PoissonLogLikelihoodWithLinearModelForMeanAndProjData norm set_up problem #56
Assignees
Comments
KrisThielemans
added a commit
to KrisThielemans/STIR-1
that referenced
this issue
Dec 23, 2023
make sure that currect normalisation::set_up is used for either sensitivity, log-likelihood or Hessian calculations Still need to do listmode Partially fixes NikEfth#56
|
Last 2 bullets are fixed in #52 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
We have an option to compute the sensitivity without TOF. There are currently 3 problems:
This should only be enabled if the norm can handle non-TOF data (e.g. TOF
BinNormalisationFromProjDatawill fail), but we have no way of checking that capability ATM.In bcab3d6 I introduced a rather complicated scheme that if nonTOF sensitivity is enabled, norm is set_up with non-TOF proj_data_info, as that is what is needed for the sensitivity calculation. This is fine for OSMAPOSL. However, OSSPS calls the Hessian (and a user could), which needs the norm. In the current version, it therefore errors out in
BinNormalisation::check()as the norm is still set_up with the non-TOF proj_data_info.Norm is currently only set_up if sensitivity needs to be computed.
The text was updated successfully, but these errors were encountered: