bartorch.nufft_adjoint

Contents

bartorch.nufft_adjoint#

bartorch.nufft_adjoint(input, traj, image_shape=None, *, weights=None, basis=None, **extra)#

Adjoint of nufft(): samples along traj back to an image.

Parameters:

  • input (torch.Tensor) – Samples, as nufft() returns them.

  • traj (torch.Tensor) – Trajectory of shape (..., samples, 3) in grid units.

  • image_shape (tuple of int, optional) – Spatial shape of the image, C order (z, y, x) or (y, x) (-d). By default BART estimates it from the trajectory.

  • weights (torch.Tensor, optional) – Diagonal applied to the samples (-p); its conjugate is applied here.

  • basis (torch.Tensor, optional) – Subspace basis over frames and coefficients (-B).

  • **extra – Further bart nufft options, by name.

Contents