this seems to be a data format independent problem (tried nrrd as well).

Please can you give a more complete description?

I can run command line modules with Nifti images, and they are loaded back into the GUI. So, which command line program are you using? What are the inputs, and outputs etc?

Thanks

Matt

The problem occurs if there are no input images from MITK, but the image data is loaded inside the external program.

If there are no images loaded within mitkWorkbench (or the diffusion app), then there is no Geometry3D defined.

So, are you running an external command line app, with absolutely no images in the mitkWorkbench? In this case, if the output image name was specified, correctly in the XML, then it should still be loaded. You will see it appear in the DataManager. Can you right click on it and do "Global Re-init".

If you already have images or surfaces loaded in mitkWorkbench, then you would have already got a valid Geometry3D, so the new images, produced as a result of the external program should be visible if the volume of interest overlaps with the existing geometry. Again, Global Re-init should help.

Can you test this please?

Thanks

Matt

Global Re-init does help. Thanks!

Could this be triggered on loading the data to the data manager?

I don't know. I presume it can, but then you have the problem .... is it a change request, or is it a bug? What about people that do not want an automatic global re-init? It could be made a preference to the Command Line Modules plugin.

As this is sounding like a "change request" to me, I will await Sascha.

M

Thanks Matt for the quick analysis. There is actually T12954 for this issue (which is just a missing initial geometry, as Matt pointed out).

The File->Open action explicitly initializes the render windows after loading data if the datastorage initially was empty, so the problem is only apparent if data is loaded "manually".