search

Mouse-Imaging-Centre / pydpiper

Python code for flexible pipeline control
Other
24 stars 10 forks source link

Should stats-kernel default depend on resolution of model (or --resolution)? #418

Open gdevenyi opened 5 years ago

gdevenyi commented 5 years ago

Should the stats kernel chosen depend on the resolution of the data, rather than a hard-coded default?

gdevenyi commented 5 years ago

With our recent work with adult, neonate and embryo mice I was wondering if anyone could provide any feedback on this?

bcdarwin commented 5 years ago

Probably, but I haven't done any experimentation to determine what is optimal. In principle this might depend on the registration algorithm as well.

It's trivial to change the default for a pipeline, but having not investigated this I've just been using 200 micron blurring for rodent imaging, although it's likely that for embryos and possible neonates something else should be used.