spyrit.core.prep.UnsplitRescaleEstim.estim_alpha

UnsplitRescaleEstim.estim_alpha(y: tensor) → tensor

Estimate the gain from the raw measurements.

This calls either pinv_estim() or mean_estim() depending on self.estim_mode.

Args:

y (torch.tensor): The measurements with shape \((*, 2M)\), where \(*\) can be any number of dimensions and \(M\) matches the number of measurements defined by meas_op.M.

Returns:

torch.tensor: The estimated gain value of shape \((*, 1)\).