A Random Estimation Framework for Variable Coincidence Sorting Methods in Digital PET

In positron emission tomography (PET), the number of random events must be estimated accurately to get a good image contrast and an exact quantitative analysis. Random estimation method like the well-known Singles Rate (SR) usually corresponds to a specific coincidence sorting method (CSM). Furthermore, because the digital PET allows the coincidence to be selected post-acquisition in software, different CSMs can be used in it. In this work, we present a random estimation framework for variable CSMs. This framework needs to count the frequency of opening the coincidence window (CW) and calculate the expectation of the random events in a CW according to the CSM. It is easier to design the random estimation method for some CSMs using this framework. We give the complete expression in two CSMs. In the first CSM, we get the same formula with SR, so we mainly evaluated that in the second CSM via some Monte Carlo simulations. The result shows it gets a good and robust performance in the coincidence data and reconstructed image between different source activity and distribution.