Speech detection apparatus under noise environment and method thereof
PURPOSE: A method for detecting a voice signal in noise environment and an apparatus thereof are provided to find out the voice signal without errors by using a basis function trained by an independent component analysis. CONSTITUTION: A mike signal(302) including a voice is inputted. A generation c...
Gespeichert in:
Hauptverfasser: | , |
---|---|
Format: | Patent |
Sprache: | eng ; kor |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | PURPOSE: A method for detecting a voice signal in noise environment and an apparatus thereof are provided to find out the voice signal without errors by using a basis function trained by an independent component analysis. CONSTITUTION: A mike signal(302) including a voice is inputted. A generation coefficient is estimated from the mike signal. A likelihood value is calculated from the estimated generation coefficient and a basis function. Judging information I is calculated from a likelihood value of a voice signal and a likelihood value of a noise signal. A voice beginning point judging unit(304) judges a voice beginning point from the judging information. When the voice begins, the mike signal is inputted to a voice recognizer(305). A generation coefficient is estimated from the voice signal to detect a voice ending point. A likelihood value is calculated from the estimated generation coefficient and the basis function. Judging information II is calculated from the likelihood value of the voice signal and the likelihood value of the noise signal. A voice ending point judging unit(306) judges a voice ending point.
독립 요소 해석 기법의 기저 함수를 이용한 음성 신호 검출 장치 및 그 방법이 개시되어 있다. 본 발명은 소정의 학습 법칙에 의해 음성 신호의 기저함수와 잡음 신호의 기저함수를 훈련하는 과정, 마이크로 입력되는 잡음 특성을 반영하여 상기 과정에서 훈련된 잡음신호의 기저함수를 현재 환경에 적응시키는 과정, 상기 훈련된 음성신호의 기저함수와 상기 환경에 적응된 잡음신호의 기저함수를 바탕으로 음성 구간의 판별정보를 추출하는 과정, 상기 과정에서 추출된 판별정보로 입력되는 마이크 신호의 음성 시작점과 음성 끝점을 검출하는 과정을 포함한다. |
---|