USER FRIENDLY SPEAKER ADAPTATION FOR SPEECH RECOGNITION

• US 20100088097A1
• Filed: 10/03/2008
• Published: 04/08/2010
• Est. Priority Date: 10/03/2008
• Status: Abandoned Application

• Current Assignee: Nokia Corporation
• Original Assignee: Nokia Corporation
• Inventors: Marila, Juha, Leppanen, Jussi, Vainio, Janne, Mikkola, Hannu, Tian, Jilei
• Application Number:
  US12/244,919
• Publication Number:
  US 20100088097A1
• Time in Patent Office: Days
• US Class Current: 704/251
• CPC Class Codes: G10L 15/07 to the speaker
• CPC Class Codes: G10L 2015/0631 Creating reference templates; Clustering

Abstract

Improved performance and user experience for speech recognition application and system by utilizing for example offline adaptation without tedious effort by a user. Interactions with a user may be in the form of a quiz, game, or other scenario wherein the user may implicitly provide vocal input for adaptation data. Queries with a plurality of candidate answers may be designed in an optimal and efficient way, and presented to the user, wherein detected speech from the user is then matched to one of the candidate answers, and may be used to adapt an acoustic model to the particular speaker for speech recognition.

Claims

1. A method comprising:

presenting a query to a user;

presenting to the user a plurality of possible answers to the query;

receiving a vocal response from the user;

matching the vocal response to one of the plurality of possible answers presented to the user; and

using the matched vocal response to adapt an acoustic model for the user for a speech recognition application.

2. The method of claim 1 further including selecting the query based on phonetic content of the possible answers.

3. The method of claim 1 further including selecting the query based on an interactive game for the user.

4. The method of claim 1 wherein matching the vocal response includes performing a forced alignment between the vocal response and one of the plurality of possible answers to the query.

5. The method of claim 1 wherein matching the vocal response includes selecting a potential match, and receiving a confirmation from the user that the potential match is correct.

6. The method of claim 1 wherein the plurality of possible answers to the query are phonetically balanced.

7. The method of claim 1 wherein the plurality of possible answers to the query are substantially phonetically distinguishable.

8. The method of claim 1 wherein the plurality of possible answers are created to minimize an objective function value among the plurality of possible answers.

9. The method of claim 1 wherein the process of matching the vocal response to one of the plurality of possible answers includes determining if one of the plurality of possible answers exceeds an adaptation threshold.

10. The method of claim 1 wherein the process of presenting a query, presenting a plurality of possible answers, receiving a vocal response, and matching the vocal response, is repeated multiple times.

11. The method of claim 4 wherein a forced alignment likelihood ratio (R) between the vocal response (S) and a first possible answer W_ans1 and a second possible answer W_ans2 is calculated using:

$R<br><br>\left(W_{\mathrm{ans}<br><br><br><br>1},W_{\mathrm{ans}<br><br><br><br>2},S\right)=\frac{P<br><br>\left(W_{\mathrm{ans}<br><br><br><br>1}<br><br>S\right)}{P<br><br>\left(W_{\mathrm{ans}<br><br><br><br>2}<br><br>S\right)}=\frac{P<br><br>\left(S<br><br>W_{\mathrm{ans}<br><br><br><br>1}\right)·<br><br>P<br><br>\left(W_{\mathrm{ans}<br><br><br><br>1}\right)}{P<br><br>\left(S<br><br>W_{\mathrm{ans}<br><br><br><br>2}\right)·<br><br>P<br><br>\left(W_{\mathrm{ans}<br><br><br><br>2}\right)}.$

12. The method of claim 1 wherein the process of using the matched vocal response to adapt an acoustic model includes using the matched vocal response only if the matched vocal response exceeds a predetermined threshold value.

13. The method of claim 12 wherein adjusting the predetermined threshold value adjusts a quality of the matched vocal responses used to adapt the acoustic model.

14. An apparatus comprising:

a processor; and

a memory, including machine executable instructions, that when provided to the processor, cause the processor to perform;

presenting a query to a user;

presenting to the user a plurality of possible answers to the query;

receiving a vocal response from the user;

matching the vocal response to one of the plurality of possible answers presented to the user; and

using the matched vocal response to adapt an acoustic model for the user for a speech recognition application.

15. The apparatus of claim 14 further including instructions for the processor to perform selecting the query based on phonetic content of the possible answers.

16. The apparatus of claim 14 further including instructions for the processor to perform selecting the query based on an interactive game for the user.

17. The apparatus of claim 14 wherein matching the vocal response includes performing a forced alignment between the vocal response and one of the plurality of possible answers to the query.

18. The apparatus of claim 14 wherein matching the vocal response includes selecting a potential match, and receiving a confirmation from the user that the potential match is correct.

19. The apparatus of claim 14 wherein the plurality of possible answers to the query are phonetically balanced.

20. The apparatus of claim 14 wherein the plurality of possible answers to the query are substantially phonetically distinguishable.

21. The apparatus of claim 14 wherein the process of matching the vocal response to one of the plurality of possible answers includes determining if one of the plurality of possible answers exceeds an adaptation threshold.

22. The apparatus of claim 14 wherein the apparatus includes a mobile terminal.

23. A computer readable medium including instructions that when provided to a processor cause the processor to perform:

presenting a query to a user;

presenting to the user a plurality of possible answers to the query;

receiving a vocal response from the user;

matching the vocal response to one of the plurality of possible answers presented to the user; and

using the matched vocal response to adapt an acoustic model for the user for a speech recognition application.

24. The computer readable medium of claim 23 further including instructions for the processor to perform selecting the query based on phonetic content of the possible answers.

25. The computer readable medium of claim 23 further including instructions for the processor to perform selecting the query based on an interactive game for the user.

26. The computer readable medium of claim 23 including instructions wherein matching the vocal response to one of the plurality of possible answers includes determining if one of the plurality of possible answers exceeds an adaptation threshold.

27. An apparatus comprising:

means for presenting a query to a user;

means for presenting to the user a plurality of possible answers;

means for receiving a vocal response from the user;

matching means for matching a vocal response received from the user to one of the plurality of possible answers presented to the user; and

means for adapting an acoustic model for the user for a speech recognition application based on the matched vocal response.

28. The apparatus of claim 27 wherein the matching means includes means for performing a forced alignment between the vocal response and one of the plurality of possible answers.