Optimisation of the event-based TOF filtered back-projection for online imaging in total-body J-PET

We perform a parametric study of the newly developed time-of-flight (TOF) image reconstruction algorithm, proposed for real-time imaging in total-body Jagiellonian PET (J-PET) scanners. The asymmetric 3D filtering kernel is applied at each most likely position electron-positron annihilation, estimated from emissions back-to-back γ-photons. optimisation its parameters studied using Monte Carlo simulations 1-mm spherical source, NEMA IEC and XCAT phantoms inside ideal J-PET scanner. combination high-pass filters which included TOF filtered back-projection (FBP), resulted spatial resolution, 1.5 times higher axial direction than conventional FBP. For realistic 10-minute scans XCAT, require trade-off between noise need Gaussian components, coupled with median post-filtering, demonstrated. best sets filter were obtained by Nelder-Mead minimisation mean squared error resulting reference images. approach allows training algorithm custom scans, phantom, when temporal resolution below 50 ps. quality parameters, outcomes, systematically better non-TOF