Solutions to Improve NQR Detection

This chapter seeks to investigate solutions to improve NQR detection using the optimized AlexNet algorithm, proposed in a previous chapter. Initially, the generalization of the algorithm using transfer learning is studied, so that it has performances at least as good as those presented so far, on spectrometers with different transfer characteristics, which can influence the response signal. Then the application of autoencoders to reduce the noise affecting the NQR signal is investigated.