Anti-noise performance of the pulse coupled neural network applied in discrimination of neutron and gamma-ray

Abstract In this study, the anti-noise performance of a pulse-coupled neural network (PCNN) was investigated in neutron and gamma-ray ( n − $$\gamma$$ γ ) discrimination field. The experiments were conducted two groups. first group, radiation pulse signals pre-processed using Fourier filter to reduce original noise signals, whereas second left untouched simulate an extremely high-noise scenario. For each part, artificial Gaussian with different intensity levels added prior process. aforementioned conditions, PCNN evaluated compared five other commonly used methods discrimination: (1) zero crossing, (2) charge comparison, (3) vector projection, (4) falling edge percentage slope, (5) frequency gradient analysis. experimental results showed that method significantly outperforms outstanding $$\mathrm{FoM}$$ FoM -value at all levels. Furthermore, fluctuations better than those obtained via most only slightly worse comparison zero-crossing under extreme conditions. Additionally, changing patterns Hence, based on results, parameter selection strategy presented. conclusion, is suitable for use application scenarios because its stability remarkable performance. It does not rely strict settings can realize satisfactory over wide range.