GATE projections to Castor Format, to perform scatter correction

Hello Castor Experts, I've been trying to replicate some older discussion at performing scatter correction in GATE projection files then reconstruct with castor.
Now, i can't perform them in the GATEROOTCASTOR path, it only seeks for Singles.
I'm looking foward to do Triple Energy Window Correction with iodine 131;
I wrote a code in Python, but i cant find in the CASTOR source code how to write the .Cdf binary.
I am able to reconstruct the images showing no errors, but when I open the image it turns into nothing.
I'm attaching the code, as notebook and as python script (Please help me part :slight_smile: ), also I'm including the images from the GATEROOTCASTOR path, and the image generated by the Python Code Cdh and Cdf, i would be very pleased by your help.
It's possible to perform the same but with Simind, but I don't have the access to the sm2castor source code.

João Henrique Castelo
University of Rio de Janeiro, UFRJ , Brazil
IRD, Rio de Janeiro, Brazil (5.72 MB)

Hi João,

I do not have time to look at your program that build the CDF.

However, you say you can’t find how to write the binary datafile.

For SPECT, you can find the complete description in Table 5, section 6.2.2, page 35 of the current documentation.

Beware that scatter correction should be provided as a rate, i.e. counts per seconds. If you directly input the number of scattered counts of your simulation and if your simulation is longer than 1 seconds, it will completely overestimate the scatter correction and will likely result in an empty image.

Finally, the scatter rates in CASToR are used to be added to the forward model in the reconstruction, so it is supposed that this quantity is free of noise. If you have much noise in your simulations, you may smooth the scatter window projection before using it as the scatter correction rates. Otherwise, you can simply remove the number of scatter counts from the prompts, but it will result in positive bias in your images.

Hope this helps !



Envoyé : mercredi 16 mai 2018 04:32