Coincidence identification in PET using neural networks

We have implemented a coincidence identification method based on Artificial Neural Networks, ANN. It was used to extract coincidences from four sets of GATE simulated data from a generic small animal PET scanner. For comparison purposes, a conventional sorting method was also implemented for two different time coincidence windows, τ = Δt and τ = 2Δt, where Δt = 10 ns is the time resolution of the scanner.