Re: Filter coefficients
- From: ImageAnalyst <imageanalyst@xxxxxxxxxxxxxx>
- Date: Mon, 09 Jul 2007 16:23:45 -0700
On Jul 8, 12:14 pm, Giper <gip...@xxxxxxx> wrote:
Dear colleagues, I have to solve one interest task.
Suppose you have source signal S.
You must decompose it to two parts: low frequency (L) and high
frequency (H).
I must to build such decomposition where the sizes of L and H may not
be equal, i.e. to get 'L' I use filter F1 with frequency band [0
0.3],
and to get 'H'I use filter F2 with frequency band [0.3 1].
Then I have to get source signal from L and H, using reconstruction
filters.
How can I build filters F1 and F2 that allows perfect
reconstruction??? Where I can found coefficients for such filters???
And how can I realize such decomposition???
I have read theory of wavelet transform. But in this theory the
source
signal S decomposed to two parts with equal lengths.
With best regards, Gipaim.
Giper/Gipaim:
I agree this sounds like a class homework assignment. Can you answer
these questions:
1. Is this question for your college class? (If so, this doesn't
prevent me from further answers or reponses, but might guide me in
making them.)
2. Is the signal 1D, 2D, or 3D?
3. Do you have to use wavelets or can you use Fourier instead?
4. What do you mean by length? This leads me to think you have a 1D
signal, like an audio signal or something. If it's an image, this
terminology is unclear. You could take the FFT of the image and then
have a filter #1 pick out an inner square and filter #2 pick out a
square annulus of the same area. Reconstruction is extremely
straightforward after this. Same concept would apply to Fourier
filtering of a 1D signal.
Regards,
ImageAnalyst
.
- Follow-Ups:
- Re: Filter coefficients
- From: Giper
- Re: Filter coefficients
- References:
- Filter coefficients
- From: Giper
- Filter coefficients
- Prev by Date: NICE.I: Noisy Iris Challenge Evaluation - Part I. (contest with Journal Special Issue)
- Next by Date: WANTED watec LCL-211H camera
- Previous by thread: Re: Filter coefficients
- Next by thread: Re: Filter coefficients
- Index(es):
Relevant Pages
|