On Sunday 12 June 2011 07:19 AM,
[hidden email] wrote:
> Dear Divakar,
>
> I am Bhuvnesh, doing my PhD at TU Berlin in the field of nanochemisry.
> I saw your various informative posts on imageJ forum and wanted to contact you and discuss my problem with you.
>
> I am new at using imageJ, so don't know much details about it hence seeking your help.
> Actually I am analyzing some microscope images, and I am interested in extracting the scattering profiles of the particles (in the image).
> I try to read many of the earlier discussion but being a chemist I don't know much about the physics and maths of how to convert 2D-spatial image to 1D- inverse space image.
> I tried to do the FFT of the image and then took the radial profile of the FFT, although it looks like the scattering profile of the spherical particles I am using but I am not sure what is on X and Y axis. Moreover I don't know if this is the exact way of doing it.
I am not sure here what exactly you mean by scattering profiles. If you
are looking for periodicities in the spatial distribution of an image of
a bunch of particles, the FFT is what you require. If not, what the FFT
shows will largely depend on the contrast mechanism used to record the
image. In the FFT, the x and y directions are the reciprocal space axes
corresponding to the real space x and y directions. The scaling along
these axes in the FFT is inverse wrt the image space That is, if you see
a peak in the FFT and measure its distance from the centre of the FFT as
R, then the periodicity in the image space that gives rise to it is C/R
where C is a constant, which in this case will depend on the spatial
extent of the original image. If your image is calibrated in ImageJ,
then the FFT is also suitably calibrated such that when you move the
cursor over the FFT, the spatial distance is directly displayed in the
ImageJ status bar.
HTH,
Divakar
PS: It would be a good idea to post your questions directly on the
ImageJ mailing list. Advantages are: (i) better informed people or
people with better ideas / links can respond, and (ii) the discussion
that emerges would be of use to others with similar problems.
> I would be thankful to you if you can take out some time and help me out in this matter.
>
> Looking forward for your suggestions.
>
> Many regards
>
> Bhuvnesh Bharti
Locked
