Debug about noisy background in recon images( using .lut scanner interation)

General Discussions
1

Dear CASToR users:
I reconstructed 3 images by castor-recon successfully, but I encountered some bugs. They are all from the same ‘.lut’ scanner system interation. The shapes of 3 recon images are right, but they all seems to have one common problem: Severe background noises. The 3 result images are attached as follow:
(1) image quality phantom data from real experimental data collection. Reconstruction command ：“castor-recon -df /home/castor/workDir/cdf/HiPET_IQP.Cdh -opti MLEM -it 2:10 -proj joseph -conv gaussian,1,1,3::psf -dim 300,300,400 -vox 0.336,0.336,0.336 -dout /home/castor/workDir/img/HiPET_IQP -th 0”
(2) point source(Na-22) data from real experimental data collection. Reconstruction command ：“castor-recon -df /home/castor/workDir/cdf/181024-5-x-10mm-step5mm-axial-0mm-singlesV927-nemaPtsrc-3600s-idRun-2-sdatAnaV24-datAnaV10-dat2lmfV4.Cdh -opti MLEM -it 20:10 -proj joseph -conv gaussian,2,2,3.5::psf -dim 220,220,340 -vox 0.336,0.336,0.336 -dout /home/castor/workDir/img/181024-5-x-10mm-step5mm-axial-0mm-singlesV927-nemaPtsrc-3600s-idRun-2-sdatAnaV24-datAnaV10-dat2lmfV4 -th 0”
(3) point source(Na-22) data from GATE simulation. Reconstruction command ：“castor-recon -df /home/castor/workDir/cdf/HiPET_nemaPtsrc_10uCi_10s_csh-noCLCS-lld-0.35-uld-0.65.Cdh -opti MLEM -it 20:10 -proj joseph -conv gaussian,2,2,3.5::psf -dim 220,220,340 -vox 0.336,0.336,0.336 -dout /home/castor/workDir/img/HiPET_nemaPtsrc_10uCi_10s_csh-noCLCS-lld-0.35-uld-0.65 -th 0”

1.png

3.png
3.png1097×522 188 KB

2.png
2.png1125×597 207 KB

2

Hi,

For real datasets, do you include random, scatter and attenuation corrections ? Did you compute a normalization correction ?

Do you have an estimation of the ratio of random and scatter events ?

For GATE data, did you include only true coincidences ? Then did you compute a normalization correction and include it in the reconstruction ?

For the quality phantom data, I suggest that you run more iterations. The “stars” artifacts are likely due to that.

Simon