tfmri.coils.estimate_sensitivities

estimate_sensitivities(input_, coil_axis=- 1, method='walsh', **kwargs)[source]

Estimate coil sensitivity maps.

This method supports 2D and 3D inputs.

Parameters

  • input – A Tensor. Must have type complex64 or complex128. Must have shape [height, width, coils] for 2D inputs, or [depth, height, width, coils] for 3D inputs. Alternatively, this function accepts a transposed array by setting the coil_axis argument accordingly. Inputs should be images if method is 'walsh' or 'inati', and k-space data if method is 'espirit'.

  • coil_axis – An int. Defaults to -1.

  • method – A string. The coil sensitivity estimation algorithm. Must be one of: {'walsh', 'inati', 'espirit'}. Defaults to 'walsh'.

  • **kwargs – Additional keyword arguments for the coil sensitivity estimation algorithm. See Notes.

Returns

A Tensor. Has the same type as input_. Has shape input_.shape + [num_maps] if method is 'espirit', or shape input_.shape otherwise.

Notes

This function accepts the following method-specific keyword arguments:

  • For method="walsh":

    • filter_size: An int. The size of the smoothing filter.

  • For method="inati":

    • filter_size: An int. The size of the smoothing filter.

    • max_iter: An int. The maximum number of iterations.

    • tol: A float. The convergence tolerance.

  • For method="espirit":

    • calib_size: An int or a list of ints. The size of the calibration region. If None, this is set to input_.shape[:-1] (ie, use full input for calibration). Defaults to 24.

    • kernel_size: An int or a list of ints. The kernel size. Defaults to 6.

    • num_maps: An int. The number of output maps. Defaults to 2.

    • null_threshold: A float. The threshold used to determine the size of the null-space. Defaults to 0.02.

    • eigen_threshold: A float. The threshold used to determine the locations where coil sensitivity maps should be masked out. Defaults to 0.95.

    • image_shape: A tf.TensorShape or a list of ints. The shape of the output maps. If None, this is set to input_.shape. Defaults to None.

References

1

Walsh, D.O., Gmitro, A.F. and Marcellin, M.W. (2000), Adaptive reconstruction of phased array MR imagery. Magn. Reson. Med., 43: 682-690. https://doi.org/10.1002/(SICI)1522-2594(200005)43:5<682::AID-MRM10>3.0.CO;2-G

2

Inati, S.J., Hansen, M.S. and Kellman, P. (2014). A fast optimal method for coil sensitivity estimation and adaptive coil combination for complex images. Proceedings of the 2014 Joint Annual Meeting ISMRM-ESMRMB.

3

Uecker, M., Lai, P., Murphy, M.J., Virtue, P., Elad, M., Pauly, J.M., Vasanawala, S.S. and Lustig, M. (2014), ESPIRiT—an eigenvalue approach to autocalibrating parallel MRI: Where SENSE meets GRAPPA. Magn. Reson. Med., 71: 990-1001. https://doi.org/10.1002/mrm.24751