METHOD AND SYSTEMS HAVING IMPROVED SPEECH RECOGNITION

• US 20130332165A1
• Filed: 06/06/2012
• Published: 12/12/2013
• Est. Priority Date: 06/06/2012
• Status: Active Grant

• Current Assignee: Qualcomm, Inc.
• Original Assignee: Qualcomm, Inc.
• Inventors: BALASUBRAMANYAM, Shivakumar, BECKLEY, Jeffrey D., AGGARWAL, Pooja
• Granted Patent: US 9,881,616 B2
• Time in Patent Office: Days
• US Class Current: 704/246
• CPC Class Codes: G10L 15/20 Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech G10L21/02 takes precedence
• CPC Class Codes: G10L 17/00 Speaker identification or verification techniques
• CPC Class Codes: G10L 17/04 Training, enrolment or model building
• CPC Class Codes: G10L 2021/02166 Microphone arrays; Beamforming

Abstract

A method for improving speech recognition by a speech recognition system includes obtaining a voice sample from a speaker; storing the voice sample of the speaker as a voice model in a voice model database; identifying an area from which sound matching the voice model for the speaker is coming; providing one or more audio signals corresponding to sound received from the identified area to the speech recognition system for processing.

Claims

1. A method for improving speech recognition by a speech recognition system, comprising:

obtaining a voice sample from a speaker;

storing the voice sample of the speaker as a voice model in a voice model database;

identifying an area from which sound matching the voice model for the speaker is coming; and

providing one or more audio signals corresponding to sound received from the identified area to the speech recognition system for processing.

2. The method of claim 1, further comprising:

controlling microphones to receive sound from the identified area.

3. The method of claim 2, further comprising:

controlling the microphones to not receive sound from areas other than the identified area.

4. The method of claim 1, further comprising:

filtering sound received from areas other than the identified area.

5. The method of claim 1, further comprising:

separating ambient noise from speech in the one or more audio signals based on the voice model stored in the voice model database to obtain a resulting speech signal; and

providing the resulting speech signal to the speech recognition system for processing.

6. The method of claim 1,

wherein the identifying comprises:

receiving, from one or more microphones, one or more audio signals corresponding to sound coming from a first area of a plurality of areas; and

determining whether the received audio signals corresponding to the sound coming from the first area match the voice model for the speaker;

wherein the providing comprises;

selectively providing audio signals corresponding to sound from the first area to the speech recognition system for processing thereof based on the determination.

7. The method of claim 6,

wherein if the one or more received audio signals corresponding to the sound coming from the first area match the voice model for the speaker, one or more audio signals corresponding to sound from the first area are provided to the speech recognition system for processing thereof;

and

wherein if the one or more received audio signals corresponding to the sound coming from the first area do not match the voice model for the speaker, one or more audio signals corresponding to sound from the first area are not provided to the speech recognition system for processing thereof.

8. The method of claim 6, wherein one or more audio signals corresponding to sound from the first area received after the determination are selectively provided to the speech recognition system for processing thereof based on the determination.

9. The method of claim 6, wherein the one or more microphones comprises at least two microphones.

10. The method of claim 6, further comprising:

receiving, from the one or more microphones, one or more audio signals corresponding to sound coming from a second area of the plurality of areas;

determining whether the one or more received audio signals corresponding to the sound coming from the second area match the voice model for the speaker; and

selectively providing audio signals corresponding to sound from the second area to the speech recognition system for processing thereof based on the determination.

11. The method of claim 10,

wherein if the one or more audio signals corresponding to the sound coming from the second area match the voice model for the speaker, one or more audio signals corresponding to sound from the second area are provided to the speech recognition system for processing thereof;

and

wherein if the one or more audio signals corresponding to the sound coming from the second area do not match the voice model for the speaker, one or more audio signals corresponding to sound from the second area are not provided to the speech recognition system for processing thereof.

12. The method of claim 6, wherein the determining comprises:

characterizing the sound coming from the first area; and

comparing the characterized sound to the voice model for the speaker.

13. The method of claim 12,

wherein if the characterized sound compares with the voice model for the speaker, the sound coming from the first area is determined to be speech from the speaker and one or more audio signals corresponding to sound coming from the first area is provided to the speech recognition system;

and

wherein if the characterized sound does not compare with the voice model for the speaker, the sound coming from the first area is determined not to be speech from the speaker and one or more audio signals corresponding to sound coming from the first area is not provided to the speech recognition system.

14. The method of claim 12, wherein the sound coming from the first area is characterized through analysis of an acoustic spectrum of the sound.

15. The method of claim 12,

wherein the one or more microphones comprise a first microphone associated with a first area of the plurality of areas and a second microphone associated a second area of the plurality of areas;

wherein if the one or more received audio signals corresponding to the sound coming from the first area matches the voice model for the speaker, one or more audio signals corresponding to sound from the first area are received from the first microphone for providing to the speech recognition system; and

wherein if the one or more received audio signals corresponding to the sound coming from the first area do not matches the voice model for the speaker, one or more audio signals corresponding to sound from the first area are not received from the first microphone for providing to the speech recognition system.

16. The method of claim 15,

wherein if the one or more received audio signals corresponding to the sound coming from the first area matches the voice model for the speaker and the one or more received audio signals corresponding to the sound coming from the second area does not match the voice model for the speaker, one or more audio signals corresponding to sound from the first area are received from the first microphone for providing to the speech recognition system;

and

wherein if the one or more received audio signals corresponding to the sound coming from the first area does not match the voice model for the speaker and the one or more received audio signals corresponding to the sound coming from the second area matches the voice model for the speaker, one or more audio signals corresponding to sound from the second area are received from the second microphone.

17. The method of claim 1,

wherein the audio signals provided to the speech recognition system correspond to speech commands for the speech recognition system;

and

wherein the speech recognition system issues instructions based on the speech.

18. An apparatus for improving speech recognition by a speech recognition system, the apparatus comprising:

means for obtaining a voice sample from a speaker;

means for storing the voice sample of the speaker as a voice model in a voice model database;

means for identifying an area from which sound matching the voice model for the speaker is coming; and

means for providing one or more audio signals corresponding to sound received from the identified area to the speech recognition system for processing.

19. An apparatus for improving speech recognition by a speech recognition system, the apparatus comprising:

a processor configured for;

obtaining a voice sample from a speaker;

storing the voice sample of the speaker as a voice model in a voice model database;

receiving, from one or more microphones, one or more audio signals corresponding to sound coming from a first area of a plurality of areas;

identifying an area from which sound matching the voice model for the speaker is coming; and

providing one or more audio signals corresponding to sound received from the identified area to the speech recognition system for processing.

20. A computer program product for improving speech recognition by a speech recognition system, the computer program product comprising:

a computer-readable storage medium comprising code for;

obtaining a voice sample from a speaker;

storing the voice sample of the speaker as a voice model in a voice model database;

receiving, from one or more microphones, one or more audio signals corresponding to sound coming from a first area of a plurality of areas;

identifying an area from which sound matching the voice model for the speaker is coming; and

providing one or more audio signals corresponding to sound received from the identified area to the speech recognition system for processing.