Planar large field-of-view gamma camera detectors in coincidence imaging ar
e characterized by a high singles count rate, leading to a high acquired ra
ndom coincidence rate. We propose and compare two methods to correct the ra
ndom coincidences that account for the variation of the random count rate d
uring the detector rotation. One method consists of using a uniform distrib
ution for the random coincidences (UD) and the other is based on the single
photon distribution (SPD) on each detector. Unlike the UD method, the SPD
method accounts for the variation of the spatial distribution of the random
coincidences. The UD method leads to an underestimation of the random coin
cidences inside the object and an overestimation outside the object. The di
stribution of the random coincidences is better estimated with the SPD meth
od, which allows one to recover contrast values closer to those expected wi
thout randoms background.