Repository navigation
Fast polar Fourier Transform (FRFT)
This is Java/ImageJ implementation of the paper:
"Fast and accurate Polar Fourier transform" by A.Averbuch, R.R.Coifman, D.L.Donoho, M.Elad, M.Israeli,
based on the published Matlab code.
This command works on one image. If the input image is not a square, it would be padded to a square shape that has even number of rows and columns (N x N). It outputs two images (2 N x 2 N):
- Polar FFT complex and imaginary coefficients;
- Power and phase.
In the output image, the Y-axis represents the radius of the polar coordinates. The radius is equal to zero in the middle of the image, basically it changes from -N to N-1.
The X-axis represents the polar angle changing from 0 to π (pi). The zero angle is defined as 45 degrees angle made by drawing a line from top left corner of the original image (1,1) to the bottom right (N,N) rotating counterclockwise.
See the illustrating image below (left-original image; right-power spectrum).
