[Rtk-users] Reconstructing DICOM data

[Simon Rit]

Hi,
Please, as I said before "To start with, I would make sure that everything
outside the leg is at about 0 in your original Ankle CT data." I don't know
the reason but your Ankle.mha file contains crazy values (Min: 9.43646e+07
Max: 9.44526e+07). If I run
rtkforwardprojections -g geoemtry.xml -i AnkleDICOM.mha -o test.mha --like
Ankle.mha  -f CudaRayCast
rtkfdk -g geometry.xml -o Output.mha -p . -r ^test.mha$ --dimension
512,512,245 --hardware cuda --hann 0.4 --hannY 0.4
I have a more decent result. Remove or increase hann values to have a
better spatial resolution.
Cheers,
Simon

On Fri, Sep 23, 2022 at 9:36 AM Sreejith P P <sreejithputhettu at gmail.com>
wrote:

> Hi Simon,
>
> Thank you very much for your quick response.
> As suggested, I have refined the Ankle Dicom data so that everything
> outside the leg part is at 0. I got a fine projection 'Ankle.mha' after
> doing that.
> However, I didn't get much effect when I reconstructed it using rtkfdk. I
> understand that projecting and reconstructing a CT image will necessarily
> result in a loss of quality due to the multiple interpolations. But I
> doubt, maybe something I am missing here can make it better. Could you have
> a look at it, please?
>
> I used the following command for reconstruction.
> *rtkfdk -g geometry.xml -o Output.mha -p . -r Ankle.mha --dimension
> 512,512,245 --hardware cuda --hann 0.4 --hannY 0.4*
>
>  Ankle.mha
> <https://drive.google.com/file/d/1M5hey85gQhh_gCC1dEvvr_EsXkRAOWm3/view?usp=drive_web>
>  AnkleDICOM.mha
> <https://drive.google.com/file/d/14qcXeN3IL_qHT3IbiJfMaQrTIPIPu81v/view?usp=drive_web>
>  geoemtry.xml
> <https://drive.google.com/file/d/105dwnd8dqgeE2d0qz9I9FPzH1t5b2gcc/view?usp=drive_web>
>  output.mha
> <https://drive.google.com/file/d/11COT9G6jzrHALPvUs8N4jnnZt9kuIMd3/view?usp=drive_web>
>
> Regards,
> Sreejith
>
> On Tue, Sep 20, 2022 at 12:44 PM Simon Rit <simon.rit at creatis.insa-lyon.fr>
> wrote:
>
>> Hi,
>> Your projections in Ankle.mha display a very large value in air. To start
>> with, I would make sure that everything outside the leg is at about 0 in
>> your original Ankle CT data.
>> Note that projecting and reconstructing a CT image will necessarily
>> result in a loss of quality due to the multiple interpolations (during
>> projection and backprojection). I would suggest to use fine pixels in the
>> projections (e.g. --spacing .25) which you can bin during reconstruction
>> (--binning 4,4).
>> Simon
>>
>> On Mon, Sep 19, 2022 at 6:06 PM Sreejith P P <sreejithputhettu at gmail.com>
>> wrote:
>>
>>> Hi,
>>>
>>> I have created projections from 'Ankle' CT data using RTK (cuda
>>> projection filter). It looks good to me.
>>> After that, I tried rtkfdk application to reconstruct it.
>>> I used the following command and got the attached output.
>>>
>>> *rtkfdk -g geometry.xml -o Output.mha -p . -r Ankle.mha --dimension
>>> 512,512,245 --hardware cuda --pad 1.0 --hann 0.4 --hannY 0.4*
>>>
>>> Am I doing it correctly? If yes, I would like to know what I can do to
>>> improve the result.
>>>
>>>  Ankle.mha
>>> <https://drive.google.com/file/d/10ybESn9HoU9_LYDgaLUNYIf2BzTarQij/view?usp=drive_web>
>>>  geometry.xml
>>> <https://drive.google.com/file/d/1cTlvPld3iRNiRUW8HmnF29i_j6AmzaPW/view?usp=drive_web>
>>>  Output.mha
>>> <https://drive.google.com/file/d/1iI9V9TUIAKUMyfe-w6TTzFzS0DzvaGyn/view?usp=drive_web>
>>>
>>> _______________________________________________
>>> 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://www.creatis.insa-lyon.fr/pipermail/rtk-users/attachments/20220923/e93cce97/attachment.htm>