A Zernike Moment Method for Pulse Shape Discrimination in PMT-Based PET Detectors

This paper investigates a new pulse shape discrimination method (PSD) method based on Zernike moment (ZM). This method is applied to get the depth of interaction (DOI) information of scintillation detectors for a positron emission tomography (PET) scanner. The number of moments is studied in order to find the optimal moment number that gains high identification results. It is found that the moments up to order ten are enough to get similar results as that of higher orders. The ZMs based PSD gains high performance at 99.43% for 10,000 pulses of LSO (t_r ~ 40 ns)-LuYAP (Russian) ( t_r ~ 20 ns) crystals. Other recent PSD algorithms that based on the Fast Fourier Transform (FFT), Normalized least sample (NLS) and Discrete wavelet transform (DWT) give discrimination performances with rates of 98.25%, 97%, and 99.78% respectively. For measured 100,000 pulses of LSO-LuYAP (PML type2), the DWT based method has a discrimination rate of 88% while the ZM method gains a higher rate of 91.8%.