[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
On interpreting views of Fourier Transforms
From: |
isoparix |
Subject: |
On interpreting views of Fourier Transforms |
Date: |
Thu, 6 Aug 2015 06:57:47 -0700 (PDT) |
Question on Fourier transforms:
If I take a simple rectangular bar image,
<http://octave.1599824.n4.nabble.com/file/n4672032/rectangle_500x7.jpg>
and engineer an FT via ImageAnalyser
(http://meesoft.logicnet.dk/Analyzer/index.htm), I get an FT
<http://octave.1599824.n4.nabble.com/file/n4672032/classic_FFT_of_bar.png>
that conforms to what I read in the literature. See for instance, page 15
of
http://www.robots.ox.ac.uk/~az/lectures/ia/lect2.pdf
<http://www.robots.ox.ac.uk/~az/lectures/ia/lect2.pdf>
Octave produces something very different.
My question is, are we all talking at cross-purposes here? I'm sure
Octave's understanding of FFT cannot be utterly different to everyone
else's. And the Octave FFT certainly, using 'ifft', transforms exactly back
to the original image, *but* why doesn't Octave's FFT look like the FFT
produced by Image Analyser, or described in the Oxford lectures in the PDF,
or in other literature?
What am I misinterpreting? Thank you.
To see what Octave does, try:
%Import image
origimage = imread('rectangle_500x7','jpg');
%Plot image
figure, imagesc(origimage)
axis square
colormap gray
title('Original Image')
%Perform 2D FFT
fftimage = fft2(double(origimage));
%Display Image of FFT
figure, imagesc(abs(fftimage))
axis square
title('FFT of Image')
colormap gray
set(gca, 'XTick', [], 'YTick', [])
%Perform Inverse 2D FFT
backimage = ifft2(fftimage);
%Display Image of FFT
figure, imagesc(abs(backimage))
axis square
title('Inverse FFT')
colormap gray
set(gca, 'XTick', [], 'YTick', [])
--
View this message in context:
http://octave.1599824.n4.nabble.com/On-interpreting-views-of-Fourier-Transforms-tp4672032.html
Sent from the Octave - General mailing list archive at Nabble.com.
- On interpreting views of Fourier Transforms,
isoparix <=