Merge pull request #25 from ivadomed/jca/update-lesion-seg-procedure

Update documentation for manual lesion creation

README.md

@@ -87,4 +87,36 @@ To update the dataset, add all manually-corrected files `derivatives/labels/`,
 
 ## Manual lesion segmentation
 
-[TODO]
+Lesions are segmented on the PSIR or STIR contrasts (when available). The workflow is the following:
+
+- Loop across subjects,
+- Load PSIR/STIR in FSLeyes, and overlay the lesion segmentation (if it already exists),
+- User create/adjust the segmentation mask using editing tools. Toggle the lesion overlay on/off to help validate the accuracy of the label ('Command F' on a Mac),
+- Save the mask
+
+We created a script to do this, which you can download [here](https://github.com/spinalcordtoolbox/manual-correction/tree/r20230302). 
+
+Then, run:
+
+```bash
+python manual_correction.py -path-in <INPUT_PATH> -config <CONFIG_FILE> -path-out <OUTPUT_PATH>
+```
+
+- `INPUT_PATH`: BIDS dataset from which manual lesions will be identified
+- `CONFIG_FILE`: YML file that lists all subjects to be included in the manual lesion segmentation. This file can either be generated by SCT's QC report (see section [3. Manual correction of spinal cord segmentation](#3-manual-correction-of-spinal-cord-segmentation) or manually. See example of a YML file below:
+- `OUTPUT_PATH`: Optional. If not specified, output segmentations will be generated in the `INPUT_PATH`.
+
+  <details><summary>Terminal output</summary>
+
+  ```yaml
+  FILES_LESION:
+    - sub-edm005_ses-M0_PSIR.nii.gz
+    - sub-edm008_ses-M0_PSIR.nii.gz
+    - sub-edm010_ses-M0_PSIR.nii.gz
+    - sub-edm011_ses-M0_PSIR.nii.gz
+    - sub-edm013_ses-M0_PSIR.nii.gz
+  ```
+
+  </details>
+
+Here are several helpful videos reviewing how to correct manual lesion segmentations: [1](https://www.dropbox.com/s/j1f81vtmmmkddtv/Screen%20Recording%202023-01-09%20at%209.45.40%20AM.mov?dl=0), [2](https://www.dropbox.com/s/bm6vpcqe062t2j0/Screen%20Recording%202023-01-11%20at%201.54.36%20PM.mov?dl=0), [3](https://www.dropbox.com/s/00xjsk917wwkp7b/Screen%20Recording%202023-01-11%20at%202.20.58%20PM.mov?dl=0), [4](https://www.dropbox.com/s/3gkrfslf6gflsjg/Screen%20Recording%202023-01-11%20at%203.38.47%20PM.mov?dl=0).