stnava
It has nothing to do with the metric - it's the initialization.
@ntustison let me know what you think about this ... short story is the itk moments calculator easily produces nonsense if the intensity is a zero centered gaussian .... did...
@MaximilianHoffmann yes - that is an issue if intensity is close to zero mean
I believe this will be resolved when we update itk + ants tags - but that may not happen too soon. does "regular" ants compile on the M1?
ok - we have not updated the itk/ants tags for antspyx for quite some time. it generally takes me a day or so to update pypi ... don't have that...
I pushed a bunch of changes to pybind11 as well as various antspyx things .... I had to remove some functionality that is probably not used much anyway as I...
ok - I just pushed a few fixes minutes ago .... I think this latest push gets us very close to full functionality and probably successful on M1.
looks like a PIL issue - maybe reinstall PIL? anyway, I cannot reproduce this one. it may be a couple months or so before we have our own M1 machines...
ok - that symbol `_xcb_connect` is not something I've ever seen in connection with ITK or ANTs.
that's interesting - I never heard of that before.