nifty7.library.wiener_filter_curvature module

WienerFilterCurvature(R, N, S, iteration_controller=None, iteration_controller_sampling=None, data_sampling_dtype=<class 'numpy.float64'>, prior_sampling_dtype=<class 'numpy.float64'>)[source]

The curvature of the WienerFilterEnergy.

This operator implements the second derivative of the WienerFilterEnergy used in some minimization algorithms or for error estimates of the posterior maps. It is the inverse of the propagator operator.

Parameters
  • R (LinearOperator) – The response operator of the Wiener filter measurement.

  • N (EndomorphicOperator) – The noise covariance.

  • S (EndomorphicOperator) – The prior signal covariance.

  • iteration_controller (IterationController) – The iteration controller to use during numerical inversion via ConjugateGradient.

  • iteration_controller_sampling (IterationController) – The iteration controller to use for sampling.

  • data_sampling_dtype (numpy.dtype or dict of numpy.dtype) – Data type used for sampling from likelihood. Conincides with the data type of the data used in the inference problem. Default is float64.

  • prior_sampling_dtype (numpy.dtype or dict of numpy.dtype) – Data type used for sampling from likelihood. Coincides with the data type of the parameters of the forward model used for the inference problem. Default is float64.

Note

It must be possible to set the sampling dtype of N and S with the help of an SamplingDtypeSetter. In practice this means that data_sampling_dtype is not None, N must be a ScalingOperator, a DiagonalOperator, or something similar.