spyrit.core.prep.UnsplitRescaleEstim.forward

UnsplitRescaleEstim.forward(y: tensor) → tensor

Unsplit and rescale raw measurements

\[m = \frac{y_+ - y_-}{\alpha}\]

where \(y_-\) \(y_+\) are the raw measurements and \(\alpha\) is the intensity estimated by calling estim_alpha().

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: Rescaled unsplit measurements with shape \((*, M)\).