Skip to content

BCD Manual Corrections

axelsone edited this page Oct 16, 2017 · 5 revisions

NOTES

These are notes to assist with knowing how to manually intervien when BCD encounters one of it's rare failure cses.

step1_forceFailure.sh step2_runEMSP.sh step3_finalRun.sh

HINTS OF WHERE TO LOOK

JIRA

https://apps.icts.uiowa.edu/jira/browse/PREDICTIMG-3181 https://apps.icts.uiowa.edu/jira/browse/PREDICTIMG-3811

Create an input list from the "denoised" images:

find /Shared/johnsonhj/HDNI/Neuromorphometrics/2012Subscription/Data -name "*_UNM_denoised.nii.gz" >> inputFiles.txt

I ran the "1_forceReportFailure.sh" file, so BCD creates EMSP.nrrd/ EMSP.fcsv files. The bash script is attached here.

For those cases that BCD could estimate a good MSP image: A- I Loaded EMSP.nrrd, EMSP.fcsv files by Slicer and define eyes. B- Renamed the EMSP.nrrd to "${SUBJECT}_maneyes_denoised.nrrd". C- Created the following file adjacent to the nrrd file to indicates that denoising should be skipped: touch ${SUBJECT}_maneyes_denoised.nrrd_noDenoise

For many cases the "EMSP.nrrd" image is just a crap because of bad MSP estimation. For such cases: A- I removed the created EMSP.fcsv file B- I re-ran the 1_forceReportFailure.sh script on the previously created "EMSP.nrrd" file! C- Now, new EMSP.nrrd/ EMSP.fcsv files are created with good qualities. D- I ran all steps of the stage 3 on the above EMSP.nrrd/ EMSP.fcsv files.

Beta-Source code scripts that may be useful can be found at https://github.com/BRAINSia/BRAINSTools/tree/master/BRAINSConstellationDetector/docs