I have data that is ordered as follows
[ z, y, x, i, j ]
where [z,y,x]
represent physical dimensions and [i,j]
represent tensor dimensions. I.e. it stores a second-order tensor at points in grid with shape [z,y,x]
.
I would like to calculate the Fourier transform of each tensor component [i,j]
along the physical dimensions [z,y,x]
, using FFTW in C.
The NumPy-equivalent of this operation would be
np.fft.fftn(a, axes=(0, 1, 2))
Neither the documentation nor Google have provided the answer.