Word boundary acoustic units

• US 6,606,594 B1
• Filed: 09/29/1999
• Issued: 08/12/2003
• Est. Priority Date: 09/29/1998
• Status: Expired due to Term

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: ScanSoft, Inc. n/k/a Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Sarukkai, Ramesh, Lynch, Tom, Sejnoha, Vladimir
• Primary Examiner(s): ABEBE, DANIEL DEMELASH
• Application Number:
  US09/408,388
• Time in Patent Office: 1,413 Days
• Field of Search: 704/231, 704/232, 704/242, 704/243, 704/245, 704/256, 704/257
• US Class Current: 704/250
• CPC Class Codes: G10L 15/05 Word boundary detection
• CPC Class Codes: G10L 15/187 Phonemic context, e.g. pronunciation rules, phonotactical constraints or phoneme n-grams
• CPC Class Codes: G10L 2015/022 Demisyllables, biphones or triphones being the recognition units

Abstract

A speech recognition system recognizes an input utterance of spoken words. The system includes a set of word models for modeling vocabulary to be recognized, each word model being associated with a word in the vocabulary, each word in the vocabulary considered as a sequence of phones including a first phone and a last phone, wherein each word model begins in the middle of the first phone of its associated word and ends in the middle of the last phone of its associated word; a set of word connecting models for modeling acoustic transitions between the middle of a word'"'"'s last phone and the middle of an immediately succeeding word'"'"'s first phone; and a recognition engine for processing the input utterance in relation to the set of word models and the set of word connecting models to cause recognition of the input utterance.

Claims

1. A speech recognition system for recognizing an input utterance of spoken words, the system comprising:

a set of word models for modeling vocabulary to be recognized, each word model being associated with a word in the vocabulary, each word in the vocabulary considered as a sequence of phones including a first phone and a last phone, wherein each word model begins in the middle of the first phone of its associated word and ends in the middle of the last phone of its associated word;

a set of word connecting models for modeling acoustic transitions between the middle of a word'"'"'s last phone and the middle of an immediately succeeding word'"'"'s first phone; and

a recognition engine for processing the input utterance in relation to the set of word models and the set of word connecting models to cause recognition of the input utterance.

2. A system as in claim 1, wherein each word model uses context-dependent phone models to represent the sequence of phones.

3. A system as in claim 2, wherein the context-dependent phone models are triphones.

4. A system as in claim 1, wherein the acoustic transitions include a pause.

5. A system as in claim 1, wherein the acoustic transitions include a period of silence.

6. A system as in claim 1, wherein the acoustic transitions include a period of noise.

7. A system as in claim 1, wherein each word connecting model further includes a previous word identification field which represents the word associated with the word model immediately preceding the word connecting model.

8. A system as in claim 1, wherein each word connecting model further includes an ending score field which represents a best score from the beginning of the input utterance to reach the word connecting model.

9. A system as in claim 1, wherein each word connecting model further includes a type field which represents specific details of the word connecting model.

10. A method of a speech recognition system for recognizing an input utterance of spoken words, the method comprising:

modeling vocabulary to be recognized with a set of word models, each word model being associated with a word in the vocabulary, each word in the vocabulary being considered as a sequence of phones including a first phone and a last phone, wherein each word model begins in the middle of the first phone of its associated word and ends in the middle of the last phone of its associated word;

modeling acoustic transitions between the middle of a word'"'"'s last phone and the middle of an immediately succeeding word'"'"'s first phone with a set of word connecting models; and

processing with a recognition engine the input utterance in relation to the set of word models and the set of word connecting models to cause recognition of the input utterance.

11. A method as in claim 10, wherein each word model uses context-dependent phone models to represent the sequence of phones.

12. A method as in claim 11, wherein the context-dependent phone models are triphones.

13. A method as in claim 10, wherein the acoustic transitions include a pause.

14. A method as in claim 10, wherein the acoustic transitions include a period of silence.

15. A method as in claim 10, wherein the acoustic transitions include a period of noise.

16. A method as in claim 10, wherein each word connecting model further includes a previous word identification field which represents the word associated with the word model immediately preceding the word connecting model.

17. A method as in claim 10, wherein each word connecting model further includes an ending score field which represents a best score from the beginning of the input utterance to reach the word connecting model.

18. A method as in claim 10, wherein each word connecting model further includes a type field which represents specific details of the word connecting model.

19. An improved speech recognition system of the type employing word models, wherein the improvement comprises:

a set of word models for modeling vocabulary to be recognized, each word model being associated with a word in the vocabulary, each word in the vocabulary considered as a sequence of phones including a first phone and a last phone, wherein each word model begins in the middle of the first phone of its associated word and ends in the middle of the last phone of its associated word; and

a set of word connecting models for modeling acoustic transitions between the middle of a word'"'"'s last phone and the middle of an immediately succeeding word'"'"'s first phone.

20. A system as in claim 19, wherein each word model uses context-dependent phone models to represent the sequence of phones.

21. A system as in claim 20, wherein the context-dependent phone models are triphones.

22. A system as in claim 19, wherein the acoustic transitions include a pause.

23. A system as in claim 19, wherein the acoustic transitions include a period of silence.

24. A system as in claim 19, wherein the acoustic transitions include a period of noise.

25. A system as in claim 19, wherein each word connecting model further includes a previous word identification field which represents the word associated with the word model immediately preceding the word connecting model.

26. A system as in claim 19, wherein each word connecting model further includes an ending score field which represents a best score from the beginning of the input utterance to reach the word connecting model.

27. A system as in claim 19, wherein each word connecting model further includes a type field which represents specific details of the word connecting model.

28. An improved method of a speech recognition system for recognizing an input utterance of spoken words, the improvement comprising:

modeling vocabulary to be recognized with a set of word models, each word model being associated with a word in the vocabulary, each word in the vocabulary being considered as a sequence of phones including a first phone and a last phone, wherein each word model begins in the middle of the first phone of its associated word and ends in the middle of the last phone of its associated word; and

modeling acoustic transitions between the middle of a word'"'"'s last phone and the middle of an immediately succeeding word'"'"'s first phone with a set of word connecting models.

29. A method as in claim 28, wherein each word model uses context-dependent phone models to represent the sequence of phones.

30. A method as in claim 29, wherein the context-dependent phone models are triphones.

31. A method as in claim 28, wherein the acoustic transitions include a pause.

32. A method as in claim 28, wherein the acoustic transitions include a period of silence.

33. A method as in claim 28, wherein the acoustic transitions include a period of noise.

34. A method as in claim 28, wherein each word connecting model further includes a previous word identification field which represents the word associated with the word model immediately preceding the word connecting model.

35. A method as in claim 28, wherein each word connecting model further includes an ending score field which represents a best score from the beginning of the input utterance to reach the word connecting model.

36. A method as in claim 28, wherein each word connecting model further includes a type field which represents specific details of the word connecting model.