LogPoissonLikelihoodDistance

class deepinv.optim.LogPoissonLikelihoodDistance(N0=1024.0, mu=1 / 50.0)

Bases: Distance

Log-Poisson negative log-likelihood.

\[\distance{z}{y} = N_0 (1^{\top} \exp(-\mu z)+ \mu \exp(-\mu y)^{\top}x)\]

Corresponds to LogPoissonNoise with the same arguments N0 and mu. There is no closed-form of the prox known.

Parameters:

• N0 (float) – average number of photons

• mu (float) – normalization constant

fn(x, y, *args, **kwargs)

Computes the distance \(\distance{x}{y}\).

Parameters:

• x (torch.Tensor) – Variable \(x\).

• y (torch.Tensor) – Observation \(y\).

Returns:

(torch.Tensor) distance \(\distance{x}{y}\) of size B with B the size of the batch.