Open ckolbPTB opened 3 weeks ago
Coverage Report
Tests | Skipped | Failures | Errors | Time |
---|---|---|---|---|
828 | 0 :zzz: | 0 :x: | 0 :fire: | 1m 10s :stopwatch: |
Allow for a linear operator in the regularization: ... + L*(Bx-x0)^2 ? With B defaulting to identity?
With B defaulting to identity?
Would we then still omit an implementation of the identity operator and realise it with setting B: LinearOperator | float = 1. and making use of the OperatorTensorSum?
B as float does not make sense IMHO.
I would make it LinearOperator | None with None being the identity (this also avoids creating an object as a default operator)
We can still have an explicit IdentityOperator, that can be used there (or used as a dummy linear operator)
Gesendet von Outlook für Androidhttps://aka.ms/AAb9ysg
From: Christoph Kolbitsch @.> Sent: Thursday, September 5, 2024 9:25:23 AM To: PTB-MR/mrpro @.> Cc: Felix F Zimmermann @.>; Comment @.> Subject: Re: [PTB-MR/mrpro] Add RegularizedIterativeSENSEReconstruction (PR #388)
With B defaulting to identity?
Would we then still omit an implementation of the identity operator and realise it with setting B: LinearOperator | float = 1. and making use of the OperatorTensorSum?
— Reply to this email directly, view it on GitHubhttps://github.com/PTB-MR/mrpro/pull/388#issuecomment-2330798330, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ABH3DV47JR4YBFURG2LKKP3ZVABOHAVCNFSM6AAAAABNU4MUICVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMZQG44TQMZTGA. You are receiving this because you commented.Message ID: @.***>
can we jusit add the regularization example to the existing iterative sense example? these two examples seem quite redundant.
can we jusit add the regularization example to the existing iterative sense example? these two examples seem quite redundant.
yes there is sume redundancy but from a teaching perspective I think it is easier to understand to build up the complexity of the reconstruction step by step.
closes #369 closes #321 closes #6