is an ONB and the corresponding spaces V j are nested. In addition, let ϕ be such that
| ϕ ( x ) | | x | N + 1 d x < ,

and let at least one of the following assumptions hold:

  1. ϕ W q N ( I R ) for some 1 q <
  2. x n ψ ( x ) d x = 0 , n = 0 , 1 , ... , N , where ψ is the mother wavelet associated to ϕ .

Then, if f belongs to the Sobolev space W p N + 1 ( I R ) , we have:

| | P j f - f | | p = O ( 2 - j ( N + 1 ) ) , as j ,

where P j is the projection operator onto V j .

Approximation of functions in hölder spaces

Here we assume for simplicity that ψ has compact support and is C [ 0 . 1 e x ] 0 . 05 e m 1 . 25 e x 1 (the formulation of the theorems are slightly different for more general ψ ).

If f is Hölder continuous with exponent α , 0 < α < 1 at x 0 , then
max k { | < f , ψ j , k > | d i s t ( x 0 , s u p p ( ψ j , k ) ) - α } = O ( 2 - j ( 1 / 2 + α ) )

The reverse of theorem [link] does not exactly hold: we must modify condition [link] slightly. More precisely, we have the following:

Define, for ϵ > 0 , the set
S ( x 0 , j , ϵ ) = { k Z Z | s u p p ( ψ j , k ) ] x 0 - ϵ , x 0 + ϵ [ } .

If, for some ϵ > 0 and some α ( 0 < α < 1 ) ,

max k S ( x 0 , j , ϵ ) | < f , ψ j , k > | = O ( 2 - j ( 1 / 2 + α ) ) ,

then f is Hölder continuous with exponent α at x 0 .

Questions & Answers

How we are making nano material?
LITNING Reply
what is a peer
LITNING Reply
What is meant by 'nano scale'?
LITNING Reply
What is STMs full form?
LITNING
scanning tunneling microscope
Sahil
what is Nano technology ?
Bob Reply
write examples of Nano molecule?
Bob
The nanotechnology is as new science, to scale nanometric
brayan
nanotechnology is the study, desing, synthesis, manipulation and application of materials and functional systems through control of matter at nanoscale
Damian
Is there any normative that regulates the use of silver nanoparticles?
Damian Reply
what king of growth are you checking .?
Renato
What fields keep nano created devices from performing or assimulating ? Magnetic fields ? Are do they assimilate ?
Stoney Reply
why we need to study biomolecules, molecular biology in nanotechnology?
Adin Reply
?
Kyle
yes I'm doing my masters in nanotechnology, we are being studying all these domains as well..
Adin
why?
Adin
what school?
Kyle
biomolecules are e building blocks of every organics and inorganic materials.
Joe
anyone know any internet site where one can find nanotechnology papers?
Damian Reply
research.net
kanaga
sciencedirect big data base
Ernesto
Introduction about quantum dots in nanotechnology
Praveena Reply
what does nano mean?
Anassong Reply
nano basically means 10^(-9). nanometer is a unit to measure length.
Bharti
do you think it's worthwhile in the long term to study the effects and possibilities of nanotechnology on viral treatment?
Damian Reply
absolutely yes
Daniel
how to know photocatalytic properties of tio2 nanoparticles...what to do now
Akash Reply
it is a goid question and i want to know the answer as well
Maciej
characteristics of micro business
Abigail
for teaching engĺish at school how nano technology help us
Anassong
How can I make nanorobot?
Lily
Do somebody tell me a best nano engineering book for beginners?
s. Reply
there is no specific books for beginners but there is book called principle of nanotechnology
NANO
how can I make nanorobot?
Lily
what is fullerene does it is used to make bukky balls
Devang Reply
are you nano engineer ?
s.
fullerene is a bucky ball aka Carbon 60 molecule. It was name by the architect Fuller. He design the geodesic dome. it resembles a soccer ball.
Tarell
what is the actual application of fullerenes nowadays?
Damian
That is a great question Damian. best way to answer that question is to Google it. there are hundreds of applications for buck minister fullerenes, from medical to aerospace. you can also find plenty of research papers that will give you great detail on the potential applications of fullerenes.
Tarell
what is the Synthesis, properties,and applications of carbon nano chemistry
Abhijith Reply
Mostly, they use nano carbon for electronics and for materials to be strengthened.
Virgil
is Bucky paper clear?
CYNTHIA
carbon nanotubes has various application in fuel cells membrane, current research on cancer drug,and in electronics MEMS and NEMS etc
NANO
how did you get the value of 2000N.What calculations are needed to arrive at it
Smarajit Reply
Privacy Information Security Software Version 1.1a
Good
Got questions? Join the online conversation and get instant answers!
Jobilize.com Reply

Get the best Algebra and trigonometry course in your pocket!





Source:  OpenStax, Multiresolution analysis, filterbank implementation, and function approximation using wavelets. OpenStax CNX. Sep 14, 2009 Download for free at http://cnx.org/content/col10568/1.2
Google Play and the Google Play logo are trademarks of Google Inc.

Notification Switch

Would you like to follow the 'Multiresolution analysis, filterbank implementation, and function approximation using wavelets' conversation and receive update notifications?

Ask