Reproducing Polynomials with B-Splines: Unterschied zwischen den Versionen

Aus NOBAQ
Zur Navigation springenZur Suche springen
Zeile 1: Zeile 1:
 +
<section begin="head"/>
 
A B-Spline of order <math>N</math> is known to be able to reproduce any polynomial up to order <math>N</math><ref>I.J. Schoenberg: "Cardinal interpolation and spline functions", ''J. Approx. Theory volume 2'', pp. 167-206, 1969</ref>:
 
A B-Spline of order <math>N</math> is known to be able to reproduce any polynomial up to order <math>N</math><ref>I.J. Schoenberg: "Cardinal interpolation and spline functions", ''J. Approx. Theory volume 2'', pp. 167-206, 1969</ref>:
  
Zeile 5: Zeile 6:
 
</math>
 
</math>
  
In words, a proper linear combination of shifted versions of a B-Spline can reproduce any polynomial up to order <math>N</math>. This is needed for different applications, for example, for the Sampling at Finite Rate of Innovation (FRI) framework. In this case any kernel <math>\varphi</math> reproducing polynomials (that is, satisfying the Strang-Fix conditions) can be used. However, among all possible kernels, the B-Splines have the smallest possible support.
+
In words, a proper linear combination of shifted versions of a B-Spline can reproduce any polynomial up to order <math>N</math>. This is needed for different applications, for example, for the Sampling at Finite Rate of Innovation (FRI) framework<ref>P.L. Dragotti, M. Vetterli, T.Blu: "Sampling Moments and Reconstructing Signals of Finite Rate of Innovation: Shannon Meets Strang-Fix", ''IEEE Transactions on Signal Processing'', vol. 55, No. 5, May 2007</ref>. In this case any kernel <math>\varphi</math> reproducing polynomials (that is, satisfying the Strang-Fix conditions) can be used. However, among all possible kernels, the B-Splines have the smallest possible support.
  
 
An important question is how to obtain the coefficients <math>c_{m,n}</math> for the reproduction-formula. In this small article, I describe one way.
 
An important question is how to obtain the coefficients <math>c_{m,n}</math> for the reproduction-formula. In this small article, I describe one way.
 +
<section end="head"/>
 +
 +
Starting from
 +
 +
<math>
 +
\sum_{n \in \mathbb{Z}} c_{m,n} \varphi(t - n) = t^m
 +
</math>
 +
 +
the coefficients can be obtained using the dual of <math>\varphi</math>, <math>\tilde{\varphi}</math> (I set <math>\beta_N = \varphi</math> for consistency with my notes):
 +
 +
<math>
 +
c_{m,n} = \int_{-\infty}^{\infty} t^m \tilde{\varphi}(t - n)\,dt
 +
</math>
 +
 +
However, even if the dual would be known, solving the infinite integral is only feasible when the dual has finite support. This is the case with the B-Spline itself but not with its dual!
 +
 +
A closer look at the formula tells that this is nothing more than a convolution (under the assumption that <math>\tilde{\varphi}</math> is symmetric which is the case):
 +
 +
<math>
 +
c_{m,n} = \int t^m \tilde{\varphi}(-(n-t))\,dt = \int t^m \tilde{\varphi}(n-t)\,dt = (t^m * \tilde{\varphi})(n)
 +
</math>
 +
 +
Now, this can be transformed to fourier domain:
 +
 +
<math>
 +
(t^m * \tilde{\varphi})(n) = \mathcal{F}^{-1}\left\{ \mathcal{F}\left\{t^m\right\} \tilde{\Phi}(\omega)\right\} =  \mathcal{F}^{-1}\left\{ j^m \sqrt{2\pi} \delta^{(n)}(\omega) \tilde{\Phi}(\omega) \right\} = j^m \sqrt{2\pi} \mathcal{F}^{-1}\left\{\delta^{(n)}(\omega) \tilde{\Phi}(\omega) \right\}
 +
</math>
 +
 +
Writing the inverse of this expression yields:
 +
 +
<math>
 +
j^m \sqrt{2\pi} \frac{1}{\sqrt{2\pi}} \int_{-\pi}^{\pi} \delta^{(n)}(\omega) \tilde{\Phi}(\omega) e^{j\omega n}\,d\omega = j^m \int_{-\infty}^{\infty} \delta^{(n)}(\omega) \underbrace{\tilde{\Phi}(\omega) e^{j\omega n}}_{f(\omega)}\,d\omega
 +
</math>
 +
 +
It is known that<ref>http://en.wikipedia.org/wiki/Dirac_delta_function</ref>:
 +
 +
<math>
 +
\int \delta^{(n)}(x) f(x)\,dx = (-1)^n f^{(n)}(0)
 +
</math>
 +
 +
so that
 +
 +
<math>
 +
j^m \int_{-\infty}^{\infty} \delta^{(n)}(\omega) f(\omega)\,d\omega = j^m (-1)^m \left. \frac{\partial^m}{\partial \omega^m} f(\omega) \right|_{\omega = 0}
 +
</math>
 +
 +
Now the whole procedure has been reduced to calculate the derivative of <math>f(\omega)</math> and set the result to zero.
 +
 +
An open question is how to obtain the dual of <math>\varphi</math>. As the reproduction formula spans a vector space, the <math>\varphi</math> must be at least bi-orthogonal to <math>\tilde{\varphi}</math>. This translates in fourier domain to<ref>S. Mallat: "A Wavelet Tour of Signal Processing", ''Academic Press'' 1999</ref>:
 +
 +
<math>
 +
\tilde{\Phi}(\omega) = \frac{\Phi(\omega)}{\sum_{k \in \mathbb{Z}} |\Phi(\omega + 2\pi k)|^2}
 +
</math>
 +
 +
The fourier transform of a B-Spline of order <math>N</math> is (e.g. <ref>M.Unser: "Splines - A Perfect Fit for Signal and Imaging Processing", ''IEEE Signal Processing Magazine'' Nov. 1999</ref>):
 +
 +
<math>
 +
\Beta_N(\omega) = \Phi(\omega) = \left( \frac{\sin(\omega/2)}{\omega/2} \right)^{N+1} =
 +
\mathrm{sinc}^{N+1}(\omega/2)
 +
</math>
 +
 +
The following derivation of the sum is borrowed from <ref>M.J.C.S. Reis, P.J.S.G. Ferreira, S.F.S.P. Soares: "Linear combinations of B-splines as generating functions for signal approximation", ''Elsevier Digital Signal Processing 15'', 2005</ref>. For this derivation to work, I set <math>L=N+1</math> temprarily:
 +
 +
<math>
 +
\sum_{k \in \mathbb{Z}} |\Phi(\omega + 2\pi k)|^2 =
 +
\sum_{k \in \mathbb{Z}} \left|\mathrm{sinc}\left(\frac{1}{2}(\omega + 2\pi k)\right)^L \right|^2 =
 +
\sum_{k \in \mathbb{Z}} \left|\mathrm{sinc}\left(\frac{1}{2}(\omega + 2\pi k)\right) \right|^{2L}
 +
</math>
 +
 +
and because <math>2L</math> is always even:
 +
 +
<math>
 +
= \sum_{k \in \mathbb{Z}}\frac{\sin^{2L}(\frac{1}{2}(\omega + 2\pi k))}{\left(\frac{1}{2}(\omega + 2\pi k)\right)^{2L}}
 +
= \sum_{k \in \mathbb{Z}}\frac{\sin^{2L}(\frac{\omega}{2} + \pi k))}{(\frac{\omega}{2} + \pi k)^{2L}}
 +
</math>
 +
 +
Because of the periodicity it is known that
 +
 +
<math>
 +
\sin^{2L}(x + \pi k) = \sin^{2L}(x)
 +
</math>
 +
 +
such that
 +
 +
<math>
 +
= \sin^{2L}(\frac{\omega}{2}) \sum_{k \in \mathbb{Z}}\frac{1}{(\frac{\omega}{2} + \pi k)^{2L}}
 +
</math>
 +
 +
And finally the following relation is used:
 +
 +
<math>
 +
\sum_k \frac{1}{(x + \pi k)^{2L}} = -\frac{1}{(2L-1)!} \frac{d^{2L-1}}{dx^{2L-1}} \cot{x}
 +
</math>
 +
 +
in order to finally obtain:
 +
 +
<math>
 +
\sum_{k \in \mathbb{Z}} \left|\mathrm{sinc}\left(\frac{1}{2}(\omega + 2\pi k)\right)^L \right|^2 =
 +
-\sin^{2L}\left(\frac{\omega}{2}\right) \frac{1}{(2L-1)!} \frac{d^{2L-1}}{d\left(\frac{\omega}{2}\right)^{2L-1}} \cot{\left(\frac{\omega}{2}\right)}
 +
</math>
 +
 +
and with <math>L = N+1</math>:
 +
 +
<math>
 +
\sum_{k \in \mathbb{Z}} |\Phi(\omega + 2\pi k)|^2 =
 +
-\sin^{2(N+1)}\left(\frac{\omega}{2}\right) \frac{1}{(2N+1)!} \frac{d^{2N+1}}{d\left(\frac{\omega}{2}\right)^{2N+1}} \cot{\left(\frac{\omega}{2}\right)}
 +
</math>
 +
 +
  
 
<references/>
 
<references/>

Version vom 19. Juli 2010, 15:33 Uhr

A B-Spline of order is known to be able to reproduce any polynomial up to order [1]:

In words, a proper linear combination of shifted versions of a B-Spline can reproduce any polynomial up to order . This is needed for different applications, for example, for the Sampling at Finite Rate of Innovation (FRI) framework[2]. In this case any kernel reproducing polynomials (that is, satisfying the Strang-Fix conditions) can be used. However, among all possible kernels, the B-Splines have the smallest possible support.

An important question is how to obtain the coefficients for the reproduction-formula. In this small article, I describe one way.


Starting from

the coefficients can be obtained using the dual of , (I set for consistency with my notes):

However, even if the dual would be known, solving the infinite integral is only feasible when the dual has finite support. This is the case with the B-Spline itself but not with its dual!

A closer look at the formula tells that this is nothing more than a convolution (under the assumption that is symmetric which is the case):

Now, this can be transformed to fourier domain:

Writing the inverse of this expression yields:

It is known that[3]:

so that

Now the whole procedure has been reduced to calculate the derivative of and set the result to zero.

An open question is how to obtain the dual of . As the reproduction formula spans a vector space, the must be at least bi-orthogonal to . This translates in fourier domain to[4]:

The fourier transform of a B-Spline of order is (e.g. [5]):

The following derivation of the sum is borrowed from [6]. For this derivation to work, I set temprarily:

and because is always even:

Because of the periodicity it is known that

such that

And finally the following relation is used:

in order to finally obtain:

and with :


  1. I.J. Schoenberg: "Cardinal interpolation and spline functions", J. Approx. Theory volume 2, pp. 167-206, 1969
  2. P.L. Dragotti, M. Vetterli, T.Blu: "Sampling Moments and Reconstructing Signals of Finite Rate of Innovation: Shannon Meets Strang-Fix", IEEE Transactions on Signal Processing, vol. 55, No. 5, May 2007
  3. http://en.wikipedia.org/wiki/Dirac_delta_function
  4. S. Mallat: "A Wavelet Tour of Signal Processing", Academic Press 1999
  5. M.Unser: "Splines - A Perfect Fit for Signal and Imaging Processing", IEEE Signal Processing Magazine Nov. 1999
  6. M.J.C.S. Reis, P.J.S.G. Ferreira, S.F.S.P. Soares: "Linear combinations of B-splines as generating functions for signal approximation", Elsevier Digital Signal Processing 15, 2005