poission-gaussian observation model

[dalessioluca]

In fluorescent microscopy the observation model should be poisson-gaussian since the brighter a pixel the more photon the more poisson noise. Things get complicated b/c CCD cameras have offsets. Porbavbly the parameter of the observation model (constant term, linear term and offset) should be learned. Ask Mehrtash and Tianle

[dalessioluca]

Another possibility is to do Anscombe transformation and trasnform the data to have Gaussian noise only.
See paper:
Poisson-Gaussian noise + Anscombe transformation: https://arxiv.org/abs/1812.10366

[dalessioluca]

looking at the empirical variance vs average plot it does not seems that the Poisson-Gaussian model is correct.