Identical NIFTI files between M0 and M12 #39
Comments
|
I wrote a script to compare the files from M0 and M12. This script is stored in the canproco dataset in The analysis gives the following: out of 1752 files from M12, 177 are identical. The output file can be found here: What are the next steps? |
|
Thank you! I checked the TXT file you generated, and it looks that for the following 32 subjects, the entire List of identical subjects: |
|
This comment summarizes the comparison of the "fixed" dataset (received 2023-11-08 from Erin) with our git-annex dataset. Clone a clean copy of the git-annex canproco datasetCopy the data received ErinCheck for changed files--> 496 changed files Some changes seem to be minor and related mainly to the nii header (possibly due to the usage of another version of the BIDSification tool?). JSON: 
nii header: 
TODOWrite and run a script to compare images subject-wise between |
|
Currently running on |
|
Currently adding the files to branch New folder created:
Files modified:modified: sub-cal084/ses-M0/anat/sub-cal084_ses-M0_STIR.json The above shows that 494 files were modified. Untracked filessub-cal056/ses-M0/anat/sub-cal056_ses-M0_acq-MT_MTS.json The above files are marked as new as they are not yet split into MTon and MToff. Using @valosekj script
|
|
After further investigation, I was able to track the files which have changes in their data : List of files with data changes (in image.data)sub-edm054_ses-M0_T2w.nii.gz This is a total of 183 files (or 33 subjects). This means that only for those files, derivatives need to be reviewed. Output of data transfer : |
|
Joplin script Visual verification of changed images Subjects :
Only two subjects M0 lesion seg on PSIR didn't correspond to the M12 image
Next step :
Question :
|
|
Images This was pushed to branch Note : Subject |
|
Thanks for the investigation, @plbenveniste! 🚀
Great, this is good!
Okay! In other words, it means that the current git-annex lesion segmentation could be renamed from
I will take care of the T2w SC segs (I will check if |
I generated QC to overlay code to generate QClist_of_subjects=(sub-edm054 sub-tor029 sub-mon002 sub-mon005 sub-mon003 sub-edm065 sub-edm035 sub-edm031 sub-mon006 sub-tor051 sub-edm088 sub-tor036 sub-mon009 sub-tor016 sub-mon032 sub-tor043 sub-tor010 sub-mon007 sub-mon060 sub-tor039 sub-tor044 sub-mon011 sub-mon041 sub-tor012 sub-mon014 sub-edm005 sub-tor038 sub-mon031 sub-mon010 sub-mon052 sub-tor021)
cd ~/data/canproco
for sub in ${list_of_subjects};do git annex get ${sub}/ses-M12/anat/${sub}_ses-M12_T2w.nii.gz; sct_qc -i ${sub}/ses-M12/anat/${sub}_ses-M12_T2w.nii.gz -s derivatives/labels/${sub}/ses-M0/anat/${sub}_ses-M0_T2w_seg-manual.nii.gz -p sct_deepseg_sc -qc qc -qc-subject ${sub};done |
|
Thanks, @valosekj, for your investigation :)
#!/bin/bash
list_of_subjects=(sub-edm054 sub-tor029 sub-mon002 sub-mon005 sub-mon003 sub-edm065 sub-edm035 sub-edm031 sub-mon006 sub-tor051 sub-edm088 sub-tor036 sub-mon009 sub-tor016 sub-mon032 sub-tor043 sub-tor010 sub-mon007 sub-mon060 sub-tor039 sub-tor044 sub-mon011 sub-mon041 sub-tor012 sub-mon014 sub-edm005 sub-tor038 sub-mon031 sub-mon010 sub-mon052 sub-tor021)
for sub in ${list_of_subjects[@]} ;do
# Go to subject directory
cd derivatives/labels/${sub}
echo ${sub}
# Copy M0 segmentation to M12
mv ses-M0/anat/${sub}_ses-M0_PSIR_lesion-manual.nii.gz ses-M12/anat/${sub}_ses-M12_PSIR_lesion-manual.nii.gz
mv ses-M0/anat/${sub}_ses-M0_PSIR_lesion-manual.json ses-M12/anat/${sub}_ses-M12_PSIR_lesion-manual.json
mv ses-M0/anat/${sub}_ses-M0_PSIR_seg-manual.nii.gz ses-M12/anat/${sub}_ses-M12_PSIR_seg-manual.nii.gz
mv ses-M0/anat/${sub}_ses-M0_PSIR_seg-manual.json ses-M12/anat/${sub}_ses-M12_PSIR_seg-manual.json
mv ses-M0/anat/${sub}_ses-M0_T2w_seg-manual.nii.gz ses-M12/anat/${sub}_ses-M12_T2w_seg-manual.nii.gz
mv ses-M0/anat/${sub}_ses-M0_T2w_seg-manual.json ses-M12/anat/${sub}_ses-M12_T2w_seg-manual.json
mv ses-M0/anat/${sub}_ses-M0_T2w_labels-manual.nii.gz ses-M12/anat/${sub}_ses-M12_T2w_labels-manual.nii.gz
mv ses-M0/anat/${sub}_ses-M0_T2w_labels-manual.json ses-M12/anat/${sub}_ses-M12_T2w_labels-manual.json
# remove disc label segmentation file
rm ses-M0/anat/${sub}_ses-M0_PSIR_labels-disc.nii.gz
rm ses-M0/anat/${sub}_ses-M0_PSIR_labels-disc.json
ls ses-M0/anat
# Go back
cd ../../..
done
Next steps :
|
|
Currently manully segmenting the M0 files with the following config file : Config file usedFILES_LESION:
- sub-edm054_ses-M0_PSIR.nii.gz
- sub-tor029_ses-M0_PSIR.nii.gz
- sub-mon005_ses-M0_PSIR.nii.gz
- sub-mon002_ses-M0_PSIR.nii.gz
- sub-mon003_ses-M0_PSIR.nii.gz
- sub-edm065_ses-M0_PSIR.nii.gz
- sub-edm035_ses-M0_PSIR.nii.gz
- sub-edm031_ses-M0_PSIR.nii.gz
- sub-mon006_ses-M0_PSIR.nii.gz
- sub-tor051_ses-M0_PSIR.nii.gz
- sub-edm088_ses-M0_PSIR.nii.gz
- sub-tor036_ses-M0_PSIR.nii.gz
- sub-mon009_ses-M0_PSIR.nii.gz
- sub-tor016_ses-M0_PSIR.nii.gz
- sub-mon032_ses-M0_PSIR.nii.gz
- sub-tor043_ses-M0_PSIR.nii.gz
- sub-tor010_ses-M0_PSIR.nii.gz
- sub-mon007_ses-M0_PSIR.nii.gz
- sub-tor014_ses-M0_PSIR.nii.gz
- sub-mon060_ses-M0_PSIR.nii.gz
- sub-tor039_ses-M0_PSIR.nii.gz
- sub-tor044_ses-M0_PSIR.nii.gz
- sub-mon011_ses-M0_PSIR.nii.gz
- sub-mon041_ses-M0_PSIR.nii.gz
- sub-tor012_ses-M0_PSIR.nii.gz
- sub-mon014_ses-M0_PSIR.nii.gz
- sub-edm005_ses-M0_PSIR.nii.gz
- sub-tor038_ses-M0_PSIR.nii.gz
- sub-mon031_ses-M0_PSIR.nii.gz
- sub-mon010_ses-M0_PSIR.nii.gz
- sub-mon052_ses-M0_PSIR.nii.gz
- sub-tor021_ses-M0_PSIR.nii.gzCode used : python manual_correction.py -path-img /Users/plbenveniste/tmp_romane/updating_canproco/canproco -config /Users/plbenveniste/Desktop/canproco_erin_M0_lesion_seg.yaml -path-label /Users/plbenveniste/tmp_romane/updating_canproco/canproco/derivatives/labels -suffix-files-lesion _lesion-manual -fsleyes-dr="-40,70"
|
|
Great! Thanks for the command! Cross-refed to manual-correction wiki. |
|
Subject sub-tor016 at M0 : I am wondering what should be done in this case ? (should the line in between lesions be segmented or should only the lesions be kept or is it only one big lesion ?) |
|
|
|
FYI: I merged the script |
jcohenadad
changed the title
Related to #37 in (this comment).
It seems that some images are identical from time point M0 to M12.
Some examples are:
To do:
The text was updated successfully, but these errors were encountered: