sliding FFT

[fabrice senger-2]

Dear all,

I'm analysing periodic structures using FFT (actually to extract the period)
and would like to use a sliding FFT approach...Has allready someone
developped such an approach ?
The objects I'm looking at are filaments wich might be straight or curved.
By now and just to validate the approach, I was cutting my structures in
little pieces but now it would be great if I could automatize it ...

Thank you,

Fabrice.

[Robert Dougherty]

Fabrice,

Bill Mohler asked me to write a plugin for a similar problem in 2005.  I applied the Helmholtz equation to a filtered version of the image to estimate the wavenumber and therefore the wavelength, or period. The plugin was called Muscle Tone to reflect the application.  I just posted the code on the OptiNav site, giving it the new name Helmholtz Analysis. It is the not the FFT approach, but I think it is actually better for estimating a single period for each region of an image. It includes additional processing to give the orientation of the structures, in terms of an angle, and locally averaged values of both the wavelength and the angle.

http://www.optinav.com/Helmholtz_Analysis.htm

Bob

On Oct 11, 2011, at 8:31 AM, fabrice senger wrote:

> Dear all,
>
> I'm analysing periodic structures using FFT (actually to extract the period)
> and would like to use a sliding FFT approach...Has allready someone
> developped such an approach ?
> The objects I'm looking at are filaments wich might be straight or curved.
> By now and just to validate the approach, I was cutting my structures in
> little pieces but now it would be great if I could automatize it ...
>
> Thank you,
>
> Fabrice.

Robert Dougherty, Ph.D.
President, OptiNav, Inc.
1414 127th Place NE #106
Bellevue, WA 98005
[phone]
FAX [phone]
www.optinav.com
[hidden email]

[fabrice senger-2]

Hello Bob,

i gave this approach a try and I could not really understand the output of
the analysis.
Unfortunatly I did not get acces to the original paper.
Would it be possible to get more info about that ? Should I go for a search
on Helmholtz ?
Thank you for any advice,

Fabrice.

2011/10/12 Robert Dougherty <[hidden email]>

> Fabrice,
>
> Bill Mohler asked me to write a plugin for a similar problem in 2005.  I
> applied the Helmholtz equation to a filtered version of the image to
> estimate the wavenumber and therefore the wavelength, or period. The plugin
> was called Muscle Tone to reflect the application.  I just posted the code
> on the OptiNav site, giving it the new name Helmholtz Analysis. It is the
> not the FFT approach, but I think it is actually better for estimating a
> single period for each region of an image. It includes additional processing
> to give the orientation of the structures, in terms of an angle, and locally
> averaged values of both the wavelength and the angle.
>
> http://www.optinav.com/Helmholtz_Analysis.htm
>
> Bob
>
> On Oct 11, 2011, at 8:31 AM, fabrice senger wrote:
>
> > Dear all,
> >
> > I'm analysing periodic structures using FFT (actually to extract the
> period)
> > and would like to use a sliding FFT approach...Has allready someone
> > developped such an approach ?
> > The objects I'm looking at are filaments wich might be straight or
> curved.
> > By now and just to validate the approach, I was cutting my structures in
> > little pieces but now it would be great if I could automatize it ...
> >
> > Thank you,
> >
> > Fabrice.
>
> Robert Dougherty, Ph.D.
> President, OptiNav, Inc.
> 1414 127th Place NE #106
> Bellevue, WA 98005
> [phone]
> FAX [phone]
> www.optinav.com
> [hidden email]
>

[Robert Dougherty]

Fabrice,

On Oct 14, 2011, at 12:20 AM, fabrice senger wrote:

> Hello Bob,
>
> i gave this approach a try and I could not really understand the output of
> the analysis.

Point taken.


> Unfortunatly I did not get acces to the original paper.
> Would it be possible to get more info about that ? Should I go for a search
> on Helmholtz ?

You can search for Helmholtz if you want to. He is a very important figure in acoustics, and apparently also several other fields. You won't find much about using the Helmholtz equation in image processing, since I think the idea is novel.  I created a writeup to try to explain it: http://www.optinav.com/helmholtz.pdf.  It is now linked from the web page about the plugin.

Bob

> Thank you for any advice,
>
> Fabrice.
>
> 2011/10/12 Robert Dougherty <[hidden email]>
>
>> Fabrice,
>>
>> Bill Mohler asked me to write a plugin for a similar problem in 2005.  I
>> applied the Helmholtz equation to a filtered version of the image to
>> estimate the wavenumber and therefore the wavelength, or period. The plugin
>> was called Muscle Tone to reflect the application.  I just posted the code
>> on the OptiNav site, giving it the new name Helmholtz Analysis. It is the
>> not the FFT approach, but I think it is actually better for estimating a
>> single period for each region of an image. It includes additional processing
>> to give the orientation of the structures, in terms of an angle, and locally
>> averaged values of both the wavelength and the angle.
>>
>> http://www.optinav.com/Helmholtz_Analysis.htm
>>
>> Bob
>>
>> On Oct 11, 2011, at 8:31 AM, fabrice senger wrote:
>>
>>> Dear all,
>>>
>>> I'm analysing periodic structures using FFT (actually to extract the
>> period)
>>> and would like to use a sliding FFT approach...Has allready someone
>>> developped such an approach ?
>>> The objects I'm looking at are filaments wich might be straight or
>> curved.
>>> By now and just to validate the approach, I was cutting my structures in
>>> little pieces but now it would be great if I could automatize it ...
>>>
>>> Thank you,
>>>
>>> Fabrice.
>>
>> Robert Dougherty, Ph.D.
>> President, OptiNav, Inc.
>> 1414 127th Place NE #106
>> Bellevue, WA 98005
>> [phone]
>> FAX [phone]
>> www.optinav.com
>> [hidden email]
>>

Robert Dougherty, Ph.D.
President, OptiNav, Inc.
1414 127th Place NE #106
Bellevue, WA 98005
Tel. [phone]
FAX [phone]
www.optinav.com
[hidden email]

[Robert Dougherty]

The Helmholtz Analysis plugin has been significantly updated to improve the user interface and the accuracy. I have not yet edited helmholtz.pdf  to reflect the changes.
http://www.optinav.com/Helmholtz_Analysis.htm