METHOD AND SYSTEM FOR CORRECTING SPEAKER DIARISATION USING SPEAKER CHANGE DETECTION BASED ON TEXT

Disclosed are a method and system for correcting a speaker diarisation using a speaker change detection based on a text. The method for correcting the speaker diarisation according to one embodiment may comprise: a step of performing the speaker diarisation for an inputted audio stream; a step of re...

Full description

Saved in:
Bibliographic Details
Main Authors GEONMIN KIM, HEO HEE SOO, KWON YOUNGKI, LEE BONG JIN, JUNG NAMKYU, LEE CHAN KYU
Format Patent
LanguageEnglish
Korean
Published 17.08.2022
Subjects
ACOUSTICS
MUSICAL INSTRUMENTS
PHYSICS
SPEECH ANALYSIS OR SYNTHESIS
SPEECH OR AUDIO CODING OR DECODING
SPEECH OR VOICE PROCESSING
SPEECH RECOGNITION
Online AccessGet full text

Cover

More Information
Summary:Disclosed are a method and system for correcting a speaker diarisation using a speaker change detection based on a text. The method for correcting the speaker diarisation according to one embodiment may comprise: a step of performing the speaker diarisation for an inputted audio stream; a step of recognizing a speech included in the inputted audio stream to convert into text; a step of detecting a change in speaker based on the converted text; and a step of correcting the speaker diarisation based on the detected change in speaker. Therefore, the present invention is capable of correcting an error for a point of change in speaker. 텍스트 기반의 화자변경검출을 활용한 화자분할 보정 방법 및 시스템을 개시한다. 일실시예에 따른 화자분할 보정 방법은 입력된 오디오 스트림에 대한 화자분할을 수행하는 단계, 상기 입력된 오디오 스트림에 포함된 음성을 인식하여 텍스트로 변환하는 단계, 상기 변환된 텍스트에 기반하여 화자 변경을 검출하는 단계 및 상기 검출된 화자 변경에 기반하여 상기 화자분할을 보정하는 단계를 포함할 수 있다.
Bibliography:Application Number: KR20210017814