Improvement of energy spectrum characteristics of CdZnTe semiconductor detector by means of digital waveform processing with neural network algorithm

We have studied the improvement of gamma-ray energy spectrum characteristics of a CdZnTe semiconductor detector by means of a neural network algorithm. The neural network recognizes the pulse shapes and determines the corrective magnification factors of pulse heights. After the corrective procedure with the network, the gamma-ray energy spectra of some radioisotope sources become more ideal with better energy resolution.