Instruments and Systems: Monitoring, Control, and Diagnostics Annotation << Back Application of Technology of Neural Networks in the Security System of Information Systems T.I. KASATKINA, L.V. ROSSIKHINA, A.V. DUSHKIN, R.V. KUZMENKO, N.I. GONCHAROV The article shows a variant of using neural network technologies as a tool for converting biometric data into codes for access keys of various lock-key type devices, as well as for upgrading existing biometric technologies. The paper highlights the advantages of using neural network technologies to create access keys. A number of features inherent in authentication systems designed for limited access to confidential information and information resources, and based on the use of neural network technologies to create access keys based on biometric data, are presented. A universal algorithm for recovering the user input biometric information vector is proposed and the distribution for the zero generation of Hamming codes in comparison with the original and next generations of biometric images acting as input data of a neural network is presented. A scheme is presented and the approach of pairwise morphing of biometric images based on the concept of pairwise morphing with the choice of biometric images belonging to different sets and most distant from each other is proposed. It is shown that biometric authentication based on neural network technology can be used as a standalone user authentication system, and as part of the security system of automated control systems. Keywords: neural network, biometric image, access code, authentication, morphing. DOI: 10.25791/pribor.06.2019.693 Contacts: E-mail: rossihina_lv@mail.ru Pp. 31-38.