[Rtk-users] SART vs FDK

Wed Apr 10 10:09:48 EDT 2013

Thanks for your answer. I'll crop the FDK and SART volumes to the FOV and see if initializing my algorithm this way yields better results.
Regarding ROI + TV, I don't know the literature on the topic, but it turns out it's quite abundant : a simple google search yields many results (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3055906/, http://iopscience.iop.org/0031-9155/54/9/014)

Regards,
Cyril
De : simon.rit at gmail.com [mailto:simon.rit at gmail.com] De la part de Simon Rit
Envoyé : mercredi 10 avril 2013 14:54
À : MORY, CYRIL
Cc : vilaoliva at creatis.insa-lyon.fr; rtk-users at openrtk.org
Objet : Re: SART vs FDK

Hi,
This is a difficult question and I don't know the answer. I think what you observe is the famous edge artifact of SART reconstructions. My feeling is that that one can not trust values outside the FOV in any algorithm. I would mask the FOV if I were you, whatever the algorithm. On real data, I know it costs a bit of time but only a test with each initialization will give you an answer.
BTW, do you know any publication on ROI reconstruction+TV ?
Simon

PS : I'm investigating the maling list issue, I'll let you know when I get an answer.

On Wed, Apr 10, 2013 at 12:59 PM, MORY, CYRIL <Cyril.Mory at philips.com<mailto:Cyril.Mory at philips.com>> wrote:
Hi Simon, Hi Marc,

I've tried to send this mail to the mailing list, but it contains images, so a manual approval from the list's manager is required.
And last time I talked with Simon, he couldn't find the interface to approve or reject the emails. So I forward it.

Regards,
Cyril

De : MORY, CYRIL
Envoyé : mercredi 10 avril 2013 11:54
À : rtk-users at openrtk.org<mailto:rtk-users at openrtk.org>
Objet : SART vs FDK

Hi,

I'm wondering whether the reconstruction results I get with SART and FDK are normal or not. I've performed reconstructions of a moving Shepp and Logan phantom, using 308 projections equidistributed on a 360° circular trajectory. The phantom is fully contained in the FOV. I've used the default projection and back projection operators for both SART (30 iterations of SART instead of the default 5) and FDK. And there are very noticeable differences outside the FOV, especially at the edges of the projection cone, where SART produces high errors while FDK doesn't.

Is this a normal behavior ? I'm using these reconstructions to initialize an iterative algorithm, and my real data is truncated, so I'd prefer to use SART than FDK. These high errors outside the FOV make me hesitate, though.

SART :
[cid:image001.jpg at 01CE3604.D3EA4D80]

FDK :
[cid:image002.jpg at 01CE3604.D3EA4D80]

Regards,

==========================================
Cyril Mory
PhD student at Philips Medisys and CREATIS

Groupement Hospitalier Est
Hôpital Cardiologique Louis Pradel
Laboratoire CREATIS - Bât. B13
CNRS UMR5220, INSERM U1044, INSA-Lyon, Univ. Lyon 1
28, Avenue du Doyen LEPINE
69677 Bron cedex FRANCE

Office : +33 4 72 35 74 12<tel:%2B33%204%2072%2035%2074%2012>
Cell : +33 6 69 46 73 79<tel:%2B33%206%2069%2046%2073%2079>

________________________________
The information contained in this message may be confidential and legally protected under applicable law. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, forwarding, dissemination, or reproduction of this message is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/rtk-users/attachments/20130410/6ce3a2ba/attachment-0009.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.jpg
Type: image/jpeg
Size: 92933 bytes
Desc: image001.jpg
URL: <http://public.kitware.com/pipermail/rtk-users/attachments/20130410/6ce3a2ba/attachment-0018.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image002.jpg
Type: image/jpeg
Size: 74939 bytes
Desc: image002.jpg
URL: <http://public.kitware.com/pipermail/rtk-users/attachments/20130410/6ce3a2ba/attachment-0019.jpg>