[Gate-users] Couldn't figure out the organization of sinogram data (288, 288, 1024)

[Arda Konik]

Hi,
I believe I have the answer for this question that i asked a week ago.  
I first reduced the problem to 4 rings (i.e.  , setting  "  /gate/ 
crystal/cubicArray/setRepeatNumberZ 1 " instead of 8 which corresponds  
to the 32-ring scanner) and 16 2D-sinograms. Then, i ran several short  
simulations placing spherical sources (one  or 2 sources per  
simulation) at different locations in a cylindrical water phantom. For  
each simulation, out of 16 possible sinograms only 2 (for 1 source) or  
4 (2 sources) sinograms appeared in the output data which seems  
reasonable (at least to me) when you draw the scanner and guess the  
ring pairs to create those sinograms. As a conclusion i found that
the data is organized as below, where "rd" is the ring difference.
{ rd=0, rd = 1, rd= -1 , rd = 2, rd= -2, rd = 3, rd= - 3 } or
more explicity in terms of ring numbers :
{ (1,1),(2,2),(3,3),(4,4),  (1,2),(2,3),(3,4),(2,1),(3,2),(4,3),    
(1,3),(2,4), (3,1),(2,4),   (3,4), (4,3) }
as you can see there are 16 ring pairs.
I am assuming for large ring numbers this organization will be in  
similar fashion.
To visualize these slices i used Osirix (image processing, freeware)  
on a Mac OS.
http://www.apple.com/downloads/macosx/imaging_3d/osirix.html

i hope this will be helpful for some users
cheers,

arda konik
University of Iowa


On Dec 10, 2008, at 2:23 PM, Arda Konik wrote:

> Hi,
>
> Could anyone explain the organization of sinogram output from  "  
> PET_Ecat_System.mac " example, please?
> A 32 ring system generates 1024 2D sinograms, and they must be  
> indexed by ring difference ( Ri - Rj) and axial position (Ri + Rj)/2  
> according to the info file and the Gate manual. However, i couldn't  
> even figure out which ones are direct sinograms ( Ri - Rj =0) out of  
> these 1024 sinograms.
>
> I tried to understand the organization by simulating the acquisition  
> from sources at different locations and shapes; decreasing ring  
> number but still couldn't solve this except that i can see a pattern  
> due to ring difference indexing, when i plot the total counts of the  
> sinograms...What other approach would you recommend to figure out  
> the data organization?
>
> Thanks,
> arda konik
> University of Iowa
> <ATT00001.txt>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opengatecollaboration.org/mailman/private/gate-users/attachments/20081218/f5f04776/attachment.htm>