tfmri.recon.grappa

grappa(kspace, mask, calib, kernel_size=5, weights_l2_regularizer=0.0, combine_coils=True, sensitivities=None, return_kspace=False)[source]

Reconstructs an MR image using GRAPPA.

Parameters

  • kspace – A Tensor. The k-space samples. Must have type complex64 or complex128. Must have shape [..., C, *K], where K is the shape of the spatial frequency dimensions, C is the number of coils and ... is the batch shape, which can have any rank. Note that K is the reduced or undersampled shape.

  • mask –

    A Tensor. The sampling mask. Must have type bool. Must have shape S, where S is the shape of the spatial dimensions. In other words, mask should have the shape of a fully sampled k-space. For each point, mask should be True if the corresponding k-space sample was measured and False otherwise. True entries should correspond to the data in kspace, and the result of dropping all False entries from mask should have shape K.

  • calib – A Tensor. The calibration data. Must have type complex64 or complex128. Must have shape [..., C, *R], where R is the shape of the calibration region, C is the number of coils and ... is the batch shape, which can have any rank and must be broadcastable to the batch shape of kspace. calib is required when method is "grappa". For other methods, this parameter is not relevant.

  • kernel_size –

    An int or list of ints. The size of the GRAPPA kernel. Must have length equal to the image rank or number of spatial dimensions. If a scalar int is provided, the same size is used in all dimensions.

  • weights_l2_regularizer – An optional float. The regularization factor for the L2 regularization term used to fit the GRAPPA weights. If 0.0, no regularization is applied.

  • combine_coils –

    An optional bool. If True, multi-coil images are combined. Otherwise, the uncombined images are returned. Defaults to True.

  • sensitivities – A Tensor. The coil sensitivity maps. Must have type complex64 or complex128. Must have shape [..., C, *S], where S is shape of the spatial dimensions, C is the number of coils and ... is the batch shape, which can have any rank and must be broadcastable to the batch shape of kspace. Note that sensitivities are not used for the GRAPPA computation, but they are used for adaptive coil combination. If sensitivities are not provided, coil combination will be performed using the sum of squares method.

  • return_kspace –

    An optional bool. If True, returns the filled k-space without performing the Fourier transform. In this case, coils are not combined regardless of the value of combine_coils.

Returns

A Tensor. The reconstructed images. Has the same type as kspace. Has shape [..., S], where ... is the reconstruction batch shape and S is the spatial shape.

References

1

Griswold, M.A., Jakob, P.M., Heidemann, R.M., Nittka, M., Jellus, V., Wang, J., Kiefer, B. and Haase, A. (2002), Generalized autocalibrating partially parallel acquisitions (GRAPPA). Magn. Reson. Med., 47: 1202-1210. https://doi.org/10.1002/mrm.10171