Selecting an acoustic model in a speech recognition system

• US 6,952,674 B2
• Filed: 01/07/2002
• Issued: 10/04/2005
• Est. Priority Date: 01/07/2002
• Status: Expired due to Fees

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Forand, Richard A.
• Primary Examiner(s): Lerner, Martin
• Application Number:
  US10/041,754
• Publication Number:
  US 20030130840A1
• Time in Patent Office: 1,366 Days
• Field of Search: 704/275, 704/243, 704/244, 704/246, 704/247, 704/250, 704/261, 704/270, 704/270.1, 379/88.01, 379/88.02
• US Class Current: 704/243
• CPC Class Codes: G10L 15/07 to the speaker

Abstract

A speech-recognition system learns a speech profile of a user whose speech is to be recognized. The system plays audible speech samples, stored in sound files, so that the file that most resembles the user'"'"'s speech may be selected. After receiving the selection, the system identifies an acoustic model that is associated with the chosen sound file. The system may also select a subset of sound files based on information indicative of the user'"'"'s speech. The system may then play a subset of sound files so that the file that most resembles the user'"'"'s speech may be selected.

Claims

1. A method for a speech-recognition system to select an acoustic model for recognizing a user'"'"'s speech, the method comprising:

making a plurality of speech samples available for listening so that the speech sample most resembling the user'"'"'s speech may be selected;

receiving the selection; and

identifying an acoustic model associated with the selected speech sample for recognizing the user'"'"'s speech.

2. The method of claim 1, further comprising:

asking a question prior to making available the plurality of speech samples, an answer to the question being indicative of the user'"'"'s speech; and

using the answer to select the plurality of speech samples to be made available for listening.

3. The method of claim 2, wherein the question relates to at least one from the group consisting of the user'"'"'s gender, nationality, geographic origin and age.

4. The method of claim 1, further comprising:

obtaining information indicative of the user'"'"'s speech and using the information to select the plurality of speech samples to be made available for listening.

5. The method of claim 4, wherein obtaining the information comprises:

making speech samples available for listening; and

receiving the selection of which speech sample most resembles the user'"'"'s speech.

6. The method of claim 4, wherein obtaining the information comprises registering the user'"'"'s speech.

7. The method of claim 1, wherein at least one of the speech samples comprises synthetic speech.

8. The method of claim 1, wherein the speech-recognition system resides on a personal computer.

9. The method of claim 1, wherein the speech-recognition system resides on a networked computer system.

10. The method of claim 1, wherein the plurality of speech samples is made available via telephone.

11. Computer-readable medium having stored thereon instructions that when executed perform the following functions:

makes a plurality of speech samples available for listening so that the speech sample most resembling a user'"'"'s speech may be selected; and

identifies an acoustic model associated with the selected speech sample so that the acoustic model may thereafter be used in recognizing the user'"'"'s speech.

12. The computer-readable medium of claim 11, further comprising the functions:

asks a question prior to making available the plurality of speech samples, an answer to the question being indicative of the user'"'"'s speech; and

uses the answer to select the plurality of speech samples.

13. The computer-readable medium of claim 12, wherein the question relates to at least one from the group consisting of the user'"'"'s gender, nationality, geographic origin and age.

14. The computer-readable medium of claim 11, further comprising the functions:

obtains information indicative of the user'"'"'s speech; and

uses the information to select the plurality of speech samples.

15. The computer-readable medium of claim 14, wherein obtaining the information comprises the following functions:

makes speech samples available for listening; and

receives the selection of which speech sample most resembles the user'"'"'s speech.

16. The computer-readable medium of claim 14, wherein obtaining the information comprises the following function:

registers the user'"'"'s speech.

17. A method for a speech-recognition system to select an acoustic model for recognizing a user'"'"'s speech, the method comprising:

obtaining information indicative of the user'"'"'s speech and using the information to select a plurality of speech samples;

making the plurality of speech samples available for listening so that the speech sample most resembling the user'"'"'s speech may be selected;

receiving the selection; and

identifying an acoustic model associated with the selected speech sample for recognizing the user'"'"'s speech.

18. The method of claim 17, wherein obtaining the information comprises asking a question and receiving an answer to the question.

19. The method of claim 17, wherein obtaining the information comprises:

making speech samples available for listening so that the speech sample most resembling the user'"'"'s speech may be selected; and

receiving the selection.

20. The method of claim 17, wherein obtaining the information comprises registering the user'"'"'s speech.

21. A speech-recognition system comprising:

a plurality of acoustic models for recognizing speech;

a setup module that makes at least some of a plurality of speech samples available for listening so that the speech sample most resembling a user'"'"'s speech may be selected; and

a speech-recognition module that uses an acoustic model associated with the selected speech sample to recognize the user'"'"'s speech.

22. The speech-recognition system of claim 21, wherein the setup module obtains information indicative of the user'"'"'s speech and uses the information to select the speech samples that are to be made available for listening.

23. The speech-recognition system of claim 22, wherein the setup module asks a question and receives an answer.

24. The speech-recognition system of claim 22, wherein the setup module makes speech samples available for listening so that a speech sample most resembling a user'"'"'s speech may be selected, receives the selection, and uses the selection to determine which speech samples of the plurality of speech samples that are to be made available.

25. The speech-recognition system of claim 22, wherein the setup module registers the user'"'"'s voice.

26. The speech-recognition system of claim 21, residing on a personal computer.

27. The speech-recognition system of claim 21, residing on a networked computer system.

28. The speech-recognition system of claim 21, wherein the setup module makes speech samples available for listening via telephone.