Faizulaieva O. Methods for improving quality of voice signal extraction for voice authentication of users

The object is the process of user voice signal extraction on the background of external spatially-anisotropic noise interference in systems of voice authentication. The goal is improving the quality of the user speech signal extraction on the background of external spatially-anisotropic noise interference. The methods are observation, comparison, measurement, decision theory, mathematical modeling, full-scale experiment and mathematical statistics. The results are the development of a method for user voice signal recording with selection of a sampling frequency; the development of a method for removal of the first-mode function of the Hilbert-Huang conversion; the development of a method for user voice signal extraction that provides adaptive compensation of spatially-anisotropic noise interference using digital two-channel space-time processing. It was implemented in the research work DR 0113U000358; in the learning process of the Kharkiv National University of Radioelectronics. The field of use is telecommunication systems and networks