search

pnlbwh / dMRIharmonization

Multi-site dMRI harmonization
Other
43 stars 14 forks source link

pipeline_test.sh results different than readme #7

Closed AndrewDDavis closed 5 years ago

AndrewDDavis commented 5 years ago

I ran the pipeline_test.sh, and it did indeed finish in an hour with 16 cores and 8 GB of RAM. The readme says the result should have been this:

CONNECTOM mean FA: 0.675917931134666 PRISMA mean FA before harmonization: 0.8008729791383536 PRISMA mean FA after harmonization: 0.6366103024088171

However, my result values were different:

CONNECTOM mean FA: 0.6432159691396834 PRISMA mean FA before harmonization: 0.7096586917137321 PRISMA mean FA after harmonization: 0.6152543085182999

Is this difference expected?

tashrifbillah commented 5 years ago

The discrepancy can be attributed to non deterministic registration by ANTs. I wouldn't be worried about this since meanFA changed in the right direction and came closer to that of reference site.

ITKv5 and ANTs 2.3.0 allows deterministic registration. Utilizing that, we shall come up with concrete numbers on a future release.

AndrewDDavis commented 5 years ago

OK, that's good to hear -- thanks for the explanataion.