ELECTRONIC APPARATUS FOR SPEAKER RECOGNITION AND OPERATION METHOD THEREOF

A speaker recognition method of an electronic device of the present disclosure comprises: a step of acquiring the first information generated from a speech signal and the second information generated from the speech signal; a step of identifying the first identification information for identifying a...

Full description

Saved in:
Bibliographic Details
Main Authors JUNG DAE JIN, LIM HYOUNG WOO
Format Patent
LanguageEnglish
Korean
Published 09.02.2023
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:A speaker recognition method of an electronic device of the present disclosure comprises: a step of acquiring the first information generated from a speech signal and the second information generated from the speech signal; a step of identifying the first identification information for identifying a speaker of the speech signal based on an output of a first neural network to which the first information is inputted, and identifying the second identification information for identifying the speaker of the speech signal based on the output of a second neural network to which the second information is inputted; and a step of identifying the speaker of the speech signal based on the first identification information and the second identification information. 본 개시의 전자 장치의 화자 인식 방법은, 음성 신호로부터 생성된 제1 정보 및 음성 신호로부터 생성된 제2 정보를 획득하는 단계; 제1 정보가 입력된 제1 뉴럴 네트워크의 출력을 기초로 음성 신호의 화자를 확인하기 위한 제1 확인 정보를 확인하고, 제2 정보가 입력된 제2 뉴럴 네트워크의 출력을 기초로 음성 신호의 화자를 확인하기 위한 제2 확인 정보를 확인하는 단계; 및 제1 확인 정보 및 제2 확인 정보를 기초로 음성 신호의 화자를 확인하는 단계를 포함한다.
Bibliography:Application Number: KR20220087549