On-clock non-paralyzable count-loss model

Abstract

This paper proposes a new count-loss model that is applicable to radiation detection systems in the field of nuclear medicine. The proposed model, represented by n = (1 − exp(−τn0))/τ (n0: input rate, n: output rate,τ: system clock duration), gives higher count rates than the well-known standard non-paralyzable (non-extensible) model represented by n = n0/(1 + τn0) does, being equally simple. This model assumes that a part of the data processing is performed on a fixed system clock, requiring a data buffering function (latch) to retain digital event information temporarily, and that input timing is random. The model's simple interpretation and expression also offers an advantage over various other advanced count-loss models proposed so far. This model was found to be a better fit to a data-acquisition system for a positron emission tomography scanner.