Library Subscription: Guest
Begell Digital Portal Begell Digital Library eBooks Journals References & Proceedings Research Collections
Telecommunications and Radio Engineering

ISSN Print: 0040-2508
ISSN Online: 1943-6009

Volumes:
Volume 76, 2017 Volume 75, 2016 Volume 74, 2015 Volume 73, 2014 Volume 72, 2013 Volume 71, 2012 Volume 70, 2011 Volume 69, 2010 Volume 68, 2009 Volume 67, 2008 Volume 66, 2007 Volume 65, 2006 Volume 64, 2005 Volume 63, 2005 Volume 62, 2004 Volume 61, 2004 Volume 60, 2003 Volume 59, 2003 Volume 58, 2002 Volume 57, 2002 Volume 56, 2001 Volume 55, 2001 Volume 54, 2000 Volume 53, 1999 Volume 52, 1998 Volume 51, 1997

Telecommunications and Radio Engineering

DOI: 10.1615/TelecomRadEng.v75.i12.80
pages 1121-1128

ARTIFICIAL NEURAL NETWORKS IN GAMMA-SPECTRUM BASED RADIONUCLIDE IDENTIFICATION

Anna I. Skrypnyk
National Science Center "Kharkiv Institute of Physics and Technology"

ABSTRACT

Artificial neural networks are widely used to solve a variety of classification problems. In this work, we consider the possibilities to apply them in gamma-spectrum based identification of nuclides in mixed radioactive compounds. We used a combination of the Geant4 software and the Monte-Carlo method to simulate response functions of the CdZnTe detector toward the radiation from 57Co, 137Cs, 131I, 133Ba and 241Am. The next step was to find the numerical values of the CdZnTe detector response toward mixed source radiation. The obtained spectra were applied to train and validate the network. Multilayer feedforward network was constructed using the Neuroph framework. In course of the study the network architecture characteristics were improved.