Hi Jerome,
Interesting analysis problem. MATLAB with all its signal processing
tools should be adequate.
Just to be sure that I understood your goal correctly, I try to rephrase it:
" Goal is to see temporal frequencies at which the cilia beat.
Therefore, you want to see temporal frequencies observed in each pixel
of the field of images of cilia. The temporal frequency of a given
pixel is a good indication of frequency with which 'cilia crosses'
that pixel. "
Following is a pseudo-code in MATLAB, which may work.
for i=1:noSlices
timeseq(:,:,i)=imread([fileprefixstring num2str(i)]); %Read files in
a 3D matrix. % XY represent space, Z represents time.
end
% Obtain the spectrum along the Z-dimension.
centershiftedtimeseq=ifftshift(timeseq,3); %second argument specifies
shift along 3rd dim.
centershiftedspectrum=fft(timeseq,3); %Take 1D FFT along 3rd dim.
recenteredspectrum=fftshift(centershiftedspectrum,3); % Recenter the spectrum.
The last 3D matrix contains the magnitude and phase of the spectrum
along each XY position in field of view. You may wonder why bother
with ifftshift and fftshift. I have some thoughts on that on this
post:
http://shalin.wordpress.com/2009/12/06/fftifft/hope this works,
Shalin
mobile: +65-90694182
blog: shalin.wordpress.com
Bioimaging Lab, Block-E3A, #7-10
Div of Bioengineering, NUS Singapore 117574
website:
http://www.bioeng.nus.edu.sg/optbioimaging/colin/people.asp#shalinmOn Sat, Feb 6, 2010 at 11:52 AM, Jerome Carpenter <
[hidden email]> wrote:
> Hello all,
> I have an image sequence and would like to take a Fourier Transform of each
> pixel in time (through the sequence). The goal is to generate frequency and
> phase maps of a field of beating cilia. I'm currently using ImageJ and the
> time series analyzer to get average intensity over ROIs, and Matlab for the
> fourier transform, but would like to go pixel by pixel. Any advice on image
> analysis packages/plugins to accomplish this?
> Thanks!
> Jerome
Locked
