[Rtk-users] Reconstruction problems

[Cyril Mory]

Arvid,

You can find a description of the geometry here :
http://www.creatis.insa-lyon.fr/~srit/geometry.pdf
In rtk/applications, several applications are simple converters from the
cone beam manufacturer's raw geometry format to RTK geometry format, so
you can also have a look at the code of these apps.

If you know that you have an almost parallel beam, you can indeed try to
use a dummy geometry file, but be careful: rtksimulatedgeometry
generates a cone beam geometry, with large fan and cone angles if you
leave the default parameters. You can adjust the 'sdd' and 'sid' options
to simulate a very distant source, making rays almost parallel to each
other (be careful with numerical precision, though).

Hope it helps,
Cyril

Le 12/10/2015 21:59, Arvid Piehl Lauritsen Böttiger a écrit :
> Hello Cyril
>
> Thank you for getting back to me. I do have the geometry (albeit not
> right here next to me).
> However, my problem was that when I looked in the geometry file, I
> didn't find it obvious how to define declare it - is it documented
> somewhere?
>
> That said - the measurements I played around with is measured with an
> almost parallel beam. I hoped I could reconstruct it (with minor
> artifacts) using a dummy geometry file.
>
> best
>
> Arvid
>
> On Mon, Oct 12, 2015 at 9:52 PM, Cyril Mory <cyril.mory at uclouvain.be> wrote:
>> Hi Arvid,
>>
>> Unfortunately, you cannot do that. You have to obtain the exact geometry of
>> this particular acquisition from the cone beam device you used.
>> If you need help to convert it to the RTK format, well, that's something you
>> can ask for on this mailing list.
>> Can you get the geometry data ?
>>
>> Regards,
>> Cyril
>>
>>
>> Le 12/10/2015 12:36, Arvid Piehl Lauritsen Böttiger a écrit :
>>> Sure.
>>>
>>> I'll admit I have yet to understand the specifics of the geometry
>>> file. Currently I just used a simulated one because that was what the
>>> SheppLogan example used, and the input projections looked similar.
>>>
>>> Here is a complete list of the commands I executed:
>>>
>>> rtksimulatedgeometry.exe -n 361 -o
>>> c:\Users\aplb\Work\out_dir\geometry360.xml
>>>
>>> rtkprojections.exe --path c:\Users\aplb\Work\tomography_data\dataset1\
>>> --output c:\Users\aplb\Work\out_dir\cylinder.mha --regexp .tif
>>>
>>> rtkfdk.exe -p c:\Users\aplb\Work\out_dir\ -r cylinder.mha -o
>>> c:\Users\aplb\Work\out_dir\cylinder_recon.mha -g
>>> c:\Users\aplb\Work\out_dir\geometry360.xml --spacing 2 --dimension 256
>>>
>>> I did not do anything to the contrast and that is most likely the
>>> issue. I downloaded and installed vv and I can see some (very unclear)
>>> features inside the SheppLogan reconstruction - so that's nice.
>>> However, currently I'm mostly interested in getting the geometry right :)
>>>
>>> best
>>>
>>> On Mon, Oct 12, 2015 at 11:54 AM, louie L <ghostcz at hotmail.com> wrote:
>>>> Hi,
>>>>
>>>> You can specify the size in each dimension. Can you post the arguments
>>>> you used for the projections and reconstruction? Did you try to adjust the
>>>> contrast of the output?
>>>>
>>>> Cheers,
>>>> Louie
>>>>
>>>> Sent from my iOS
>>>>
>>>>> On 12 Oct 2015, at 11:42, Arvid Piehl Lauritsen Böttiger
>>>>> <bottiger at gmail.com> wrote:
>>>>>
>>>>> Hi there
>>>>>
>>>>> I hope someone can help me me get started with RTK.
>>>>>
>>>>> First of all I should mention that I have compiled RTK on Windows 7
>>>>> using Visual Studio 2013. Everything seems to compile and run just
>>>>> fine.
>>>>>
>>>>> Secondly: I tried to reproduce this example:
>>>>> http://wiki.openrtk.org/index.php/RTK/Scripts/FDK
>>>>>
>>>>> This was somewhat successful - I think. I could generate some
>>>>> projections, and reconstruct them. However, it just looks like an egg
>>>>> and not like a Shepp Logan phantom. See attachment.
>>>>>
>>>>> However, I'm unsure if this is expected or not. I took some homemade
>>>>> projections (of a small cylinder) and converted them into a MHA file
>>>>> (using rtkprojections.exe). The file looks correct in ImageJ. The
>>>>> dimensions are 370 pixels * 195 pixels * 361 projections
>>>>>
>>>>> You can get the file here:
>>>>> https://www.dropbox.com/s/p8tsgsh5xxb468f/cylinder.mha?dl=0
>>>>>
>>>>> The problem is that the reconstruction is wrong (see cylinder_recon).
>>>>> It looks like the dimensions are not read correctly, but I am unsure
>>>>> how I fix that? The rtkfdk.exe has a single --dimension parameter, but
>>>>> I am unsure which dimension that is suppose to be, but neither 195,
>>>>> 370 or any other value I could think of works for me.
>>>>>
>>>>> I hope someone can help me in the right direction.
>>>>>
>>>>> best
>>>>>
>>>>> Arvid
>>>>> <SheppLogan.PNG>
>>>>> <cylinder_recon.PNG>
>>>>> _______________________________________________
>>>>> 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
>>
>>
>> ---
>> L'absence de virus dans ce courrier électronique a été vérifiée par le
>> logiciel antivirus Avast.
>> https://www.avast.com/antivirus
>>




---
L'absence de virus dans ce courrier électronique a été vérifiée par le logiciel antivirus Avast.
https://www.avast.com/antivirus