Open AccessVOICE SPEAKER IDENTIFICATION AS ONE OF THE CURRENT BIOMETRIC METHODS OF IDENTIFICATION OF A PERSON
Author(s) -
K. Lutsenko,
K. Nikulin
Publication year - 2020
Publication title -
teorìâ ta praktika sudovoï ekspertizi ì krimìnalìstiki
Language(s) - English
Resource type - Journals
eISSN - 2708-5171
pISSN - 1993-0917
DOI - 10.32353/khrife.1.2019.18
Subject(s) - biometrics , computer science , identification (biology) , speech recognition , speaker recognition , realization (probability) , speaker identification , key (lock) , artificial intelligence , computer security , mathematics , statistics , botany , biology
The article deals with the most widespread biometric identification systems of individuals, including voice recognition of the speaker on video and sound recordings. The urgency of the topic of identification of a person is due to the active informatization of modern society and the increase of flows of confidential information.
The branches of the use of biometric technologies and their general characteristics are given. Here is an overview of the use of identification groups that characterize the voice. Also in the article the division of voice identification systems into the corresponding classes is given.
The main advantages of voice biometrics such as simplicity of system realization are considered; low cost (the lowest among all biometric methods); No need for contact, the voice biometry allows for long-range verification, unlike other biometric technologies.
The analysis of existing methods of speech recognition recognition identifying a person by a combination of unique voice characteristics, determining their weak and strong points, on the basis of which the choice of the most appropriate method for solving the problem of text-independent recognition, Namely the model of Gaussian mixtures, was carried out.
The prerequisite for the development of speech technologies is a significant increase in computing capabilities, memory capacity with a significant reduction in the size of computer systems. It should also be Noted the development of mathematical methods that make it possible to perform the Necessary processing of an audio signal by isolating informative features from it.
It has been established that the development of information technologies, and the set of practical applications, which use voice recognition technologies, make this area relevant for further theoretical and practical research.