Spatially variant resolution modelling for iterative list-mode PET reconstruction

A spatially variant resolution modelling technique is presented which estimates the system matrix on-the-fly. The method randomly redistributes the line-of-response endpoints according to probability density functions describing the detector response and photon acollinearity effects. A list-mode OSEM algorithm is used for the reconstruction. We demonstrate that this model agrees with measured PSFs and we present results showing an improvement in resolution recovery as compared to using a Gaussian convolution to model the resolution, although with an increase in noise. We also present results from applying this method to event-by-event rigid motion correction with list-mode reconstruction using the microPET Focus220 scanner.