Open AccessNatural Speech Signal Recognition Algorithm
Author(s) -
Oleksandr Ruslanovych Osadchuk
Publication year - 2021
Publication title -
elektronna ta akustična ìnženerìâ
Language(s) - English
Resource type - Journals
eISSN - 2617-0965
pISSN - 2524-2725
DOI - 10.20535/2617-0965.eae.228077
Subject(s) - computer science , speech recognition , variety (cybernetics) , reliability (semiconductor) , voice activity detection , control (management) , speaker recognition , quality (philosophy) , speech processing , artificial neural network , noise (video) , natural (archaeology) , artificial intelligence , power (physics) , philosophy , physics , archaeology , epistemology , quantum mechanics , image (mathematics) , history
Speech recognition technologies are becoming more and more part of our lives, providing a convenient way to control a variety of electronic devices - voice control. One of the current problems that is solved in the development of such control systems is the problem of insufficient accuracy of voice command recognition. Improvements are being made to increase reliability, independence from individual voice characteristics, and reduce the negative impact of background noise on recognition quality.The paper presents an algorithm for recognizing and processing user intentions using a neural network built on the principle of understanding natural language and processing audio signals for use in the user support system.