Got Questions? Get Answers.
Discover MakerZone

MATLAB and Simulink resources for Arduino, LEGO, and Raspberry Pi

Learn more

Discover what MATLAB® can do for your career.

Opportunities for recent engineering grads.

Apply Today

Thread Subject:
Wavelet based denoising

Subject: Wavelet based denoising

From: chourasiav@gmail.com

Date: 25 Aug, 2008 08:28:08

Message: 1 of 7

Hi, I am doing a work on wavelet based denoising of fetal heart sound
signals. I want to know that how to choose the coefficients while
reconstruction of original signal. Also how to choose the level of
thresholding,

Subject: Wavelet based denoising

From: NZTideMan

Date: 25 Aug, 2008 19:54:08

Message: 2 of 7

On Aug 25, 8:28=A0pm, chouras...@gmail.com wrote:
> Hi, I am doing a work on wavelet based denoising of fetal heart sound
> signals. I want to know that how to choose the coefficients while
> reconstruction of original signal. Also how to choose the level of
> thresholding,

There's no unique answer to your question. The optimum denoising
algorithm depends on your data.
You will need to do some reading, then some experimentation.

Here's an elegant paper that has some things you may like to try:
Katul, G., and Vidakovic, B. (1996). =93The partitioning of attached and
detached eddy motion in the atmospheric surface layer using Lorentz
wavelet filtering.=94 Boundary-Layer Meteorology, 77,: 153-172.

Subject: Wavelet based denoising

From: Chaos

Date: 3 Apr, 2009 16:54:01

Message: 3 of 7

chourasiav@gmail.com wrote in message <c2af553d-6f83-4e6f-9578-e9bea69b29cb@l33g2000pri.googlegroups.com>...
> Hi, I am doing a work on wavelet based denoising of fetal heart sound
> signals. I want to know that how to choose the coefficients while
> reconstruction of original signal. Also how to choose the level of
> thresholding,

i don't like wavelets, they ring more that the book "For Whom the Bell Tolls". you can't just plugin and let run like an fft or dct. if you're lucky with the right data set with right parameters, you might get something useful.

Subject: Wavelet based denoising

From: ImageAnalyst

Date: 3 Apr, 2009 17:11:55

Message: 4 of 7

On Apr 3, 12:54=A0pm, "Chaos" <rothko....@gmail.com> wrote:
> chouras...@gmail.com wrote in message <c2af553d-6f83-4e6f-9578-e9bea69b2.=
..@l33g2000pri.googlegroups.com>...
> > Hi, I am doing a work on wavelet based denoising of fetal heart sound
> > signals. I want to know that how to choose the coefficients while
> > reconstruction of original signal. Also how to choose the level of
> > thresholding,
>
> i don't like wavelets, they ring more that the book "For Whom the Bell To=
lls". you can't just plugin and let run like an fft or dct. =A0if you're lu=
cky with the right data set with right parameters, you might get something =
useful.

----------------------------------------------
I don't know if the original poster is still looking at this after 7
months, but one of the best noise reduction methods found so far uses
wavelets:
http://www.cs.tut.fi/~foi/GCF-BM3D/

Subject: Wavelet based denoising

From: Chaos

Date: 3 Apr, 2009 17:44:01

Message: 5 of 7

ImageAnalyst <imageanalyst@mailinator.com> wrote in message <59a2a512-5cb1-4793-9de7-df5615c8f490@e38g2000yqa.googlegroups.com>...
> On Apr 3, 12:54=A0pm, "Chaos" <rothko....@gmail.com> wrote:
> > chouras...@gmail.com wrote in message <c2af553d-6f83-4e6f-9578-e9bea69b2.=
> ..@l33g2000pri.googlegroups.com>...
> > > Hi, I am doing a work on wavelet based denoising of fetal heart sound
> > > signals. I want to know that how to choose the coefficients while
> > > reconstruction of original signal. Also how to choose the level of
> > > thresholding,
> >
> > i don't like wavelets, they ring more that the book "For Whom the Bell To=
> lls". you can't just plugin and let run like an fft or dct. =A0if you're lu=
> cky with the right data set with right parameters, you might get something =
> useful.
>
> ----------------------------------------------
> I don't know if the original poster is still looking at this after 7
> months, but one of the best noise reduction methods found so far uses
> wavelets:
> http://www.cs.tut.fi/~foi/GCF-BM3D/

this method relies on how much 'self-similar' objects can be found and stacked up, if you have a image that supports that kind of 'self-similar' object, this code works very well, i tried it. it doesn't not work well with very large images super 4k x 6k, the space is too large and you'll die of old age before you a meaningful result.

Subject: Wavelet based denoising

From: ImageAnalyst

Date: 3 Apr, 2009 20:06:50

Message: 6 of 7

On Apr 3, 1:44=A0pm, "Chaos" <rothko....@gmail.com> wrote:
> ImageAnalyst <imageanal...@mailinator.com> wrote in message <59a2a512-5cb=
1-4793-9de7-df5615c8f...@e38g2000yqa.googlegroups.com>...
> > On Apr 3, 12:54=3DA0pm, "Chaos" <rothko....@gmail.com> wrote:
> > > chouras...@gmail.com wrote in message <c2af553d-6f83-4e6f-9578-e9bea6=
9b2.=3D
> > ....@l33g2000pri.googlegroups.com>...
> > > > Hi, I am doing a work on wavelet based denoising of fetal heart sou=
nd
> > > > signals. I want to know that how to choose the coefficients while
> > > > reconstruction of original signal. Also how to choose the level of
> > > > thresholding,
>
> > > i don't like wavelets, they ring more that the book "For Whom the Bel=
l To=3D
> > lls". you can't just plugin and let run like an fft or dct. =3DA0if you=
're lu=3D
> > cky with the right data set with right parameters, you might get someth=
ing =3D
> > useful.
>
> > ----------------------------------------------
> > I don't know if the original poster is still looking at this after 7
> > months, but one of the best noise reduction methods found so far uses
> > wavelets:
> >http://www.cs.tut.fi/~foi/GCF-BM3D/
>
> this method relies on how much 'self-similar' objects can be found and st=
acked up, if you have a image that supports that kind of 'self-similar' obj=
ect, this code works very well, i tried it. =A0it doesn't not work well wit=
h very large images super 4k x 6k, the space is too large and you'll die of=
 old age before you a meaningful result. =A0- Hide quoted text -
>
> - Show quoted text -

---------------------------------------------------------------------------=
-------------------------------------------------------
Chaos:
You're probably right. It may well be slow for large images, and may
work best for 'self-similar' scenes. It's good that you pointed that
out. However I do know that many people are very impressed with its
performance and are using it as a benchmark when they try to devise a
better method. Unfortunately I don't have the required wavelet
toolbox and so I couldn't try the MATLAB code they posted. Maybe some
day...
ImageAnalyst

Subject: Wavelet based denoising

From: Chaos

Date: 3 Apr, 2009 21:17:01

Message: 7 of 7


> -------------------------------------------------------
> Chaos:
> You're probably right. It may well be slow for large images, and may
> work best for 'self-similar' scenes. It's good that you pointed that
> out. However I do know that many people are very impressed with its
> performance and are using it as a benchmark when they try to devise a
> better method. Unfortunately I don't have the required wavelet
> toolbox and so I couldn't try the MATLAB code they posted. Maybe some
> day...
> ImageAnalyst

ImageAnalyst

the theory is very sound, it is actually a technique that astro photographers have been using for years, they stack hundreds of images to reduce noise, but when you do that you are basically crushing out time, so any movement in the frame will not be handled. so if the blurr/motion from something or someone moving is supposed to be in frame the codes goes off into the 'complex' wilderness.

it performs very well if the image is fairly small, uniform and 8bit. i threw it against scenes from death valley with lots of rocks, sand dunes, mountains, actors. the program basically got 'lost' in the noise of one rock looking like another. i think i need to develop a theory of 'big uint16' images, it's a completely different world. that extra byte really kills lots of programs efficiency, they are used to searching a very small binned space. it's like comparing Mount Everest to Olympus Mons. even on static normal scene with structure it got 'lost'.

you can get most of the wavelet tools from Rice, search for Wavelab 8.5.

Chaos

Tags for this Thread

No tags are associated with this thread.

What are tags?

A tag is like a keyword or category label associated with each thread. Tags make it easier for you to find threads of interest.

Anyone can tag a thread. Tags are public and visible to everyone.

Contact us