Skip to content

Latest commit

 

History

History
37 lines (27 loc) · 1.61 KB

README.md

File metadata and controls

37 lines (27 loc) · 1.61 KB

Segment SC using contrast-agnostic MONAI model from PSIR contrast and perform vertebral labeling

  1. Create a conda virtual environment and install dependencies
conda create -n monai python=3.8
conda activate monai
pip install -r segment_sc_contrast-agnostic/requirements.txt
  1. Segment SC using the contrast-agnostic MONAI model from PSIR contrast and perform vertebral labeling

The sct_run_batch wrapper script is used to run the segment_sc_contrast-agnostic/segment_sc_contrast-agnostic.sh across multiple subjects in parallel.

sct_run_batch -config config.json

Example config.json file:

{
  "path_data"   : "<PATH_TO_DATASET>/canproco",
  "path_output" : "<PATH_TO_DATASET>canproco_contrast-agnostic_2023-10-06",
  "script"      : "${HOME}/code/canproco/segment_sc_contrast-agnostic/segment_sc_contrast-agnostic.sh",
  "jobs"        : 1,
  "include"     : "ses-M0",
  "exclude"     : "sub-mon118",
  "script_args" : "${HOME}/code/canproco/segment_sc_contrast-agnostic/run_inference_single_image.py ${HOME}/data/models/contrast-agnostic_final_monai_model/nnunet_nf=32_DS=1_opt=adam_lr=0.001_AdapW_CCrop_bs=2_64x192x320_20230918-2253"
}

ℹ️ The conda environment with MONAI is required to run the segment_sc_contrast-agnostic/segment_sc_contrast-agnostic.sh script.

ℹ️ Note that the segment_sc_contrast-agnostic/run_inference_single_image.py script is just a copy of the contrast-agnostic-softseg-spinalcord/monai/run_inference_single_image.py script.