Creating PET acquisition model without a template

[Kiarashdnsh]

Hi

I’m having trouble creating a PET acquisition model. I have read through most of the introductory files, and in all of them, a template_sinogram.hs is used to create the acquisition model. I don't have sinograms for my PET files, and using the provided template doesn't seem to be the correct approach for image reconstruction. How should I proceed from here?

I would greatly appreciate your help.

[KrisThielemans]

There's https://github.com/SyneRBI/SIRF/blob/master/examples/Python/PET/acquisition_data_from_scanner_info.py
but maybe you can elaborate what you want to do. If you have no sinograms, do you want to create some for a scanner that SIRF/STIR already supports? If so, the xample will help.

[Kiarashdnsh]

All my images are 256 * 256 * 256 with voxel sizes of 1 * 1 * 1, with voxel values between 0 and 1 (I attached a sample interfile extracted by XMedCon). I want to create different varieties of low-res and high-res PET images from one original PET image for PVC, whether with AC or other methods possible with SIRF.

m000-31013_AV451_S_E.zip

[KrisThielemans]

sorry, those dimensions are not your "standard" PET, so some resampling has gone. I cannot give you any further recommendations on what scanner to use. You'll have to pick one (and resample the images before projection). You'd need a mumap as well.

Best of luck

[Kiarashdnsh]

@KrisThielemans Yes, these images had undergone preprocessing. However, I have one original PET image. I can also request and gain access to others as well if I could make this work. This image has a timeframe of size 26, dimensions of 256 * 256 * 127, and voxel sizes of 1.4017 * 1.4017 * 2.0313. I also attached the header of this file. (I can also get rid of the timeframe if it complicates things.)
Thanks!

30001_ses-d2430_acq-AV45_pet.zip

[KrisThielemans]

that looks like a Siemens scanner. Have a look if you can find one in Scanner.cxx that has 64 rings (hence 127 image planes) and ring-distance 2*2.0313. There might be more than one of course.

best of luck

[Kiarashdnsh]

@KrisThielemans I looked up the Scanner.cxx master file. There's only one scanner with 64 rings and a ring spacing of 4.0625, which is the Siemens mMR scanner. (If there are other files I must look up, please tell me.) I generated a template and ran the code. Got this error:

ERROR: ProjMatrixByBinUsingRayTracing sadly doesn't support shifted x/y origin yet  
.
.
.
error: ??? '"ProjMatrixByBinUsingRayTracing sadly doesn\'t support shifted x/y origin yet" exception caught at line 677 of /home/sirfuser/devel/buildVM/sources/SIRF/src/xSTIR/cSTIR/cstir.cpp; the reconstruction engine output may provide more information'

I used the original unprocessed PET image for this run.

Many thanks!

• Update: I get this error for almost all of the scanners I generate a template file with.