« previous | Friday, March 28, 2014 | next »
Last Time
Haar Multi-Resolution Analysis (MRA)
Properties
Nesting
Note: , — there are functions in that are not in .
Density
Separation
Scaling
if and only if
Orthonormal Basis
is an orthonormal basis for
Haar Function Decomposition
Two Bases
, is orthogonal
can be defined as
(where)
Haar:
We know . What is ?
Question: How are the 's and 's related?
Theorem. is the average of its "double" term and the odd that follows it:
Proof. In the arbitary sense, projection in any vector space is given by:
In this case,
Recall the scaling relation:
Plugging this back into our integral gives
quod erat demonstrandum
Example
What's ?
- for or
Hence
- if , then
- if , then
- If , then
- if , then
Then the function is given by
is just the projection of onto the space
Wavelet Spaces
and
We saw that ,given , we can find
We know is orthogonal to (it cannot possibly be in unless )
What is ?
is an open secret... it stands for "wavelet"
Let
What is a basis for ?
We'll work with , ,
this means that
, BUT . Therefore
Now and .
This is our wavelet ! (thunderous applause)
We can show that is an orthonormal basis for .
, then
where is the level.