spyrit.core.prep.SplitPoisson.sigma

SplitPoisson.sigma(x: tensor) → tensor

Estimates the variance of the preprocessed measurements

The variance is estimated as \(\frac{4}{\alpha^2} H(x[0::2]+x[1::2])\)

Args:

x: batch of images in the Hadamard domain

Shape:

• Input: \((*,2*M)\) \(*\) indicates one or more dimensions

• Output: \((*, M)\)

Example:

>>> x = torch.rand([10,2*400], dtype=torch.float)
>>> v = split_op.sigma(x)
>>> print(v.shape)
torch.Size([10, 400])