Speech recognition using word-in-phrase command

• US 7,085,716 B1
• Filed: 10/26/2000
• Issued: 08/01/2006
• Est. Priority Date: 10/26/2000
• Status: Expired due to Term

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Li, Li, Even, Stijn Van, Du, Xianju, Zhan, Puming
• Primary Examiner(s): Storm, Donald L.
• Application Number:
  US09/696,685
• Time in Patent Office: 2,105 Days
• Field of Search: 704/235, 704/251
• US Class Current: 704/235
• CPC Class Codes: G10L 15/1815 Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2015/228 of application context

Abstract

A method is described that corrects incorrect text associated with recognition errors in computer-implemented speech recognition. The method includes the step of performing speech recognition on an utterance to produce a recognition result for the utterance. The command includes a word and a phrase. The method includes determining if a word closely corresponds to a portion of the phrase. A speech recognition result is produced if the word closely corresponds to a portion of the phrase.

Claims

1. A method of computer-implemented speech recognition, the method comprising:

performing speech recognition on an utterance to produce a recognition result for the utterance, the recognition result including a command, a word, and a phrase;

determining if the word is similar to a portion of the phrase; and

producing a speech recognition result if the word is similar to a portion of the phrase.

2. The method of claim 1 wherein the recognition result comprises a command in the Chinese language.

3. The method of claim 1 wherein the recognition result comprises “

Write <

word>

as in <

phrase>

”

in the English language.

4. The method of claim 1 further comprising extracting the word and the phrase from the recognition result.

5. The method of claim 1 wherein determining if the word is similar to a portion of the phrase comprises determining if the word matches a substring of the phrase.

6. The method of claim 5 wherein producing the speech recognition result comprises producing the word.

7. The method of claim 1 wherein determining if the word is similar to a portion of the phrase comprises determining if the word sounds similar to a substring of the phrase.

8. The method of claim 7 wherein producing the speech recognition result comprises producing the substring of the phrase that sounds similar to the word.

9. The method of claim 1 further comprising producing no speech recognition result if the word is not similar to a portion of the phrase.

10. The method of claim 1 further comprising determining if previously-recognized text has been selected.

11. The method of claim 10 further comprising replacing selected text with the produced speech recognition result if text has been selected.

12. The method of claim 11 further comprising inserting the produced speech recognition result into the text at a predetermined location if text has not been selected.

13. A speech recognition system comprising:

an input device for receiving user input;

an output device for presenting information to the user;

a processor having communications links for transmitting information to and from the output and input devices; and

memory storing software instructions performed by the processor (i) for performing speech recognition on an utterance to produce a recognition result for the utterance, the recognition result including a command, a word, and a phrase, (ii) for determining if the word is similar to a portion of the phrase, and (iii) for producing a speech recognition result if the word is similar to a portion of the phrase.

14. The system of claim 13 wherein the recognition result comprises a command in the Chinese language.

15. The system of claim 13 wherein the recognition result comprises “

Write <

word>

as in <

phrase>

”

in the English language.

16. The system of claim 13 wherein the memory further comprises software instructions for extracting the word and the phrase from the recognition result.

17. The system of claim 13 wherein the software instruction for determining if the word is similar to a portion of the phrase comprises a software instruction for determining if the word matches a substring of the phrase.

18. The system of claim 17 wherein the software instruction for producing the speech recognition result comprises a software instruction for producing the word.

19. The system of claim 13 wherein the software instruction for determining if the word is similar to a portion of the phrase comprises a software instruction for determining if the word sounds similar to a substring of the phrase.

20. The system of claim 19 wherein the software instruction for producing the speech recognition result comprises a software instruction for producing the substring of the phrase that sounds similar to the word.

21. The system of claim 13 wherein the memory further comprises a software instruction for producing no speech recognition result if the word is not similar to a portion of the phrase.

22. The system of claim 13 wherein the memory further comprises a software instruction for determining if previously-recognized text has been selected.

23. The system of claim 22 wherein the memory further comprises a software instruction for replacing selected text with the produced speech recognition result if text has been selected.

24. The system of claim 23 wherein the memory further comprises a software instruction for inserting the produced speech recognition result into the text at a predetermined location if text has not been selected.

25. A computer-readable medium, comprising software instructions for speech recognition, the software instructions comprising:

a first code segment to perform speech recognition on an utterance to produce a recognition result for the utterance, the recognition result including a command, a word, and a phrase;

a second code segment to determine if the word is similar to a portion of the phrase; and

a third code segment to produce a speech recognition result if the word is similar to a portion of the phrase.

26. The computer readable medium of claim 25 wherein the recognition result comprises a command in the Chinese language.

27. The computer readable medium of claim 25 wherein the recognition result comprises “

Write <

word>

as in <

phrase>

”

in the English language.

28. The computer readable medium of claim 25 wherein the software instructions comprise a fourth code segment to extract the word and the phrase from the recognition result.

29. The computer readable medium of claim 25 wherein the second code segment comprises a code segment to determine if the word matches a substring of the phrase.

30. The computer readable medium of claim 29 wherein the third code segment comprises a code segment to produce the word.

31. The computer readable medium of claim 25 wherein the second code segment comprises a code segment to determine if the word sounds similar to a substring of the phrase.

32. The computer readable medium of claim 31 wherein the third code segment comprises a code segment to produce the substring of the phrase that sounds similar to the word.

33. The computer readable medium of claim 25 wherein the software instructions comprise a fourth code segment to produce no speech recognition result if the word is not similar to a portion of the phrase.

34. The computer readable medium of claim 25 wherein the software instructions comprise a fourth code segment to determine if previously-recognized text has been selected.

35. The computer readable medium of claim 34 wherein the software instructions comprise a fifth code segment to replace selected text with the produced speech recognition result if text has been selected.

36. The computer readable medium of claim 35 wherein the software instructions comprise a sixth code segment to insert the produced speech recognition result into the text at a predetermined location if text has not been selected.