ApplyApplying Deep Learning in gamma-spectroscopy for radionuclide identificationng Deep-learning in gamma-spectroscopy for radionuclide identification

André Gomes Lamas Otero; Ademar José Potiens Júnior; Julio Takehiro Marumo

doi:10.15392/2319-0612.2025.2945

Authors

André Gomes Lamas Otero IPEN - Instituto de Pesquisas Energéticas e Nucleares , Nuclear Energy Research Institute https://orcid.org/0000-0002-7712-164X (unauthenticated)
Ademar José Potiens Júnior Nuclear Energy Research Institute https://orcid.org/0000-0002-4098-0272 (unauthenticated)
Julio Takehiro Marumo Nuclear Energy Research Institute https://orcid.org/0000-0003-3010-9691 (unauthenticated)

DOI:

https://doi.org/10.15392/2319-0612.2025.2945

Keywords:

Deep learning, Nuclide identification, Automated Isotope Identification, Gamma Spectroscopy, Transfer Learning, Monte Carlo Simulation

Abstract

This study presents the results of using a Deep Convolutional Neural Network model on gamma spectrum classification for radioactive waste management. The approach uses a modified version of the VGG-19 architecture, originally developed for image recognition with 1000 mutually exclusive classes. The modified VGG-19 architecture uses a gamma spectrum as input and classifies, on a nonexclusive basis, ten classes representing the ten most common radionuclides at IPEN's Radioactive Waste Management Department (Am-241, Ba-133, Cd-109, Co-57, Co-60, Cs-137, Eu-152, Mn-54, Na-22, Pb-210). Gamma spectra were generated using Monte Carlo simulations created with PENELOPE/PenEasy, simulating an HPGe detector with sources inside a steel drum filled with paper, representing the common content of the drums managed at IPEN's Radioactive Waste Management Department. The data set was augmented by mixing these simulated spectra into new spectra containing up to four radionuclides. Several distances from the detector to the drum (41 cm, 46 cm, 51 cm, and 56 cm) were used to create a representative data set. The data from 56 cm (originally 150 spectra after the argumentation process, 375 spectra) was used for validation. After 250 training epochs, the model achieved consistent performance in the training set, demonstrating the efficiency of the method.

Downloads

Download data is not yet available.

References

[1] LECUN, Y.; BENGIO, Y.; HINTON, G. Deep learning. Nature, Nature Publishing Group, v. 521, n. 7553, p. 436, 2015. DOI: https://doi.org/10.1038/nature14539

[2] SIMONYAN, K.; ZISSERMAN, A. Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556, 2014.

[3] HE, K. et al. Mask r-cnn. In: Proceedings of the IEEE international conference on computer vision. [S.l.: s.n.], 2017. p. 2961–2969.

[4] KALCHBRENNER, N. et al. Neural machine translation in linear time. arXiv preprint arXiv:1610.10099, 2016.

[5] GRAVES, A.; JAITLY, N. Towards end-to-end speech recognition with recurrent neural networks. In: International conference on machine learning. [S.l.: s.n.], 2014. p. 1764–1772.

[6] CHIU, C. et al. State-of-the-art speech recognition with sequence-to-sequence models. CoRR, abs/1712.01769, 2017. Disponível em: <http://arxiv.org/abs/1712.01769>.

[7] SUN, C. et al. Revisiting unreasonable effectiveness of data in deep learning era. In: The IEEE International Conference on Computer Vision (ICCV). [S.l.: s.n.], 2017. DOI: https://doi.org/10.1109/ICCV.2017.97

[8] PAN, S. J.; YANG, Q. A survey on transfer learning. IEEE Transactions on knowledge and data engineering, IEEE, v. 22, n. 10, p. 1345–1359, 2009. DOI: https://doi.org/10.1109/TKDE.2009.191

[9] KELLER, P. E. et al. Nuclear spectral analysis via artificial neural networks for waste handling. IEEE transactions on nuclear science, IEEE, v. 42, n. 4, p. 709–715, 1995. DOI: https://doi.org/10.1109/23.467888

[10] YOSHIDA, E. et al. Application of neural networks for the analysis of gamma-ray spectra measured with a Ge spectrometer. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2002. ISSN 01689002. DOI: https://doi.org/10.1016/S0168-9002(01)01962-3

[11] CHEN, L.; WEI, Y.-X. Nuclide identification algorithm based on k–l transform and neural networks. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, v. 598, n. 2, p. 450 – 453, 2009. ISSN 0168-9002. DOI: https://doi.org/10.1016/j.nima.2008.09.035

[12] KAMUDA, M.; SULLIVAN, C. J. An automated isotope identification and quantification algorithm for isotope mixtures in low-resolution gamma-ray spectra. Radiation Physics and Chemistry, Elsevier, v. 155, p. 281–286, 2019. DOI: https://doi.org/10.1016/j.radphyschem.2018.06.017

[13] Metropolis, N., & Ulam, S. (1949). The Monte Carlo method. Journal of the American Statistical Association, 44(247), 335-341. DOI: https://doi.org/10.1080/01621459.1949.10483310

[14] SEMPAU, J. Peneasy, a structured main program from penelope. Freely available from http://www. upc. es/inte/downloads/penEasy. htm, v. 2006, p. 06–01, 2006.

[15] Householder, Alston S. (June 1941). "A theory of steady-state activity in nerve-fiber networks: I. Definitions and preliminary lemmas". The Bulletin of Mathematical Biophysics. 3 (2): 63–69. doi:10.1007/BF02478220. ISSN 0007-4985. DOI: https://doi.org/10.1007/BF02478220

[16] Bridle, John S. (1990a). Soulié F.F.; Hérault J. (eds.). Probabilistic Interpretation of Feedforward Classification Network Outputs, with Relationships to Statistical Pattern Recognition. Neurocomputing: Algorithms, Architectures and Applications (1989). NATO ASI Series (Series F: Computer and Systems Sciences). 68. Berlin, Heidelberg: Springer. pp. 227–236. doi:10.1007/978-3-642-76153-9_28. DOI: https://doi.org/10.1007/978-3-642-76153-9_28

[17] Soti, Z., Magill, J., & Dreher, R. (2019). Karlsruhe Nuclide Chart – New 10th edition 2018. EPJ Nuclear Sciences & Technologies, 10(1), doi:10.1051/epjn/2024009. DOI: https://doi.org/10.1051/epjn/2019004

[18] CHOLLET, F. et al. Keras. 2015. <https://keras.io>.

[19] ABADI, M. et al. TensorFlow: Large-Scale Machine Learning on Heterogeneous Systems. 2015. Software available from tensorflow.org. Disponível em: <https://www.tensorflow.org/>.

[20] NICKOLLS, J. et al. Scalable parallel programming with cuda. Queue, ACM, New York, NY, USA, v. 6, n. 2, p. 40–53, mar. 2008. ISSN 1542-7730. DOI: https://doi.org/10.1145/1365490.1365500

[21] CHETLUR, S. et al. cudnn: Efficient primitives for deep learning. arXiv preprint arXiv:1410.0759, 2014.

[22] GARCÍA-TORAÑO, E.; PEYRES, V.; SALVAT, F. Pennuc: Monte carlo simulation of the decay of radionuclides. Computer Physics Communications, Elsevier, v. 245, p.106849, 2019. DOI: https://doi.org/10.1016/j.cpc.2019.08.002

[23] KINGMA, D. P.; BA, J. Adam: A method for stochastic optimization. In: BENGIO, Y.; LECUN, Y. (Ed.). 3rd International Conference on Learning Representations, ICLR 2015, San Diego, CA, USA, May 7-9, 2015, Conference Track Proceedings. [s.n.], 2015. Disponível em: <http://arxiv.org/abs/1412.6980>.

[24] FAWCETT, T. An introduction to ROC analysis. Pattern Recognition Letters, North-Holland, v. 27, n. 8, p. 861–874, jun 2006. ISSN 01678655. DOI: https://doi.org/10.1016/j.patrec.2005.10.010

[25] ANSI. American National Standard Performance Criteria for Hand-Held Instruments for the Detection and Identification of Radionuclides. [S.l.], 2007.