A comparison of resolution recovery performed in projection-space and image-space for a high resolution small animal PET scanner

Advances in PET scanner design and image reconstruction methods have lead to the development of high resolution emission tomographic scanners. Traditional PET reconstruction methods that do not take into account the mispositioning of the lines of response (parallax error) result in blurring images and increased distortion with increased distance from the center of the field-of-view. The image quality can be improved by performing resolution recovery (RR) using a point spread function (PSF) which describes the response of the system to point source. For a high-resolution small animal PET scanner, we studied the resolution recovery modeling in projection-space using a Gaussian-PSF or Monte-Carlo (MC) simulated PSF. We also studied the RR in image-space using the Gaussian-PSF. For this work, we used data that were acquired by scanning a Micro Deluxe phantom and a healthy mouse. The resolution recovery approach which used Gaussian-PSF, with a fixed FWHM, improved the spatial resolution and enhanced image contrast compared to the standard image reconstruction. The resolution recovery method which used the MC simulated spatially variant-PSF further improved the resolution compared to the simple Gaussian-based PSF.