User-cued speech recognition

• US 6,195,635 B1
• Filed: 08/13/1998
• Issued: 02/27/2001
• Est. Priority Date: 08/13/1998
• Status: Expired due to Term

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: Dragon Systems, Inc. (Microsoft Corporation)
• Inventors: Wright, Barton D.
• Primary Examiner(s): Knepper, David D.
• Application Number:
  US09/130,342
• Time in Patent Office: 929 Days
• Field of Search: 704/270, 704/275, 704/235, 704/231
• US Class Current: 704/231
• CPC Class Codes: G10L 15/08 Speech classification or search

Abstract

Recognition of speech by a speech recognizer is improved by receiving deliberately contiguously repeated spoken utterances corresponding to a speech element and recognizing fewer instances of the speech element than the number of repeated spoken utterances. If a spoken utterance corresponding to the speech element is received and misrecognized prior to receiving the deliberately contiguously repeated spoken utterances, the spoken utterance and the repeated spoken utterances may be used to recognize the speech element.

Claims

1. A method for improving recognition of a speech element by a speech recognizer, comprising:

receiving deliberately contiguously repeated spoken instances of the speech element;

performing speech recognition on the spoken instances of the speech element; and

producing a speech recognition result that includes only a single instance of the speech element.

2. The method of claim 1, wherein performing speech recognition on the spoken instances of the speech element comprises identifying possible recognized speech elements for the repeated spoken instances of the speech element;

and

producing the speech recognition result comprises selecting one of the possible recognized speech elements as a recognized speech element.

3. The method of claim 2, wherein the selecting one of the possible recognized speech elements comprises:

developing scores for the possible recognized speech elements; and

selecting as the recognized speech element a possible recognized speech element with an optimal score.

4. The method of claim 3, wherein possible recognized speech elements are identified for a predetermined number of the repeated spoken instances of the speech element.

5. The method of claim 1, wherein performing speech recognition on the spoken instances of the speech element comprises applying a recognition process directly to representations of speech waveforms for at least two of the repeated spoken instances of the speech element without separately recognizing a speech element for each of the spoken instances.

6. The method of claim 1, wherein the speech element comprises a word.

7. The method of claim 1, wherein the speech element comprises a phrase.

8. The method of claim 1, wherein the speech element comprises a sentence.

9. The method of claim 1, wherein:

at least one of the repeated spoken instances of the speech element is repeated by a user after misrecognition of another one of the repeated spoken instances of the speech element is apparent.

10. The method of claim 1, further comprising:

if the speech element is in a predetermined class of speech elements, recognizing an instance of the speech element for each of the repeated spoken instances.

11. The method of claim 10, wherein the class comprises speech elements which may properly be repeated in a language recognized by the speech recognizer.

12. The method of claim 10, wherein the class comprises commands.

13. The method of claim 1, further comprising:

prior to receiving the deliberately contiguously repeated spoken instances of the speech element, receiving a spoken instance corresponding to the speech element; and

misrecognizing the speech element.

14. The method of claim 13, wherein the spoken instance of the speech element and the repeated spoken instances are used to recognize the speech element.

15. A computer program tangibly stored on a computer-readable medium and operable to cause a computer to improve recognition of a speech element by a speech recognizer, comprising instructions that cause the computer to:

receive deliberately contiguously repeated spoken instances of the speech element;

perform speech recognition on the spoken instances of the speech element; and

produce a speech recognition result that includes only a single instance of the speech element.

16. The computer program of claim 15, wherein instructions to perform speech recognition on the spoken instances of the speech element comprise:

instructions to identify possible recognized speech elements for the repeated spoken instances of the speech element; and

instructions to produce the speech recognition result comprises instructions to select one of the possible recognized speech elements as a recognized speech element.

17. The computer program of claim 16, wherein instructions to select comprise instructions to:

develop scores for the possible recognized speech elements; and

select as the recognized speech element a possible recognized speech element with an optimal score.

18. The computer program of claim 17, wherein possible recognized speech elements are identified for a predetermined number of the repeated spoken instances.

19. The computer program of claim 15, wherein instructions to perform speech recognition on the spoken instances of the speech element comprise instructions to apply a recognition process directly to representations of speech waveforms for at least two of the repeated spoken instances of the speech element without separately recognizing a speech element for each of the spoken instances.

20. The computer program of claim 15, wherein the speech element comprises a word.

21. The computer program of claim 15, wherein the speech element comprises a phrase.

22. The computer program of claim 15, wherein the speech element comprises a sentence.

23. The computer program of claim 15, wherein:

at least one of the repeated spoken instances of the speech element is repeated by a user after misrecognition of another one of the repeated spoken instances of the speech element is apparent.

24. The computer program of claim 15, further comprising instructions to:

recognize an instance of the speech element for each of the repeated spoken instances if the speech element is in a predetermined class of speech elements.

25. The computer program of claim 24, wherein the class comprises speech elements which may properly be repeated in a language recognized by the speech recognizer.

26. The computer program of claim 24, wherein the class comprises commands.

27. The computer program of claim 15, further comprising instructions to:

receive a spoken instance corresponding to the speech element prior to receiving the deliberately contiguously repeated spoken instances of the speech element; and

use the spoken instance and the repeated spoken instances to recognize the speech element.