[Gate-users] 3-D image reconstruction using STIR from GATE Results

[Kris]

Dear Claude

STIR does know about this offset. 

However, I think there's is something wrong with the simulation (or on how
Guifen reads the data). If you look at the screenshot if the sinogram (made
with STIR), it's completely impossible that this is a sinogram of an
existing object. So, obviously any image recon will go wrong. In particular,
if you simulate a point source in the centre of the scanner, the ECAT
conventions are such that your sinogram should show a straight-line.

I haven't looked at Guifen's GATE macros to see what he is doing though.

Do you have any sino display tools in GATE that GUifen could use to check
how GATE thinks that the sino looks like?

Kris

> -----Original Message-----
> From: gate-users-bounces at lists.healthgrid.org 
> [mailto:gate-users-bounces at lists.healthgrid.org] On Behalf Of 
> Claude Comtat
> Sent: 06 April 2007 08:57
> To: gate-users at lists.healthgrid.org
> Subject: Re: [Gate-users] 3-D image reconstruction using STIR 
> from GATE Results
> 
> Hi Guifen,
> 
> I do not know if this helps, but in a ecat7 sinogram, the 
> center of the sinogram along the radial direction is offset 
> by the half of the bin size. This is due to the fact that the 
> number of radial bins is even.
> For example, if you have 4 radial bins, numbered from 0 to 3, 
> bin 2 will be at the center. I would imagine that the 
> reconstruction algorithm should know that (at least, the 
> ecat7 based reconstruction algorithms).
> 
> Claude
> 
> Guifen Lin wrote:
> 
> > Hi
> > simulate a point source (x=0,y=0) in the file PET_Ecat.mac.I 
> > reconstructe the image in two mwthod:
> > 1.Method one
> > <1>.
> > cd /usr/local/gate/gate_v3.0.0/examples/my_example/PET
> > source /usr/local/gate/gate_v3.0.0/env_gate.csh
> > /usr/local/gate/g4work/bin/Linux-g++/Gate PET_Ecat.mac ......
> > 
> > <2>.
> > [root at localhost ~]# cd /usr/local/gate/STIR/my_test [root at localhost 
> > my_test]# /usr/local/gate/STIR/bin/ifheaders_for_ecat7
> > MySinogramFile.S
> > Attempt all data-sets (Y) or single data-set (N) [Y/N D:Y]:
> > Data number ? [0,8 D:0]:
> > WARNING: ECAT7 IO: Bin size from header.x_resolution 
> (2.2471) does not 
> > agree with expected value 2.25 for scanner ECAT 962. Using bin size 
> > from header...
> > 
> > <3>.
> > [root at localhost my_test]# /usr/local/gate/STIR/bin/display_projdata
> > MySinogramFile_S_f1g1d0b0.hs
> >  
> > The sinogram shows a clear discontinuity in the line and the 
> > reconstructing image shows  horseshoe shape . A friend said 
> that this 
> > horseshoe shape is usually a result of an offset of 1 pixel 
> along the 
> > x-axis in the sinogram.this means I need to shift my sinogram by n 
> > pixels in the +x or -x direction.How I can shift my sinogram by n 
> > pixels in the +x or -x direction.Do I modify STIR's program?
> > I'm looking forward to your help.
> > Thank you in advance!
> > guifen
> >  
> > 
> > 
> --------------------------------------------------------------
> ----------
> > Mp3疯狂搜-新歌热歌高速下 
> > <http://music.yahoo.com.cn/?source=mail_mailbox_footer>
> > 
> > 
> > 
> --------------------------------------------------------------
> ----------
> > 
> > _______________________________________________
> > Gate-users mailing list
> > Gate-users at lists.healthgrid.org
> > http://lists.healthgrid.org/mailman/listinfo/gate-users
> 
> -- 
> Claude Comtat						
> Laboratoire d'Imagerie Moléculaire Expérimentale	
> CEA/DSV/I2BM/SHFJ
> 
> 
> 
> 
> 
> _______________________________________________
> Gate-users mailing list
> Gate-users at lists.healthgrid.org
> http://lists.healthgrid.org/mailman/listinfo/gate-users
>