[Gate-users] backprojection of coincidences

[Marlen Priegnitz]

Dear Gate-Users,

I found some strange result for the backprojection of coincidences and 
cannot find any explanation for it. Maybe somebody can help me with this?
This is what I did:
I was simulating a PETscanner and saving the coincidences in an ASCII-File.
My source is a voxelized cylinder (diameter 180 mm) of water with some 
background activity (0.01 Bq per voxel) and inside there are some 
spheres (water) with activity 0.04 Bq per voxel.
I have chosen this low activity to get in principle no randoms.
I'm simulating Positrons.
When I do the backprojection of the registered coincidences, then I 
found, that approx. 25 % of the registered coincidences lie outside of 
the source. (Also if I set the background activity to zero and have only 
the hot spheres, I get this result)
How can this be? I don't believe that this are randoms, because when I 
simulate a delayed time window, then I got nearly no coincidences.
Can anybody help me with this?
Did anybody observed something similar?
I'm grateful for each hint.
Thanks in advance.

Marlen