next up previous contents index
Next: The Requirement of Commensurability Up: Multiresolution Analysis Previous: Multiscale Analysis vs Multiresolution   Contents   Index


The Pyramid Algorithm

The MSA measuring process starts with the acquisition of a signal as an element in the central (``fiducial'', ``reference'') vector space $ \mathbf{V}_0$ . This means that a signal

$\displaystyle f(t)=\sum_{\ell=-\infty}^\infty c_\ell^0\phi(t-\ell)\in \mathbf{V}_0
$

is acquired in the form of a square summable sequence of numbers

$\displaystyle \{c_\ell^0\}=
\left\{ \langle\phi(u-\ell),f(u)\rangle\right\},~~~\ell=0,\pm 1,\cdots
$

The problem is to determine the representation of the signal in each of the subsequent (``lower resolution'') approximation spaces $ \mathbf{V}_0 \supset \mathbf{V}_1\supset \mathbf{V}_2\supset \cdots$ , i.e. find

$\displaystyle \{c_\ell^1\},\{c_\ell^2\},\cdots
$

such that

$\displaystyle \sum_{\ell=-\infty}^\infty c_\ell^k\phi(t-\ell)=P_{\mathbf{V}_k}f(t)
$

is the least squares projection of $ f(t)$ onto $ \mathbf{V}_k$ for $ k=1,2,\cdots$ . This turns out to be an iterative process which terminates after a finite number of steps.

The key observation which makes this process so powerful and appealing is that the relationship between two adjacent resolution spaces, say $ \mathbf{V}_k$ and $ \mathbf{V}_{k+1}$ , is independent of the order $ k$ .

Given the fact that $ \{\mathbf{V}_k:k=0,\pm 1,\cdots\}$ is a MSA and that $ \phi(t)$ is the corresponding scaling function, we know that

Consequently, each such basis element can be expanded uniquely in terms of the $ \mathbf{V}_k$ -basis

$\displaystyle \phi(2^{-(k+1)}t-\ell')=\sum_{\ell=-\infty}^\infty \phi(2^{-k}t-\ell)~ 2^{-k}\langle\phi(2^{-k}u-\ell),\phi(2^{-(k+1)}u-\ell')\rangle$ (2108)

By changing variables in the inner product integral one finds that

$\displaystyle 2^{-k}\langle\phi(2^{-k}u-\ell),\phi(2^{-(k+1)}u-\ell')\rangle$ $\displaystyle \equiv 2^{-k}\int\limits_{-\infty}^\infty \overline{\phi}(2^{-k}u-\ell)\phi(2^{-(k+1)}u-\ell')du$    
  $\displaystyle =\int\limits_{-\infty}^\infty \overline{\phi}(u-\ell)\phi(2^{-1}u-\ell')du$    
  $\displaystyle =\int\limits_{-\infty}^\infty \overline{\phi}\left(u-(\ell-2\ell')\right)\phi(2^{-1}u)du$    
  $\displaystyle \equiv \sqrt{2}~h_{\ell-2\ell'}$    

Exercise 26.4 (ORTHONORMALITY)


a) Point out why this inner product is the $ (\ell,\ell')$ th entry of the $ \sqrt{2}$ -multiple of a unitary matrix, which is independent of $ k$ .


b) Show that $ \sum_{\ell=-\infty}^\infty \overline{h}_\ell h_{\ell-2\ell'}
=\delta_{0\ell'}$ .

When one computes the (complex) inner product of $ f$ with both sides of Eq.(2.108), one obtains

$\displaystyle \underbrace{ \langle\sqrt{2^{-(k+1)}}\phi(2^{-(k+1)}u-\ell'),f(u)\rangle }_{c_{\ell'}^{k+1}}$ $\displaystyle = \sum_{\ell=-\infty}^\infty \langle\sqrt{2^{-(k+1)}}\phi(2^{-(k+1)}u-\ell'), \sqrt{2^{-k}}\phi(2^{-k}u-\ell)\rangle$    
  $\displaystyle ~~~~~~~~~~~~\times\langle\sqrt{2^{-k}}\phi(2^{-k}u-\ell'),f(u)\rangle$    
  $\displaystyle = \sum_{\ell=-\infty}^\infty \underbrace{ \langle\sqrt{2^{-k}}\phi(2^{-k}u-\ell),f(u)\rangle }_{c_{\ell}^{k}} ~\overline{h}_{\ell-2\ell'}$    

Thus, by setting $ \overline{h}_{\ell-2\ell'}\equiv\tilde{h}_{2\ell'-\ell}$ one has

$\displaystyle \boxed{ c_{\ell'}^{k+1}=\sum_{\ell=-\infty}^\infty \tilde{h}_{2\ell'-\ell}~c_{\ell}^{k} }$ (2109)

The sum on the r.h.s. is the discrete convolution of $ \{ \tilde{h}_\ell \}$ and $ \{ c_\ell^k \}$ . It shows that the discrete approximation

$\displaystyle \{ c_\ell^{k+1} \}=
\{
\langle\sqrt{2^{-(k+1)}} \phi(2^{-(k+1)} u-\ell),f(u)\rangle
:~\ell=0,\pm 1,\cdots \}
$

of $ f$ can be computed from $ \{ c_\ell^k \}$ by convolving it with $ \{ \tilde{h}_\ell \}$ , and then keeping only every other sample from the result. Thus, if one starts out with a discrete approximation $ \{ c_\ell^k \}$ which represents $ f$ by means of a finite number of samples, then the next discrete approximation is represented by only half as many samples. After a sufficient number of such iterative steps the process stops because one has run out of samples. All successive discrete approximations of $ f$ are merely sequences of zeros. This iterative algorithm is known as the pyramid algorithm first introduced by Stephane Mallat[#!StephaneMallat1!#]. It is a rather efficient algorithm because it terminates after only

$\displaystyle \log_2(\char93 \textrm{ of sampled values of }f)~.
$

iterations.


next up previous contents index
Next: The Requirement of Commensurability Up: Multiresolution Analysis Previous: Multiscale Analysis vs Multiresolution   Contents   Index
Ulrich Gerlach 2007-04-05