> I'd like to realize the following sentence in a paper:
> "Segmented images were convoluted with the point spread function of the PET
> device".
> This procedure is a step of PVE-correction of (in my case) FDG PET images.
> How and with what sofware should/can the convolution be peformed? What
> PSF-values should be applied in the case of the Siemens ECAT EXACT scanner
> (CTI, Knoxville, Tennessee; 3D mode)?
This will depend on what filter settings were used for the filtered
back-projection reconstruction (or other things for EM reconstructed images),
as well as how the acquisition was done (septa in, with 2D reconstruction or
septa out and 3D reconstruction). I would suggest speaking with your local
PET physicists, who may have measured (and maybe modelled) the PSF(s).
If the PSF is stationary and can be well approximated by a Gaussian of known
FWHM, then you could use the Smooth button of SPM to do the convolution.
Best regards,
-John
|