Failed to fetch Sample from MONAI Label Server - Source @ MONAI or Slicer?

[HendrikBE]

Hey all,

The aim of my project is to use Deepgrow in muscle segmentation for which I'm using CT slices from a open dataset. I've adjusted the Deepgrow model with the relevant samples.

Slicer is able to load the images, however when trying to load them from the MONAI Label server it is unable to do so enven though MONAI Label does register the images and does not provide an error itself. Because similar issues were present I've downgraded slicer from 5.2.2. to 5.0.2., this presents the same error.

I've been shooting at this for some time and starting to grind to a halt, is someone able to give me a push in the right direction?

Traceback (most recent call last):
  File "C:/ProgramData/NA-MIC/Slicer 5.0.2/NA-MIC/Extensions-30822/MONAILabel/lib/Slicer-5.0/qt-scripted-modules/MONAILabel.py", line 1212, in onNextSampleButton
    self._volumeNode = slicer.util.loadVolume(image_file)
  File "C:\ProgramData\NA-MIC\Slicer 5.0.2\bin\Python\slicer\util.py", line 908, in loadVolume
    return loadNodeFromFile(filename, filetype, properties, returnNode)
  File "C:\ProgramData\NA-MIC\Slicer 5.0.2\bin\Python\slicer\util.py", line 681, in loadNodeFromFile
    raise RuntimeError(errorMessage)
RuntimeError: Failed to load node from file: **path**\liver_136.nii.gz

[diazandr3s]

Hi @HendrikBE,

Would you please attach the MONAI Label logs? I can't see what can be the error from the Slicer logs :/

[HendrikBE]

Thanks for your reply, last lines before the afore mentioned error of the log below:

[2023-05-10 06:37:00,226] [1504] [MainThread] [INFO] (timeloop:60) - Starting Timeloop..
[2023-05-10 06:37:00,228] [1504] [MainThread] [INFO] (timeloop:42) - Registered job <function MONAILabelApp.on_init_complete.<locals>.run_scheduler at 0x00000283DC0AFD90>
[2023-05-10 06:37:00,230] [1504] [MainThread] [INFO] (timeloop:63) - Timeloop now started. Jobs will run based on the interval set
[2023-05-10 06:38:11,069] [1504] [MainThread] [INFO] (monailabel.endpoints.activelearning:43) - Active Learning Request: {'strategy': 'random', 'client_id': 'user-xyz'}
[2023-05-10 06:38:11,101] [1504] [MainThread] [INFO] (monailabel.tasks.activelearning.random:47) - Random: Selected Image: liver_76; Weight: 1683725891
[2023-05-10 06:38:11,120] [1504] [MainThread] [INFO] (monailabel.endpoints.activelearning:59) - Next sample: {'id': 'liver_76', 'weight': 1683725891, 'path': 'C:\\Users\\Administrator\\workspace\\images\\liver_76.nii.gz', 'ts': 1683632648, 'name': 'liver_76.nii.gz'}

Edit:
So I did some more testing. If I upload the full CT scan they are loaded, I have a feeling the extraction of single slices using Nibabel could be the cause of my woes.