[Rtk-users] RTK code correction

[Simon Rit]

Regarding the field of view issue, I have added an option to allow
adequate FOV computation with a displaced detector
https://github.com/SimonRit/RTK/commit/ae71a4bb33a88af1ceb86f283ac215e6ba095062
This is basically what you have suggested but you have to say
explicitely that you use a displaced detector. See
https://github.com/SimonRit/RTK/commit/9a12222ae0d4f527b09fcc83513f436070538c4b
Simon

On Tue, Aug 27, 2013 at 10:59 AM, Uffe Bernchou <Uffe.Bernchou at rsyd.dk> wrote:
>
>
> Dear Simon,
>
>
>
> I’m writing you because I am a user of the nice reconstruction toolkit that
> you generously have shared on openrtk.org. Colleagues and I are working on
> imaging in radiotherapy, and have been looking for an open source algorithm
> for reconstruction of cone beam data from Elekta’s XVI system.
>
>
>
> During our work we have discovered a small bug in
> rtkFieldOfViewImageFilter.txx. The code works fine with cone beam images
> where the detector is in the “small” position. However, for M scans only a
> very small field of view is displayed. The problem seems to be related to
> the following lines:
>
>
>
> m_Radius = std::min( m_Radius, vcl_abs(
> sourceOffsetX+mag*(corner1[0]+projOffsetX-sourceOffsetX) ) );
>
> m_Radius = std::min( m_Radius, vcl_abs(
> sourceOffsetX+mag*(corner2[0]+projOffsetX-sourceOffsetX) ) );
>
>
>
> With this code, only the intersection where space imaged during every frame
> is in the FOV. When the panel is in other positions than small, this will
> created problems. A small hack that changes this code to
>
>
>
> m_Radius = std::min( m_Radius, std::max(vcl_abs(
> sourceOffsetX+mag*(corner1[0]+projOffsetX-sourceOffsetX) ),vcl_abs(
> sourceOffsetX+mag*(corner2[0]+projOffsetX-sourceOffsetX) )) );
>
>
>
> seems to solve the problem. I think you should be aware of this problem, and
> consider to change the code in a future update.
>
>
>
>
>
> One of my colleagues (Rune Slot Thing) is working on improving the image
> quality of our cone beam images by subtraction of scatter in the projection
> images using MC simulations. We are planning to test the results using the
> RTK reconstruction code in order to avoid the attempts to subtract scatter
> by the XVI software.
>
>
>
> Perhaps you would be so kind to inform me on whether any attempts are done
> to subtract scatter in your algorithm?
>
>
>
> Thanking you in advance.
>
>
>
> Best regards,
>
> Uffe
>
>
>
>
>
> ------------------------------
>
> Uffe Bernchou, Ph.d.
>
> Post doc
>
> Laboratory of Radiation Physics, OUH
>
> Institute of Clinical Research, SDU
>
> Email: uffe.bernchou at rsyd.dk
>
> Phone: +45 29796478
>
> Sdr. Boulevard 29, Indgang 97, 5000 Odense C
>
>
>
>