spyrit.core.prep.UnsplitRescaleEstim.pinv_estim

UnsplitRescaleEstim.pinv_estim(y: tensor) → tensor

Estimate gain from pseudo-inverse.

Args:

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

Returns:

torch.tensor: Estimated intensity \(\alpha\) of shape \((*, 1)\).