Bayesian methods for background modeling

[nksauter]

References:

• https://doi.org/10.1107/S1600576716013595
• https://doi.org/10.1107/S2052252517010259
• https://doi.org/10.1107/S205979831700969X

Fundamentally we have to understand how to modify the old Andrew Leslie-type treatment to account for asymmetrical shoeboxes (non-mm symmetry), and for small-value photon-counted signals.

We have the opportunity to include uncertainty in pedestal and gain from the dark run/flat field calibration runs. How to include this in these Bayesian method?