When I ran monailabel in 3D Slicer, segmentation accuracy is very low.

[ulphypro]

hello, members

I have the question.

When I ran monailabel in 3D Slicer, segmentation accuracy is very low.
And there are waring messages in python console as follows.

Please, let me know how to increase segmentation accuracy and remove warning messages regarding [QT and VTK] occurred in python console.

[SachidanandAlle]

Warning possibly due to label mask generated by the model. The model might not been training over enough epochs etc..

If you are using bundle, try to train a model from scratch using the corresponding docs. Same in case if you are rereferring monailabel app. Run for enough number of epochs and make sure samples are enough.

[ulphypro]

Dear @SachidanandAlle

Thank you for answering my question.

I'm conducting monailabel app.

The number of epoch times is 50 as showing figure.

I don't find code file to edit number of epoch times.

Can you let me know how to modify number of epoch times?

And which file should I use?

[diazandr3s]

In the Options tab in Slicer, you can change the number of epochs. Please check here: https://youtu.be/KtPE8m0LvcQ?t=6334

[diazandr3s]

BTW, I know the BRATS dataset is a multiparametric dataset. You may find this discussion useful: #1051

Also check this comment: #1051 (comment)

[ulphypro]

Dear @diazandr3s

Thank you for answering my question.

I found option tab to modify number of epoch times.

For dataset,
segmentation_brats dataset is BraTS2021 dataset.(kaggle link : https://www.kaggle.com/datasets/dschettler8845/brats-2021-task1)

Please refer to image file.

Question 1.
After I conducted tumor segmentation using 'Grow from Seeds' effect in 'Segment Editor' option, I pressed 'Submit label' button. Then I should press 'Train' button, but when I pressed 'Submit Label' button, and then it goes into next images. Please, let me know how to solve this problem.

Question 2.
'Run' button doesn't show in 'Auto Segmentation' option.
Is problem that clashed between torch version and monailabel version?
Some times when I reinstalled 3d slicer program, I can show 'Run' button in 'Auto Segmenation' option. Now despite of reinstalling 3d slicer, 'Run' button didn't show in 'Auto Segmentation' option.
Please, let me know how to also solve this problem.

monailabel version is 0.5.2.
And torch version information is as following.

[diazandr3s]

Dear @diazandr3s

Thank you for answering my question.

I found option tab to modify number of epoch times.

For dataset, segmentation_brats dataset is BraTS2021 dataset.(kaggle link : https://www.kaggle.com/datasets/dschettler8845/brats-2021-task1)

Please refer to image file.

Question 1. After I conducted tumor segmentation using 'Grow from Seeds' effect in 'Segment Editor' option, I pressed 'Submit label' button. Then I should press 'Train' button, but when I pressed 'Submit Label' button, and then it goes into next images. Please, let me know how to solve this problem.

Question 2. 'Run' button doesn't show in 'Auto Segmentation' option. Is problem that clashed between torch version and monailabel version? Some times when I reinstalled 3d slicer program, I can show 'Run' button in 'Auto Segmenation' option. Now despite of reinstalling 3d slicer, 'Run' button didn't show in 'Auto Segmentation' option. Please, let me know how to also solve this problem.

monailabel version is 0.5.2. And torch version information is as following.

Hi @ulphypro,

BRATS dataset has 4 modalities per patient. Please refer to the conversation I shared before regarding working with multiparametric datasets with 3DSlicer+MONAI Label.

Question 1:

If you don't want to automatically fetch next sample after submitting to the server, you can deactivate that function by going to Edit-Application Settings-MONAILabel and then deactivating Auto-Fetch Next Sample.

Question 2:

Please use the latest MONAI Label version (0.6.0) That should not occur on the latest version.

Hope this helps

[ulphypro]

Dear @diazandr3s

Thank your for helping me.

I want to use monailabel 0.6.0 version, but it occurred the error whenever I installed monailabel 0.6.0 package as follow.

----Error Message------

PS C:\Users\aaa> monailabel start_server --app apps/radiology --studies datasets/BraTS2021_Training_Data/brats --conf models segmentation_brats
Using PYTHONPATH=C:\Users\aaa\AppData\Local\Programs\Python;
""
2023-03-09 17:51:57,020 - USING:: version = False
2023-03-09 17:51:57,020 - USING:: app = C:\Users\aaa\apps\radiology
2023-03-09 17:51:57,020 - USING:: studies = C:\Users\aaa\datasets\BraTS2021_Training_Data\brats
2023-03-09 17:51:57,020 - USING:: verbose = INFO
2023-03-09 17:51:57,020 - USING:: conf = [['models', 'segmentation_brats']]
2023-03-09 17:51:57,020 - USING:: host = 0.0.0.0
2023-03-09 17:51:57,020 - USING:: port = 8000
2023-03-09 17:51:57,020 - USING:: uvicorn_app = monailabel.app:app
2023-03-09 17:51:57,020 - USING:: ssl_keyfile = None
2023-03-09 17:51:57,020 - USING:: ssl_certfile = None
2023-03-09 17:51:57,020 - USING:: ssl_keyfile_password = None
2023-03-09 17:51:57,020 - USING:: ssl_ca_certs = None
2023-03-09 17:51:57,020 - USING:: workers = None
2023-03-09 17:51:57,020 - USING:: limit_concurrency = None
2023-03-09 17:51:57,020 - USING:: access_log = False
2023-03-09 17:51:57,020 - USING:: log_config = None
2023-03-09 17:51:57,020 - USING:: dryrun = False
2023-03-09 17:51:57,020 - USING:: action = start_server
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_API_STR =
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_PROJECT_NAME = MONAILabel
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_APP_DIR =
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_STUDIES =
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_AUTH_ENABLE = False
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_AUTH_DB =
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_APP_CONF = '{}'
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_TASKS_TRAIN = True
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_TASKS_STRATEGY = True
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_TASKS_SCORING = True
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_TASKS_BATCH_INFER = True
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_DATASTORE =
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_DATASTORE_URL =
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_DATASTORE_USERNAME =
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_DATASTORE_PASSWORD =
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_DATASTORE_API_KEY =
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_DATASTORE_CACHE_PATH =
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_DATASTORE_PROJECT =
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_DATASTORE_ASSET_PATH =
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_DATASTORE_DSA_ANNOTATION_GROUPS =
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_DICOMWEB_USERNAME =
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_DICOMWEB_PASSWORD =
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_DICOMWEB_CACHE_PATH =
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_QIDO_PREFIX = None
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_WADO_PREFIX = None
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_STOW_PREFIX = None
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_DICOMWEB_FETCH_BY_FRAME = False
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_DICOMWEB_CONVERT_TO_NIFTI = True
2023-03-09 17:51:57,020 - ENV SETTINGS:: MONAI_LABEL_DICOMWEB_SEARCH_FILTER = '{"Modality": "CT"}'
2023-03-09 17:51:57,036 - ENV SETTINGS:: MONAI_LABEL_DICOMWEB_CACHE_EXPIRY = 180
2023-03-09 17:51:57,036 - ENV SETTINGS:: MONAI_LABEL_DICOMWEB_PROXY_TIMEOUT = 30.0
2023-03-09 17:51:57,036 - ENV SETTINGS:: MONAI_LABEL_DICOMWEB_READ_TIMEOUT = 5.0
2023-03-09 17:51:57,036 - ENV SETTINGS:: MONAI_LABEL_DATASTORE_AUTO_RELOAD = True
2023-03-09 17:51:57,036 - ENV SETTINGS:: MONAI_LABEL_DATASTORE_READ_ONLY = False
2023-03-09 17:51:57,036 - ENV SETTINGS:: MONAI_LABEL_DATASTORE_FILE_EXT = '[".nii.gz", ".nii", ".nrrd", ".jpg", ".png", ".tif", ".svs", ".xml"]'
2023-03-09 17:51:57,036 - ENV SETTINGS:: MONAI_LABEL_SERVER_PORT = 8000
2023-03-09 17:51:57,036 - ENV SETTINGS:: MONAI_LABEL_CORS_ORIGINS = '[]'
2023-03-09 17:51:57,036 - ENV SETTINGS:: MONAI_LABEL_SESSIONS = True
2023-03-09 17:51:57,036 - ENV SETTINGS:: MONAI_LABEL_SESSION_PATH =
2023-03-09 17:51:57,036 - ENV SETTINGS:: MONAI_LABEL_SESSION_EXPIRY = 3600
2023-03-09 17:51:57,036 - ENV SETTINGS:: MONAI_LABEL_INFER_CONCURRENCY = -1
2023-03-09 17:51:57,036 - ENV SETTINGS:: MONAI_LABEL_INFER_TIMEOUT = 600
2023-03-09 17:51:57,036 - ENV SETTINGS:: MONAI_LABEL_TRACKING_ENABLED = True
2023-03-09 17:51:57,036 - ENV SETTINGS:: MONAI_LABEL_TRACKING_URI =
2023-03-09 17:51:57,036 - ENV SETTINGS:: MONAI_ZOO_SOURCE = github
2023-03-09 17:51:57,036 - ENV SETTINGS:: MONAI_ZOO_REPO = Project-MONAI/model-zoo/hosting_storage_v1
2023-03-09 17:51:57,036 - ENV SETTINGS:: MONAI_ZOO_AUTH_TOKEN =
2023-03-09 17:51:57,036 - ENV SETTINGS:: MONAI_LABEL_AUTO_UPDATE_SCORING = True
2023-03-09 17:51:57,036 -
Allow Origins: ['*']
[2023-03-09 17:51:57,903] [9428] [MainThread] [INFO] (uvicorn.error:75) - Started server process [9428]
[2023-03-09 17:51:57,904] [9428] [MainThread] [INFO] (uvicorn.error:45) - Waiting for application startup.
[2023-03-09 17:51:57,904] [9428] [MainThread] [INFO] (monailabel.interfaces.utils.app:38) - Initializing App from: C:\Users\aaa\apps\radiology; studies: C:\Users\aaa\datasets\BraTS2021_Training_Data\brats; conf: {'models': 'segmentation_brats'}
[2023-03-09 17:51:58,191] [9428] [MainThread] [INFO] (monailabel.utils.others.class_utils:37) - Subclass for MONAILabelApp Found: <class 'main.MyApp'>
[2023-03-09 17:51:58,238] [9428] [MainThread] [ERROR] (uvicorn.error:119) - Traceback (most recent call last):
File "C:\Users\aaa\AppData\Local\Programs\Python\Python39\lib\site-packages\starlette\routing.py", line 635, in lifespan
async with self.lifespan_context(app):
File "C:\Users\aaa\AppData\Local\Programs\Python\Python39\lib\site-packages\starlette\routing.py", line 530, in aenter
await self.router.startup()
File "C:\Users\aaa\AppData\Local\Programs\Python\Python39\lib\site-packages\starlette\routing.py", line 612, in startup
await handler()
File "C:\Users\aaa\AppData\Local\Programs\Python\Python39\lib\site-packages\monailabel\app.py", line 106, in startup_event
instance = app_instance()
File "C:\Users\aaa\AppData\Local\Programs\Python\Python39\lib\site-packages\monailabel\interfaces\utils\app.py", line 51, in app_instance
app = c(app_dir=app_dir, studies=studies, conf=conf)
File "C:\Users\aaa\apps\radiology\main.py", line 45, in init
for c in get_class_names(lib.configs, "TaskConfig"):
File "C:\Users\aaa\AppData\Local\Programs\Python\Python39\lib\site-packages\monailabel\utils\others\class_utils.py", line 144, in get_class_names
module = importlib.import_module("." + name, package=current_module_name)
File "C:\Users\aaa\AppData\Local\Programs\Python\Python39\lib\importlib_init.py", line 127, in import_module
return _bootstrap._gcd_import(name[level:], package, level)
File "", line 1030, in _gcd_import
File "", line 1007, in _find_and_load
File "", line 986, in _find_and_load_unlocked
File "", line 680, in _load_unlocked
File "", line 850, in exec_module
File "", line 228, in _call_with_frames_removed
File "C:\Users\aaa\apps\radiology\lib\configs\deepedit.py", line 26, in
from monailabel.tasks.activelearning.tta import TTA
ModuleNotFoundError: No module named 'monailabel.tasks.activelearning.tta'

[2023-03-09 17:51:58,239] [9428] [MainThread] [ERROR] (uvicorn.error:56) - Application startup failed. Exiting.

By the way, when I installed monailabel 0.5.2 version, it perfectly doesn't work, however it didn't occur the error masseges as above.
The error only occured on monaiabel 0.6.0 version.
So, I always conducted segmentation on monailabel 0.5.2.

My laptop GPU and system information are as follows.
[display]
Operating System: Windows 11 Home, 64-bit
DirectX Version: 12.0
GPU Processor: NVIDIA GeForce RTX 3070 Laptop GPU
Driver Version: 531.14
Driver Type: DCH
Direct3D feature level: 12_1
CUDA Core: 5120
Resizable BAR Yes
Max-Q technology 3rd generation
Dynamic Boost 2.0 Yes
WhisperMode 2.0 Yes
Advanced Optimus No
Maximum graphics performance 100 W
Core clock: 1290 MHz
Memory data rate: 12.00 Gbps
Memory interface: 256-bit
Memory bandwidth: 384.06 GB/s
Total available graphics memory: 16079 MB
Dedicated video memory: 8192 MB GDDR6
System video memory: 0 MB
Shared system memory: 7887 MB
Video BIOS Version: 94.04.4F.00.19
IRQ: Not used
Bus: PCI Express x8 Gen3
Device Id: 10DE 249D 20141A58
Part number: 4735 0010

cuDNN installation information : cudnn-windows-x86_64-8.8.0.121_cuda12-archive
Python version : 3.9.13

Please, let me know how to solve the problem that ran in monailabel 0.5.2 but didn't run in monailabel 0.6.0.

[SachidanandAlle]

You are using older app.. with new version.. download the radiology app again and try

[ulphypro]

Dear @SachidanandAlle

Thank you for helping me.

I downloaded monailabel 0.6.0, and then I conducted segmentation using monailabel 0.6.0.

But 'Run' button doesn't show in 'Auto Segmentation' option.

And when I had run monailabel server in window terminal, train-loss was nan.

I don't know why train-loss is nan.

Please, let me know what I should do to solve the problem.

[tangy5]

The NaN loss probably means the labels have some issue, could you check the annotations and training samples and see if they are "ready" for training run, e.g., number of channels, classes, etc..some basic parameters.

[ulphypro]

Dear @tangy5

Thank you for helping my problem.

I checked training sample, number of channels, classes or some basic parameters. I found solution regarding the problem that 'train-loss is nan' occurred.

When I had modified "if strtobool(self.conf.get("use_pretrained_model", "true")):" into " if strtobool(self.conf.get("use_pretrained_model", "false")):
", train-loss worked normally as following.

But 'Run' button doesn't show in 'Auto Segmentation' still as following.

I already installed MONAILabel 0.6.0 version.
And information of torch+cuda117 version installed in my laptop is as following.
torch version : 1.13.1
torch vision : 0.14.1
torch audio : 0.13.1
And monailabel : 0.6.0

I don't know that 'Run' button doesn't show in 'Auto Segmentation' option .

Please, let me know what I should do to show 'Run' button in 'Auto Segmentation' option.