Hi, I am following the DevDay notebook to get an idea on how Deepgrow works. I think I got Auto-segmentation(clara_ct_seg_spleen_no_amp) to work on my CT series. When I tried clara_deepgrow by clicking a bunch of points of the ROI, I was not getting any result back from AIAA which running locally.
Any idea on how to troubleshoot it?
Thanks.
Hi
Thanks for your interest in clara train AIAA.
You should be able to load the Deep grow model similar to how you loaded the segmentation model. can you verify you can see the deepgrow model in the slicer lefthand side panel (under deepgrow section) ? --> this would confirm you loaded the model correctly
If so you should click the foreground (+ve) points button then click a point on the liver or spleen. Wait for a few minutes then see the segmentation
to trouble shoot go to your AIAA :/logs
Hope that helps
Thanks for your reply. I also tried MiTK. Following is a part of the logs that seem to indicate AIAA got the data. However I am not seeing any visual result on MITK and I don’t have a case to compare with.
[2021-01-05 19:29:09] [INFO] (nvmidl.apps.aas.actions.inference_engine) - Load Data from: /claraDevDay/tmp/AIAA//sessions/30bffbec-4f8b-11eb-bbf4-0242ac110004/tmp.0.WrXa6Y.nii.gz
[2021-01-05 19:29:09] [INFO] (nvmidl.apps.aas.actions.inference_engine) - Using Image: /claraDevDay/tmp/AIAA//sessions/30bffbec-4f8b-11eb-bbf4-0242ac110004/tmp.0.WrXa6Y.nii.gz
[2021-01-05 19:29:09] [INFO] (nvmidl.apps.aas.actions.inference_engine) - Using Params: {'foreground': '[[396,322,81],[396,322,81],[425,297,81],[425,297,81],[397,312,81],[397,312,81],[424,301,81],[424,301,81],[422,295,81],[422,295,81],[393,298,81],[393,298,81],[370,317,81],[370,317,81],[367,311,81],[367,311,81],[362,318,81],[362,318,81],[411,324,81],[411,324,81],[400,336,81],[400,336,81],[422,318,81],[422,318,81],[422,318,81],[422,318,81],[424,289,81],[424,289,81],[405,302,81],[405,302,81],[380,325,81],[380,325,81],[376,331,81],[376,331,81],[404,310,81],[404,310,81],[383,321,81],[383,321,81],[383,321,81],[383,321,81],[373,317,81],[373,317,81],[427,282,81],[427,282,81]]', 'background': '[]', 'image_original': '/claraDevDay/tmp/AIAA//sessions/30bffbec-4f8b-11eb-bbf4-0242ac110004/tmp.0.WrXa6Y.nii.gz'}
[2021-01-05 19:29:09] [INFO] (nvmidl.apps.aas.actions.inference_engine) - Run Pre Processing
[2021-01-05 19:29:09] [INFO] (nvmidl.apps.aas.actions.inference_engine) - Pre-Processing Input Keys: dict_keys(['image', 'image_path', 'params'])
[2021-01-05 19:29:09] [INFO] (nvmidl.apps.aas.inference.inference_utils) - PRE - Run Transforms
[2021-01-05 19:29:09] [INFO] (nvmidl.apps.aas.inference.inference_utils) - PRE - Input Keys: dict_keys(['image', 'image_path', 'params'])
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.inference.inference_utils) - PRE - Time consumed by Transform (LoadNifti): 0.8864603042602539
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.inference.inference_utils) - PRE - Time consumed by Transform (DeepGrow2DGetOneSlice): 0.0006926059722900391
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.inference.inference_utils) - PRE - Time consumed by Transform (DeepGrow2DScaleByResolution): 0.02539229393005371
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.inference.inference_utils) - PRE - Time consumed by Transform (DeepGrow2DCropCenterVolume): 0.000667572021484375
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.inference.inference_utils) - PRE - Time consumed by Transform (DeepGrow2DFixedMeanStdNormalization): 0.0002837181091308594
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.inference.inference_utils) - PRE - Time consumed by Transform (DeepGrow2DAddUserPoints): 0.007464170455932617
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.inference.inference_utils) - PRE - Time consumed by Transform (DeepGrow2DMergeImages): 0.000545501708984375
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.inference.inference_utils) - PRE - Time consumed by Transform (ConvertToChannelsFirst): 7.414817810058594e-05
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.actions.inference_engine) - ++ Total Time consumed for pre-processing: 0.9225304126739502
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.actions.inference_engine) - Pre-Processing Output Keys: dict_keys(['image_path', 'params', 'image.affine', 'image.original_affine', 'image.file_name', 'image.file_format', 'image.original_shape', 'image.original_shape_format', 'image.spacing', 'image.as_canonical', 'image.shape_format', 'image.slice_index', 'image.resolution_factor', 'image.start_px', 'image.end_px', 'foreground.shape_format', 'background.shape_format', 'image'])
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.inference.trtis_inference) - Run TRTIS Inference for: clara_deepgrow
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.inference.inference_utils) - ShapeFormat: CHW
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.inference.trtis_inference) - Original ShapeFormat: CHW
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.actions.inference_engine) - ++ Total Time consumed for inference: 0.10468339920043945
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.actions.inference_engine) - Inference Output Keys: dict_keys(['image_path', 'params', 'image.affine', 'image.original_affine', 'image.file_name', 'image.file_format', 'image.original_shape', 'image.original_shape_format', 'image.spacing', 'image.as_canonical', 'image.shape_format', 'image.slice_index', 'image.resolution_factor', 'image.start_px', 'image.end_px', 'foreground.shape_format', 'background.shape_format', 'image', '*.shape_format', 'model'])
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.actions.inference_engine) - Run Post Processing
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.actions.inference_engine) - Post-Processing Input Keys: dict_keys(['image_path', 'params', 'image.affine', 'image.original_affine', 'image.file_name', 'image.file_format', 'image.original_shape', 'image.original_shape_format', 'image.spacing', 'image.as_canonical', 'image.shape_format', 'image.slice_index', 'image.resolution_factor', 'image.start_px', 'image.end_px', 'foreground.shape_format', 'background.shape_format', 'image', '*.shape_format', 'model'])
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.inference.inference_utils) - POST - Run Transforms
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.inference.inference_utils) - POST - Input Keys: dict_keys(['image_path', 'params', 'image.affine', 'image.original_affine', 'image.file_name', 'image.file_format', 'image.original_shape', 'image.original_shape_format', 'image.spacing', 'image.as_canonical', 'image.shape_format', 'image.slice_index', 'image.resolution_factor', 'image.start_px', 'image.end_px', 'foreground.shape_format', 'background.shape_format', 'image', '*.shape_format', 'model'])
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.inference.inference_utils) - POST - Time consumed by Transform (DeepGrow2DGetPrediction): 0.01916813850402832
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.inference.inference_utils) - POST - Time consumed by Transform (CopyProperties): 0.00012111663818359375
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.actions.inference_engine) - ++ Total Time consumed for post-processing: 0.019738435745239258
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.actions.inference_engine) - ++ Total Time consumed for ALL: 1.0473146438598633
[2021-01-05 19:29:10] [INFO] (nvmidl.apps.aas.actions.inference_engine) - FilePath: /tmp; FileName: tmpzmk5t0ka
[2021-01-05 19:29:11] [INFO] (nvmidl.apps.aas.actions.inference_engine) - Result File: /tmp/tmpzmk5t0ka_model.nii.gz
[2021-01-05 19:29:11] [INFO] (nvmidl.apps.aas.actions.inference_engine) - Result Params: {}
[2021-01-05 19:29:11] [INFO] (nvmidl.apps.aas.www.api.api_v1) - Result Json: {}
[2021-01-05 19:29:11] [INFO] (nvmidl.apps.aas.www.api.api_v1) - Result File: /tmp/tmpzmk5t0ka_model.nii.gz
[2021-01-05 19:31:19] [INFO] (schedule) - Running job Every 5 minutes do cleanup_sessions({'30bffbec-4f8b-11eb-bbf4-0242ac110004': <nvmidl.apps.aas.actions.sessions.SessionInfo object at 0x7f4842a50f60>}) (last run: 2021-01-05 19:26:19, next run: 2021-01-05 19:31:19)
[2021-01-05 19:36:19] [INFO] (schedule) - Running job Every 5 minutes do cleanup_sessions({'30bffbec-4f8b-11eb-bbf4-0242ac110004': <nvmidl.apps.aas.actions.sessions.SessionInfo object at 0x7f4842a50a58>}) (last run: 2021-01-05 19:31:19, next run: 2021-01-05 19:36:19)```Hi
It seems you are able to debug and show files received by AIAA and inference ran correctly. I would test again with a small CT volume from may be the decatholon spleen data set. I think may be file is very large and caused some issues.
you should clicking on liver or spleen or kidneys, should get some sort of results.
Also do you get any result when running segmentation models?