Method and apparatus for recognizing speech, and method and apparatus for generating noise-speech recognition model

Patent metadata

Field	Value
Publication number	US-9626962-B2
Application number	US-201514621050-A
Country	US
Kind code	B2
Filing date	Feb 12, 2015
Priority date	May 2, 2014
Publication date	Apr 18, 2017
Grant date	Apr 18, 2017

Abstract

An apparatus and method for recognizing a speech, and an apparatus and method for generating a noise-speech recognition model are provided. The speech recognition apparatus includes a location determiner configured to determine a location of the apparatus, a noise model generator configured to generate a noise model corresponding to the location by collecting noise data related to the location, and a noise model transmitter configured to transmit the noise model to a server.

First claim

What is claimed is:

1. A speech recognition apparatus, comprising:
a location determiner configured to determine a location of the apparatus;
a noise model generator configured to generate a noise model corresponding to the location by collecting noise data related to the location;
a noise model transmitter configured to transmit the noise model to a server, and
a speech recognizer configured to perform speech recognition using a noise-speech recognition model that is generated by the server by applying the noise model to a baseline speech recognition model.

2. The speech recognition apparatus of claim 1 , wherein the noise model generator is configured to collect the noise data that is generated at the location via a microphone of the apparatus.

3. The speech recognition apparatus of claim 1 , wherein the noise model generator is configured to collect the noise data from web videos related to the location.

4. The speech recognition apparatus of claim 1 , wherein the speech recognizer is further configured to transmit a speech recognition request including the location and a speech signal to the server, and receive a result of speech recognition that is performed on the speech signal using the noise-speech recognition model.

5. The speech recognition apparatus of claim 1 , wherein the speech recognizer is configured to receive from the server the noise-speech recognition model applied with the noise model corresponding to the location, and to perform speech recognition using the noise-speech recognition model.

6. An apparatus for generating a noise-speech recognition model, the apparatus comprising:
a noise model receiver configured to receive, from a mobile terminal, a noise model corresponding to a location of the mobile terminal;
a noise-speech recognition model generator configured to generate a noise-speech recognition model corresponding to the location, by applying the noise model to a baseline speech recognition model; and
a memory storage configured to store the noise-speech recognition model.

7. The apparatus of claim 6 , further comprising:
a speech recognizer configured to, in response to receiving a speech recognition request including information on the location of the mobile terminal and speech signal from the mobile terminal, perform speech recognition on the speech signal using the noise-speech recognition model corresponding to the location; and
a speech recognition result transmitter configured to transmit a result of the speech recognition to the mobile terminal.

8. The apparatus of claim 6 , wherein the noise model comprises noise data related to the location of the mobile terminal and information on the location of the mobile terminal.

9. The apparatus of claim 6 , further comprising:
a noise-speech recognition model transmitter configured to, in response to receiving a noise-speech recognition model transmission request including the location of the mobile terminal, transmit a noise-speech recognition model corresponding to the location to the mobile terminal.

10. A method of speech recognition, the method comprising:
determining a location of a mobile terminal;
collecting noise data related to the location;
generating noise model corresponding to the location using the noise data;
transmitting the noise model to a server,
receiving a speech signal; and
performing speech recognition on the speech signal using the noise-speech recognition model that is generated by the server by applying the noise model to a baseline speech recognition model.

11. The method of claim 10 , wherein the performing of speech recognition comprises:
transmitting a speech recognition request including the location of the mobile terminal and the speech signal to the server; and
receiving a result of speech recognition performed on the speech signal using the noise-speech recognition model to which the noise model corresponding to the location of the mobile terminal at time of receipt of the speech signal is applied.

12. The method of claim 10 , wherein the collecting of noise data comprises collecting noise data generated at the location.

13. The speech recognition method of claim 10 , wherein the collecting of noise data comprises collecting noise data from a web video found by searching web videos related to the location.

14. The method of claim 10 , wherein the performing of speech recognition comprises:
determining a location of the mobile terminal at time of receiving the speech signal; and
receiving a noise-speech recognition model to which the noise model corresponding to the location of the mobile terminal using the received noise-speech recognition model.

15. A method of generating a noise-speech model, the method comprising:
receiving a noise model corresponding to a location;
generating a noise-speech recognition model corresponding to the location by applying the noise model to a baseline speech recognition model; and
storing the noise-speech recognition model.

16. The method of claim 15 , further comprising:
receiving, from a mobile terminal, a speech recognition request including the information on the location of the mobile terminal and a speech signal;
performing speech recognition on the speech signal using a noise-speech recognition model corresponding to the location of the mobile terminal; and
transmitting a result of the speech recognition to the mobile terminal.

17. The method of claim 15 , further comprising:
receiving, from the mobile terminal, a noise-speech recognition model transmission request including the information on the location of the mobile terminal; and
transmitting, to the mobile terminal, a noise-speech recognition model corresponding to the location of the mobile terminal.

18. The method of claim 15 , wherein the noise model comprises noise data related to the location and information on the location.

19. An apparatus, comprising:
a microphone configured to detect a speech signal; and
a processor configured to obtain a noise-speech recognition model corresponding to a location of the apparatus and to recognize a word in the speech signal using the noise-speech recognition model.

20. The apparatus of claim 19 , wherein the processor is configured to determine the location of the apparatus and generate a noise model.

21. The apparatus of claim 20 , wherein the apparatus further comprises a noise model transmitter configured to transmit the noise model to a server, and the processor is configured to obtain the noise-speech recognition model from the server.

22. The apparatus of claim 19 , wherein the processor is configured to collect noise data at the location from a sound signal detected by the microphone.

Assignees

• Samsung Electronics Co Ltd

Inventors

• Lee Ho Sub
• Seo Young Wan
• Choi Young Sang

Classifications

• G10L15/30

  Distributed recognition, e.g. in client-server systems, for mobile phones or network applications · CPC title

• G10L21/0208

  Noise filtering · CPC title

• G10L15/20 Primary

  Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech (G10L21/02 takes precedence) · CPC title

• G10L2015/228

  of application context · CPC title