Method and apparatus for recognizing speech
AssigneeNanjing Horizon Robotics Technology Co., Ltd.
InventorsChangbao Zhu, Jianwei Niu, Ding Liu
Life of this patent
Estimating the term…
Abstract
Disclosed is a method and an apparatus for recognizing speech, and the method comprises: separating an input audio signal into at least two separated signals; generating a denoised signal at a current frame; performing a preliminary recognition on each interesting signal at the current frame; and performing a recognition decision according to a recognition score of each interesting signal at the current frame. The method and apparatus of the present disclosure deeply integrate an array signal processing and a speech recognition and use multiway recognitions such that a good recognition rate may be obtained even in a case of a low signal-to-noise ratio.
Claims17
A method for recognizing speech, comprising:
- separating an input audio signal into at least two separated signals;
- generating a denoised signal at a current frame based on a primary separated signal and one or more secondary separated signals selected from the at least two separated signals of the audio signal at the current frame;
- performing a preliminary recognition on each of a plurality of interesting signals at the current frame, the plurality of interesting signals at the current frame including the at least two separated signals and the denoised signal at the current frame, and each of the plurality of interesting signals at the current frame having a recognition score at the current frame associated with a result of the preliminary recognition at the current frame;
- and
- performing a recognition decision according to the recognition score of each of the plurality of interesting signals at the current frame.
An apparatus for recognizing speech, comprising:
- a splitter configured to separate an input audio signal into at least two separated signals;
- a signal generator configured to generate a denoised signal at a current frame based on a primary separated signal and one or more secondary separated signals selected from the at least two separated signals of the audio signal at the current frame;
- a recognizer configured to perform a preliminary recognition on each of a plurality of interesting signals at the current frame, the plurality of interesting signals at the current frame including the at least two separated signals and the denoised signal at the current frame, and each of the plurality of interesting signals at the current frame having a recognition score at the current frame associated with the result of preliminary recognition at the current frame;
- and
- a decision device configured to perform a recognition decision according to the recognition score at the current frame of each of a plurality of interesting signals at the current frame.
Dependencies are inferred from the claim text.