Open JostMigenda opened 4 months ago
:zap: Code Analysis Results :zap:
Reverted this commit after discussing with @OrodRazeghi on Slack. Writing the mitk::Surface
to disk between steps is useful, e.g. if clinicians want to look at the mesh and check its quality before moving on to the next steps.
It looks like there are some subtle differences between the surfaces generated by MIRTK (on the development
branch) and by CemrgCommonUtils::ExtractSurfaceFromSegmentation
(our designated replacement). Note in particular the top left image in the attached screenshots, where the white line generated by MIRTK seems like a better fit.
I’m not sure whether that’s caused (at least in part) by the default parameters used for MIRTK being less well suited for the new implementation; so as discussed during previous meeetings, let’s wait for @alonsoJASL to experiment with that.
This is step 1 of the larger project to remove the dependency on MIRTK; an alternate implementation of
CemrgCommandLine::ExecuteSurf
that usesCemrgCommonUtils::ExtractSurfaceFromSegmentation()
instead of MIRTK’sextract-surface
andsmooth-surface
binaries.This is a draft PR right now, because there are a number of to-do items:
close-image
binary at the startmitk::Image
/mitk::Surface
directly, instead of writing them to disk and passing around the file path. Updating the calling code may be more work, but this could reduce overhead quite a bit.