[Rtk-users] Issues with displaced detector filter when running FDK inline with acquisition

Simon Rit simon.rit at creatis.insa-lyon.fr
Tue Sep 20 20:56:52 UTC 2022


The problem is clear and it is obviously a bug. I suggest to continue the
discussion on the Github issue, I have already added a comment
<https://github.com/SimonRit/RTK/issues/509#issuecomment-1252901973>.

On Tue, Sep 20, 2022 at 11:08 AM Nick Staut <nick.staut at smartscientific.nl>
wrote:

> Hey Simon,
>
>
>
> Thanks for the quick reply.
>
>
>
> I did not notice that manual setting was not possible. The set and get
> functions seemed to return the expected values before requesting the
> output. The alternative previous version displaced detector filter you
> suggested runs without any issues.
>
>
>
> Kind regards,
>
> Nick
>
>
>
> *Van:* Simon Rit <simon.rit at creatis.insa-lyon.fr>
> *Verzonden:* Tuesday, September 20, 2022 9:25 AM
> *Aan:* Nick Staut <nick.staut at smartscientific.nl>
> *CC:* rtk-users at public.kitware.com
> *Onderwerp:* Re: [Rtk-users] Issues with displaced detector filter when
> running FDK inline with acquisition
>
>
>
> Hi Nick,
>
> I stopped caring about the inline application and removed it as it did not
> seem to be used... One of the things that has been developed is a new
> weighting scheme for offset detectors, see doi.org/10.1118/1.4945418. In
> this filter, you cannot specify the minimum and maximum offsets manually:
>
>
> https://github.com/SimonRit/RTK/blob/master/include/rtkDisplacedDetectorForOffsetFieldOfViewImageFilter.hxx#L81-L89
>
> Things should be different with the previous version
>
>
> https://github.com/SimonRit/RTK/blob/master/include/rtkDisplacedDetectorImageFilter.hxx#L130-L149
>
> Can you try this version instead? I have opened an issue to track this
> problem:
>
> https://github.com/SimonRit/RTK/issues/509
>
> Simon
>
>
>
> On Mon, Sep 19, 2022 at 12:17 PM Nick Staut <nick.staut at smartscientific.nl>
> wrote:
>
> Hi everyone,
>
>
>
> I’ve been trying to use RTK to run a reconstruction during acquisition
> which runs without issues and returns an acceptable result if I disable the
> displacedDetectorFilter. However, when I try to enable the displaced
> detector filter before getting sufficient projections (+ more importantly
> for the ddf: geometry entries) I get the following error:
>
> “ITK ERROR: Cannot account for too large detector displacements, a part of
> space must be covered by all projections.”
>
>
>
> I saw that the old rtkinlinefdk implementation (from RTK v1.4 after which
> it seemed discontinued) manually set the minimum and maximum offsets to the
> displacedDetectorFilter, however if I do the same now it still fails.
>
>
>
> To provide a reproducible example I recreated the rtkinlinefdk application
> following the original format (with small modifications to mimic what I do
> in my code these modifications are mentioned at the comment at the top of
> the cxx file). I attached all required files for the example application to
> the e-mail, it should just compile if you recompile RTK with this unzipped
> folder in you application directory. Once I understand the issue with the
> displaced detector filter I could also create a pull request for this rtk
> application if this is considered useful.
>
>
>
> Kind regards,
>
> Nick
>
> _______________________________________________
> Rtk-users mailing list
> Rtk-users at public.kitware.com
> https://public.kitware.com/cgi-bin/mailman/listinfo/rtk-users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/rtk-users/attachments/20220920/09ffd1f0/attachment.htm>


More information about the Rtk-users mailing list