hofmannu / clahe3d Goto Github PK

Right now we still rely on those individual unit tests defined through CMake interface, we should move to a more modern approach of unit testing using e.g. catch2.

Hi, when i tried to git submodule init command it gives me error like this:
Submodule 'lib/CVolume' (https://github.com/hofmannu/CVolume.git) registered for path 'lib/CVolume'
Submodule 'lib/imgui' (https://github.com/ocornut/imgui.git) registered for path 'lib/imgui' Any idea?

generate binary masks depending on a threshold and apply those masks to other volumes again. Implementing this feature requires severe reprogramming of the volume handling since we want to store multiple representations simultaneously.

Running CLAHE3D with GPU support fails testcase #9 with the following output:

ctest --rerun-failed --output-on-failure

Start 9: test_gaussfilt
1/1 Test #9: test_gaussfilt ...................Subprocess aborted***Exception:   1.38 sec
terminate called after throwing an instance of 'char const*'


0% tests passed, 1 tests failed out of 1

Total Test time (real) =   1.38 sec

The following tests FAILED:
	  9 - test_gaussfilt (Subprocess aborted)
Errors while running CTest

Possibly related warnings that are thrown during compilation:

meanfilt.h:24:7: warning: direct base ‘cudaTools’ inaccessible in ‘meanfilt’ due to ambiguity [-Winaccessible-base]
   24 | class meanfilt :
      |       ^~~~~~~~

 warning: narrowing conversion of ‘(int)sett.gaussfiltsett::kernelSize[0]’ from ‘int’ to ‘long unsigned int’ [-Wnarrowing]
   75 |         gaussfilter.set_kernelSize(
      |        

I downloaded and built the packaged exactly as provided in the instructions. System is an updated ARCH Linux with Nvidia 3090.

Move the actual CLAHE3D filtering procedure to multicore calculation to speed up things,

For the future we want to be capable of combining different versions of the volume after applying for example different filters. For this it would be very nice to have a workspace where we can rename, copy, create, delete, etc. different volumetric representations.

Self explaining

biggest challenge: where can we store the datasets?