SPEAKER RECOGNITION SYSTEM THROUGH ACCUMULATED VOICE DATA ENTERED THROUGH VOICE SEARCH

The present invention relates to a system, method, and computing apparatus for recognizing a speaker through accumulation of voice data input during voice search. The system for recognizing a speaker through accumulation of voice data input during voice search comprises a broadcast device, a search...

Full description

Saved in:
Bibliographic Details
Main Authors CHUN, WANG SUNG, CHOI, JAE WON
Format Patent
LanguageEnglish
Korean
Published 03.05.2016
Subjects
ACOUSTICS
CALCULATING
COMPUTING
COUNTING
ELECTRIC DIGITAL DATA PROCESSING
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:The present invention relates to a system, method, and computing apparatus for recognizing a speaker through accumulation of voice data input during voice search. The system for recognizing a speaker through accumulation of voice data input during voice search comprises a broadcast device, a search word extraction server, and a speaker recognition server. The broadcast device receives voice during voice search, and includes a voice data generation unit which generates voice data from the received voice. The search word extraction server receives the voice data from the broadcast device, and includes a voice data transmission and reception unit which transmits the voice data to the speaker recognition server and a search word extraction unit which extracts a search word by analyzing the voice data. The speaker recognition server includes a voice data accumulation unit which receives the voice data from the search word extraction server and accumulates the voice data, a voice pattern analysis unit which analyzes the accumulated voice data into a voice pattern of the voice data, and a user recognition unit which extracts a user ID based on the obtained voice pattern. 음성 검색 시 입력된 음성 데이터를 누적하여 화자를 인식하는 시스템은 방송 디바이스, 검색어 추출 서버 및 화자 인식 서버를 포함하되, 상기 방송 디바이스는 음성 검색 시 음성을 입력받고, 상기 입력된 음성으로부터 음성 데이터를 생성하는 음성 데이터 생성부를 포함하고, 상기 검색어 추출 서버는 상기 방송 디바이스로부터 상기 음성 데이터를 전송받고, 상기 음성 데이터를 상기 화자 인식 서버로 전송하는 음성 데이터 송수신부 및 상기 음성 데이터를 분석하여 검색어를 추출하는 검색어 추출부를 포함하고, 상기 화자 인식 서버는 상기 검색어 추출 서버로부터 상기 음성 데이터를 전송받아 누적하는 음성 데이터 누적부, 상기 누적된 음성 데이터로부터 상기 음성 데이터의 음성 패턴을 분석하는 음성 패턴 분석부 및 상기 분석된 음성 패턴에 기초하여 사용자 ID를 추출하는 사용자 인식부를 포함한다.
Bibliography:Application Number: KR20140144269