A popular method for gamma-ray detection uses scintillation detectors made of crystals optically coupled to photomultiplier tubes (PMTs). Scintillators are widely used in medical radiation fields such as CT scanners, gamma cameras, and positron emission tomography (PET) scanners ‎[1]-‎[5]. The scintillation crystal responds to the absorption of gamma rays by emitting a light pulse. This light pulse is characterized by special properties of the crystal such as the decay time constant. Then the PMT generates an electrical pulse related to the absorbed gamma energy. The phenomenon known as parallax error or depth of interaction (DOI) error [5] which reduces the sensitivity and reconstruction quality of PET, most likely occurred when photons enter the detector at a non-perpendicular angle. Phosphorus sandwich detectors (phoswich) [1] are considered one of the methods used to reduce parallax error. The phoswich detector is a stack of two or more different scintillation crystals; i.e. with different decay time constants, optically coupled to a single PMT. Then the DOI error is reduced when the scintillated crystal is identified. Crystal identification (CI) requires the application of one of the pulse shape discrimination (PSD) methods ‎ [6] – ‎ [20]. Several PSD algorithms have been developed which can be classified into two categories: time domain and frequency domain. In the time domain, cross correlation ‎ [10], fuzzy logic ‎ [11], and neural network ‎ [12] ‎ have been employed in PSD and CI methods. On the other hand, fast Fourier transform (FFT) of pulses improved CI performance ‎‎[13]. Furthermore, the normalized minimum sample (NLS) method ‎[14], which is highly dependent on... middle of the paper.... Saleh and M.A. Ashour, “A Zernike Moment Method for Pulse Shape Discrimination in PMT-Based PET Detectors ,” Nuclear Science, IEEE Transactions on, pp. 1-9, 2013.[25] V. Vapnik, “Statistical Learning Theory”, John Wiley & Sons Inc., New York, 1998.[26] V. Vapnik, "The nature of statistical learning theory", Springer Verlag, Berlin, 1995.[27] G. Amayeh, A. Erol, G. Bebis, M. Nicolescu, "Accurate and efficient computation of high-order Zernike moments", International Symposium on Visual Computing, (LNCS, Vol.3804), Lake Tahoe, Nevada, pp . 462 -469, 2005.[28] J. Gu, H. Z. Shu, C. Toumoulin, L. M. Luo, “A new algorithm for fast computation of Zernike moments,” Pattern Recognition, vol. 35, no. 12, pp. 2905-2911, 2002.[29] S. Hwang, W. Kim, “A new approach to fast computation of Zernike moments,” Pattern Recognition, vol. 39, no. 11, page. 2065–76, 2006.