gradient_unwarped

pulserver.recon.gradient_unwarped()

Return an image resampled from where its voxels were acquired to where they belong.

The coil’s spherical-harmonic field is evaluated at the scanner position of every voxel of the reconstruction matrix, and the image is resampled there. A complex image has its real and imaginary parts resampled separately.

Geometry comes from the header’s reconstruction space and from the unit’s reference acquisition, so the correction is in the frame the series was acquired in, oblique prescriptions included.

Parameters:

• image – (..., [partitions,] lines, columns), NumPy array or Torch tensor.

• context – The plugin’s ReconContext, whose header states the table.

• data – The plugin’s ReconData, whose buffer states the geometry.

Returns:

• The corrected image, of the type and on the device it was given in.

• image itself where the header states no table, where the unit has no

• reference acquisition, or where the correction fails (an uncorrected image)

• reaches the scanner, and the reason is logged.