mri.optimizers.utils.reweight#

Reweighting optimisation strategies.

class mReweight(weights, linear_op, thresh_factor=1)[source]#

Bases: object

Ming reweighting.

This class implements the reweighting scheme described in Ming2017.

Parameters

  • weights (numpy.ndarray) – Array of weights

  • linear_op (pysap.numeric.linear.Wavelet) – A linear operator.

  • thresh_factor (float, default 1) – Threshold factor: sigma threshold.

reweight(x_new)[source]#

Update the weights.

Parameters

x_new (numpy.ndarray) – the current primal solution.

Returns

sigma_est – the variance estimate on each scale.

Return type

numpy.ndarray