[Neuroimaging] Nibabel CZI grant - starting work

[Oscar Esteban]

Hi Romain,

I definitely recommend you send a fresh email to the list with your
question.

Long story short, having the possibility to store two affines in the NIfTI
header is actually a very clever idea. But, as often happens, if you then
don't fully implement that idea the result is confusing and even dangerous.

Happy to follow up in some other thread (or in the nitransforms repo -
https://github.com/poldracklab/nitransforms)

Cheer,
Oscar


On Fri, Jul 9, 2021 at 9:32 AM VRomain <romain.valabregue at upmc.fr> wrote:

> Hi
>
> great to see new effort on dicom parsing, (among other)
>
>
> Sorry if this is a bit off topic, but seeing those new effort on nibabel
> (and the milestone on spatial transform), I wonder if you could help to
> solve the nifti ambiguity due to the 2 affine stored in the header
>
> Personnaly, I still do not understand, why this choice has been made,
> and especially  why nobody agree on the interpretation (using multiple
> software, it is easy to get silent bug just because of different sform
> and qform ...
>
> So not sure what the solution, if one get stick to the nifty and the
> possibility to store 2 affines, (I guess one will always get into
> trouble at some point. But may on could encourage an uniform
> interpretation among software ?
>
> Cheers
>
> Romain
>
>
> Le 06/07/2021 à 12:06, Matthew Brett a écrit :
> > Dear all,
> >
> > You’ve already heard echoes about this on the list, but Nibabel now
> > has a Chan Zuckerberg Initiative grant from their Essential
> > Open-source Software for Science series.
> >
> > All the details are here:
> >
> > https://github.com/matthew-brett/czi-nibabel
> >
> > We actually got the grant back in October last year, to start in
> > January, but I (Matthew) had some trouble negotiating time for the
> > work.  That is now resolved and we have the money ready to spend, and
> > the time to work on it.
> >
> > We are planning to get going on the grant work starting this week
> > (starting 5 July).  You can see the general schedule in the work plan
> > in the Github repo above.  In particular, we will get going on the
> > planned milestones:
> >
> > https://github.com/matthew-brett/czi-nibabel/blob/master/milestones.md
> >
> > I will be starting work on the axis labelling / metadata API, and
> > better DICOM integration.  Chris will be ramping up work on the
> > surface API.  Oscar will be extending the surface transform work.
> >
> > We’re emailing now to warn you that you will see more activity appear
> > quite quickly, and we would really appreciate help in reviewing. We
> > would love to hear from any of you who have particular interest in
> > these topics.
> >
> > Cheers,
> >
> > Matthew, Chris, Oscar
> > _______________________________________________
> > Neuroimaging mailing list
> > Neuroimaging at python.org
> > https://mail.python.org/mailman/listinfo/neuroimaging
> _______________________________________________
> Neuroimaging mailing list
> Neuroimaging at python.org
> https://mail.python.org/mailman/listinfo/neuroimaging
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.python.org/pipermail/neuroimaging/attachments/20210709/a6798a8d/attachment.html>