ptyrad.utils.math_ops

Math operations of FFTs, masks, affine transformation matrices, etc.

Functions

approx_torch_quantile(t, q[, sample_size])	Approximated quantile to prevent the 2^24 element (roughly 16.7M) limitation of torch.quantile as of now.
compose_affine_matrix(scale, asymmetry, ...)
decompose_affine_matrix(input_affine_mat)
exponential_decay(r, a, b)
fftshift2(x)	A wrapper over torch.fft.fftshift for the last 2 dims
get_decomposed_affine_matrix_from_bases(...)	Fit the affine matrix components from input and output matrices A and B
ifftshift2(x)	A wrapper over torch.fft.ifftshift for the last 2 dims
make_gaussian_mask(Npix, radius, std[, center])	Create a 2D Gaussian-blurred circular mask.
make_sigmoid_mask(Npix[, relative_radius, ...])	Create a 2D circular mask with a sigmoid transition.
mfft2(im)
power_law(r, a, b)
torch_phasor(phase)	Creates a complex tensor with unit magnitude using the phase.