Closed paskino closed 2 weeks ago
I wouldn't put this in the ISTA example. It certainly makes very little sense to have both OSEM-type step-size (i.e. preconditioner) and the "kappa^2" preconditioner. It would be a good candidate for SIRF-Contrib though. We could conceivably have some OSEM implementations there: hand-crafted, using SIRF, using CIL.
By the way, I'm not sure if this should be a step-size or preconditioner. I guess with the current CIL code, there's little distinction anymore.
shall we close this then?
Closing this as no intention to merge it.
Create and use the OSEM step size rule.