spyrit.core.noise.Poisson.forward

Poisson.forward(x)

Simulates measurements corrupted by Poisson noise

Args:

x: Batch of images

Shape:

• x: \((*, N)\)

• Output: \((*, M)\)

Example 1: Two noisy measurement vectors from a Linear measurement operator

>>> H = torch.rand([400,32*32])
>>> meas_op = Linear(H)
>>> noise_op = Poisson(meas_op, 10.0)
>>> x = torch.FloatTensor(10, 32*32).uniform_(-1, 1)
>>> y = noise_op(x)
>>> print(y.shape)
>>> print(f"Measurements in ({torch.min(y):.2f} , {torch.max(y):.2f})")
>>> y = noise_op(x)
>>> print(f"Measurements in ({torch.min(y):.2f} , {torch.max(y):.2f})")
torch.Size([10, 400])
Measurements in (2249.00 , 2896.00)
Measurements in (2237.00 , 2880.00)

Example 2: Two noisy measurement vectors from a HadamSplit operator

>>> Perm = torch.rand([32*32,32*32])
>>> meas_op = HadamSplit(H, Perm, 32, 32)
>>> noise_op = Poisson(meas_op, 200.0)
>>> x = torch.FloatTensor(10, 32*32).uniform_(-1, 1)
>>> y = noise_op(x)
>>> print(y.shape)
>>> print(f"Measurements in ({torch.min(y):.2f} , {torch.max(y):.2f})")
>>> y = noise_op(x)
>>> print(f"Measurements in ({torch.min(y):.2f} , {torch.max(y):.2f})")
torch.Size([10, 800])
Measurements in (0.00 , 55338.00)
Measurements in (0.00 , 55077.00)

Example 3: Two noisy measurement vectors from a LinearSplit operator

>>> H = torch.rand(24,64)
>>> meas_op = LinearSplit(H)
>>> noise_op = Poisson(meas_op, 50.0)
>>> x = torch.FloatTensor(10, 64, 92).uniform_(-1, 1)
>>> y = noise_op(x)
>>> print(y.shape)
>>> print(f"Measurements in ({torch.min(y):.2f} , {torch.max(y):.2f})")
>>> y = noise_op(x)
>>> print(f"Measurements in ({torch.min(y):.2f} , {torch.max(y):.2f})")
torch.Size([10, 48, 92])
Measurements in (500.00 , 1134.00)
Measurements in (465.00 , 1140.00)