Gaussian position-weighted center of gravity algorithm for multiplexed readout

Abstract

Readout signal multiplexing is a commonly used method to reduce the electronics cost in positron emission tomography (PET) systems, and the calculation of the scintillation coordinates typically is done by using a center of gravity (CoG) technique due to its simplicity and ease of implementation. This comes with a drawback, since CoG has a non-linear response at the periphery of the detector due to the lack of weights beyond the detector. Detectors with multiplexed readout that are based on finely segmented scintillators and coarsely segmented photosensors are known to suffer from the so-called edge effect where a pile-up of the reconstructed coordinates from the edge crystals is observed. This may lead to incorrect assignment of the events to crystal pixels and result in the formation of erroneous lines of response causing a degradation of spatial resolution and reduction of image contrast. To overcome the abovementioned limitations in gamma-ray detectors with multiplexed photosensor readout, we propose to use a modified Gaussian position-weighted center of gravity (PW-CoG) technique for the calculation of gamma-ray interaction position. Here, the proposed method is applied to PET detectors with 24 [Formula: see text] 24 LYSO crystals read out by 8 [Formula: see text] 8 SiPM array with 64:16 row/column multiplexing. Furthermore, we compared the modified Gaussian PW-CoG and truncated center of gravity coordinate reconstruction methods. It was observed that both algorithms resolve peaks corresponding to events registered in the crystal pixels on the periphery of the crystal array. However peak-to-valley ratios and crystal resolvability metrics for the PW-CoG algorithm are generally greater.