Next: Euclidean Distance in Wavelet
Up: Introduction to Wavelet Networks
Previous: Direct Calculation of Weights
Considering the optimized family of wavelets , its closed
linear span constitutes a subspace
in the
space.
With eq. (8) any
function can be orthogonally projected into that subspace. It is
interesting to ask whether constitutes a basis, because then
the projection is unique. That this is indeed
so can be shown with induction over the number of wavelets:
Consider wavelets
, that minimize
the energy functional (2) and that form a
basis. Let us choose a new wavelet that approximates best the
residual between the function and its approximation with the first
wavelets. After optimization of the
th wavelet,
the energy functional (2) is smaller than
before (for the wavelets):
Assuming now, that
we have in particular
This again means that
which implies
This, however, contradicts the choice of
in the
optimization step, where
was selected such that
Let us call the closed linear span of
(image)
subspace. The dual wavelets
are linearly
independent, and the projection
establishes an isomorphism from
(or the image space, respectively) into
which is the space of the -vectors containing the
wavelet coefficients. This space is dual to the image
subspace and we call it the wavelet subspace.
Next: Euclidean Distance in Wavelet
Up: Introduction to Wavelet Networks
Previous: Direct Calculation of Weights
Volker Krueger
2001-05-31