ENH: Bump elastix to 5.3.0#363
Conversation
|
@N-Dekker it looks like we get a slight difference in the registration, resulting in a small out-of-domain sampling at the lower left: https://open.cdash.org/tests/2368766089 perhaps we increase the baseline comparison tolerance to allow more pixels? Or add a new baseline? |
Thanks Matt. I see, you did already increase the tolerance before. Two times already: So then, I wonder, what would cause these new differences? I wonder, would they be introduced by SuperElastix/elastix, somewhere between those revisions? SuperElastix/elastix@8c4f347...798346c Is the registration result still fine anyway? Update: Maybe it's pull request SuperElastix/elastix#1361 commit SuperElastix/elastix@6f031a5 (Reduce rounding errors "GridSpacingSchedule" GetDefaultParameterMap) So then indeed, we should have a new baseline! Can someone please 🙏 help me to get it there? |
8f50a54 to
05e9abd
Compare
|
Update: this commit BUG: Replace default values of "GridSpacingSchedule" in test aims to solve the test failure without changing the baseline 🙏 |
|
@N-Dekker great work!! Looks like it will be green with a formatting fix. |
12c2c99 to
2b11358
Compare
|
Build / py-dev / build-linux-py (11, 2014-x64) (pull_request) fails, saying: I'll give it another try! |
2b11358 to
1198b6e
Compare
Follow-up to pull request InsightSoftwareConsortium#344 commit 2821abb "COMP: Use upstream elastix repository", July 2025.
1198b6e to
c5fd9ca
Compare
|
When this is merged, shall we also make a new tag for ITKElastix? What about |
Yes, we need to make sure to update the version first, here: Line 7 in f43d560 |
Follow-up to pull request #344 commit 2821abb
"COMP: Use upstream elastix repository", July 2025.