The current N4 bias correction binary is not deterministic across platforms

[ledigchr]

This is a problem as it hampers reproducibility.

• Investigate
• Consider updating N4 binary

[PichardRarker]

Hi Christian,

Would you mind elaborating on the non-deterministic effect you've observed?

I'm in the process of migrating MALPEM (+ N4) from one server to another (both running Ubuntu 10) and I'm seeing very slightly different results in bias field correction across servers. Digging around in the log files, I can see that the N4 convergence estimates differ very slightly across machines, is this similar to what you're seeing?

Best,

Richard

[ledigchr]

Hi Richard,

Yes, what you describe sounds like a confirmation of this issue. This is a problem with the N4 binary that ships with MALPEM not MALPEM itself. You can try recompiling a more recent version of ITK N4 and replace the binary.

Please let us know if that solves the problem in case you move forward with this.

Christian

[PichardRarker]

Will do, thanks!

Richard

[PichardRarker]

Hi Christian,

I ended up fixing the issue by fixing the ITK_GLOBAL_DEFAULT_NUMBER_OF_THREADS variable across servers, as per this post: https://sourceforge.net/p/advants/discussion/840261/thread/354df4c0/?limit=25#116c

Hope this is useful!

Richard

[ledigchr]

Thanks a lot for sharing this information, very useful!

Best wishes,
Christian