I am tring to recon my own scanner data by pytomography. It looks like very well. But when I wanner check normalization, it looks terrible.
At first, I would like to do normalization by Gate simulation without phantom(while source is backtoback), which means the output data only contain true and random coincidence, so I cannot attenuation. Only use true coincidence to do normalization and recon, it looks good, as below.
Then, I wanner try Gate simulation with phantom(ring from 97 to 100 mm with Polyethylene, and cylinder from 0 to 97 mm with water, while source is F18), which means the output data contain attenuation, random and scatter(like real world). I wanner use true coincidence data to do normalization and recon. But lt looks terrible, as below.
dose It mean I need to do attenuation correction before normalization?
Beside, I’m also learning the function of ‘gate.get_normalization_weights_cylinder_calibration’. It looks like component based normalization, but is different with the fomular(from the paper of ‘Normalization of Monte Carlo PET data using GATE’).
I lost in it, could you give any fomular in any reference?
The function gate.get_normalization_weights_cylinder_calibration can only be used to derive normalization weights in the case where you acquire data from a thin cylindrical phantom (shell) with radius approaching the edge of the scanner. I don’t think you can use the true coincidence data from your phantom to derive the normalization coefficients.
After I changed norm phantom from cylinder to ring (99/100 mm smaller than scanner radius 102 mm), I found there still is artifact around the transverse cylinder(raduis 85 mm), as blow. Besides, It seems same profile between norm and norm_ac.
What’s more, It confused me about the histo, why not [Nr_crystal_axial_bins, Nr_crystal_axial_bins, Nr_crystal_trans_bins, Nr_crystal_trans_bins, Nr_delta_submodule_axial_bins, Nr_delta_submodule_trans_bins, Nr_delta_module_axial_bins, Nr_delta_module_trans_bins, Nr_delta_rsector_trans_bins]. Although I got the same artifact in recon image after adding Nr_delta_submodule_trans_bins and Nr_delta_module_trans_bins.
Lastly, dose it contain the sensitivity between different detectors and crystals? What if I wanner do normalization with real PET system data?
dear Hannah,
please i am working on CT scan simulation using GATE, i got projection data output in this format ( benchmarkCT_000.dat … benchmarkCT_089.dat), i stuck all of this projections in a sinogram matrix (in matlab), i need help how can i do reconstructions using pyTomography?
thank you for your help
best regards
dear Hannah,
thank you for your reply, could you please verify my GATE CT scan code with me , i am not sure about it, i obtained this image of chest using a custom code python reconstruction, this is my email: hassan.ouhadda@ump.ac.ma
regards
