Closed tashrifbillah closed 5 years ago
Write LD_LIBRARY_PATH instruction Write MATLAB_Runtime installation instruction
DONE https://github.com/pnlbwh/dMRIharmonization/commit/1c0b3a1d6218fd77f1dbaed0c3a41b461fa0c5f2
Think about not clipping b0 to 1. The clipping is only performed during normalizing by b0: Fixed by https://github.com/pnlbwh/dMRIharmonization/commit/4cca4a10c96ef1340de6966abee90ed79644c102
Observe a harmonized image and see if there are white dots around the skull region. Since there were white dots before as a consequence of infinite scaling, we clipped scale maps at 10.
Make a note in README.md about unring fsl exectuable: requires Centos7 DONE https://github.com/pnlbwh/dMRIharmonization/commit/1c0b3a1d6218fd77f1dbaed0c3a41b461fa0c5f2#diff-04c6e90faac2675aa89e2176d2eec7d8
Think about how to best handle reproducible template in local testing. One obvious way would be to change the antsMultivariateTemplateConstruction2.sh
and roll back to pristine script later. antsMultivariateTemplateConstruction2.sh
script has been put in lib directory
Not clipping the scale at 10. results in following difference in the skull region ...
It may be an attribute of morphological operation.
Scale clipping fixed by https://github.com/pnlbwh/dMRIharmonization/commit/e8b17eca06aa05e2ee547fec6827e63f71e3807c
Update version tag when making a release
Settle on diffusivity measure extraction method:
dti.py
Think about whether to clip scale at 10:buildTempaltes.py
DONE antsRegistrationSyNQuick.sh random seed or not:cleanOutliers.py
? random seed specified for now ... Think about not clippingb0
to 1. The clipping is only performed during normalizing by b0 this is okay Update README.md to have either FSL or DIPY results