[Rtk-users] Helical cone-beam reconstruction with RTK

[Simon Rit]

Hi,
We haven't worked on helical (CB)CT and I haven't heard of anyone working
on it. Others answers are correct:
- it's easy to simulate helical CBCT by playing with the geometry
<http://www.openrtk.org/Doxygen/classrtk_1_1ThreeDCircularProjectionGeometry.html>.
I guess compared to a circular geometry where only source-to-detector and
source-to-isocenter distances matter, you'll have to adjust the
SourceOffsetY and the ProjectionOffsetY. That should work for shepp logan
type of simulations and forward projections of 3D images.
- iterative reconstruction should work since we don't have any prior on the
geometry except we use the displaced detector for weighting the central
area when required. You probably want to do something similar in the
direction of the axis of rotation to have smooth transitions between
projections but it's theoretically not mandatory (we found the result is
better in practice).
- there is no analytical reconstruction available.
Good luck and don't hesitate to contribute if you do some tests.
Simon

On Thu, Feb 26, 2015 at 10:33 PM, louie L <ghostcz at hotmail.com> wrote:

> Hi,
>
> You can change the geometry. If changed correctly, most, if not all, the
> iterative cone beam reconstruction will work.
>
> Best regards,
> Louie
>
> Greetings from my iOS!
>
> On 26 Feb 2015, at 21:19, Robert Calließ <robert.calliess at gmx.de> wrote:
>
> Hello Vahid,
>
> I think you need to add a new reconstruction filter for it. For the first
> projection the object usually is placed so that the
>
> resulting projection image covers the upper half of the detector. And you
> will need some additional handling (z-increment).
>
> I guess you could do a little work around with the current rtk. If you
> only reconstruct an rectangular roi around
>
> the detector’s center where the perspective distortion (because of the
> cone) is minimal. And then you add these
>
> seperate volumes to a final volume. This is just an idea. I do not know
> how the result will look like. Think of it like a TDI
>
> but with a wider sensor array.
>
>
>
> I do not know the code in detail, but if it were possible to have seperate
> geometry settings for every projection image it
>
> would be possible to model the helical trajectory by the changing geometry
> without changing /adding code.
>
>
>
> But let’s see what grandmasters of RTK will suggest.
>
>
>
> Best regards,
>
> Robert
>
>
>
>
>
> *Von:* Rtk-users [mailto:rtk-users-bounces at public.kitware.com
> <rtk-users-bounces at public.kitware.com>] *Im Auftrag von *vahid ettehadi
> via Rtk-users
> *Gesendet:* Donnerstag, 26. Februar 2015 20:35
> *An:* rtk-users at public.kitware.com
> *Betreff:* [Rtk-users] Helical cone-beam reconstruction with RTK
>
>
>
> Dear RTK users,
>
>
>
>
>
> I am using the RTK code to reconstruct the image for a bench-top micro-CT
> in a circular cone-beam geometry. It works well with FDK algo. Thanks RTK.
>
> I would like to change the current source trajectory from circular to
> helical. I was wondering is it possible to use the RTK for helical
> reconstruction only with some simple modification?
>
> I don't know well about reconstruction process and I don't know which type
> of modification is necessary to use RTK for a helical cone beam geometry.
> So, any suggestion and recommendation would be appreciated.
>
>
>
> Sincerely,
>
> Vahid
>
>
>
>
>
> _______________________________________________
> Rtk-users mailing list
> Rtk-users at public.kitware.com
> http://public.kitware.com/mailman/listinfo/rtk-users
>
>
> _______________________________________________
> Rtk-users mailing list
> Rtk-users at public.kitware.com
> http://public.kitware.com/mailman/listinfo/rtk-users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.creatis.insa-lyon.fr/pipermail/rtk-users/attachments/20150227/228965fd/attachment.htm>