Position manipulation in speech recognition

• US 6,601,027 B1
• Filed: 06/15/1998
• Issued: 07/29/2003
• Est. Priority Date: 11/13/1995
• 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: Young, Jonathan Hood, Parmenter, David W., Gould, Joel M., Gold, Allan, Dubach, Joev, Wright, Barton D.
• Primary Examiner(s): Banks-Harold, Marsha D.
• Assistant Examiner(s): Lerner, Martin
• Application Number:
  US09/094,611
• Time in Patent Office: 1,870 Days
• Field of Search: 704/231, 704/235, 704/251, 704/254, 704/255, 704/257, 704/275, 704/270
• US Class Current: 704/235
• CPC Class Codes: G06F 3/16 Sound input; Sound output speech processing G10L
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2015/223 Execution procedure of a spoken command
• CPC Class Codes: G10L 2015/228 of application context

Abstract

An action position is manipulated in computer-implemented speech recognition by receiving data representing a spoken command. The command includes a command identifier (e.g., insert before, insert after, resume with) and a designation of at least one previously-spoken word. Speech recognition is performed on the data to identify the command identifier and the designation. Finally, an action position is established relative to the previously-spoken word based on the command identifier. Text may be selected using a spoken selection command that includes a command identifier and a text block identifier identifying a block of previously-recognized text. At least one word included in the block of text is not included in the text block identifier.

Claims

1. A method of manipulating an action position in computer-implemented speech recognition, comprising:

receiving data representing a spoken command, the command including a command identifier and a designation of at least one previously-spoken word;

performing speech recognition on the data to identify the command identifier and the designation; and

establishing an action position measured from a position of the at least one previously-spoken word based on the command identifier;

wherein the designation comprises the at least one previously-spoken word.

2. The method of claim 1, wherein the command identifier indicates that the action position is to be before the position of the at least one previously-spoken word, and establishing the action position comprises establishing the action position immediately prior to the position of the at least one previously-spoken word.

3. The method of claim 2, wherein the command identifier comprises “

insert before”

.

4. The method of claim 1, where in the command identifier indicates that the action position is to be after the position of the at least one previously-spoken word, and establishing the action position comprises establishing the action position immediately after the position of the at least one previously-spoken word.

5. The method of claim 4, wherein the command identifier comprises “

insert after”

.

6. The method of claim 1, wherein the designation comprises at least one new word, the method further comprising replacing any words following the at least one previously-spoken word with the at least one new word.

7. The method of claim 6, wherein establishing the action position comprises establishing the action position after the position of the at least one new word.

8. The method of claim 6, wherein the command identifier comprises “

resume with”

.

9. The method of claim 6, wherein receiving data comprises receiving data generated by recording the spoken command using a recording device physically separate from a computer implementing the speech recognition.

10. The method of claim 9, wherein

the recording device comprises a digital recording device, and

receiving data comprises receiving a file from the digital recording device.

11. The method of claim 9, wherein receiving data comprises receiving signals generated by playing back the spoken command using the recording device.

12. The method of claim 11, wherein the recording device comprises an analog recording device.

13. The method of claim 1, wherein receiving data comprises receiving data generated by recording the spoken command using a recording device physically separate from a computer implementing the speech recognition.

14. The method of claim 13, wherein

the recording device comprises a digital recording device, and

receiving data comprises receiving a file from the digital recording device.

15. The method of claim 13, wherein receiving data comprises receiving signals generated by playing back the spoken command using the recording device.

16. The method of claim 15, wherein the recording device comprises an analog recording device.

17. A method of selecting a block of text in computer-implemented speech recognition, comprising:

receiving data representing a spoken selection command, the command including a command identifier and a text block identifier identifying a block of previously-recognized text, wherein at least one word included in the block of text is not included in the text block identifier;

performing speech recognition on the data to identify the command identifier and the text block identifier; and

selecting the block of text corresponding to the text block identifiers;

wherein the text block identifier includes at least a first previously-recognized word of the block of text and at least a last previously-recognized word of the block of text.

18. The method of claim 17, wherein the command identifier comprises “

select” and

the text block identifier comprises the first previously-recognized word of the block of text, “

through”

, and the last previously-recognized word of the block of text.

19. The method of claim 17, wherein performing speech recognition on the data comprises performing speech recognition using a constraint grammar.

20. The method of claim 19, wherein the constraint grammar permits the block of text to start with any word in a set of previously-recognized words and to end with any word in the set of previously-recognized words.

21. The method of claim 20, wherein the set of previously-recognized words comprises previously-recognized words displayed on a display device when the selection command is spoken.

22. The method of claim 20, wherein performing speech recognition comprises generating multiple candidates for the text block identifier and further comprises eliminating candidates for which the block of text starts with a previously-recognized word spoken after a previously-recognized word with which the block of text ends.

23. The method of claim 22, wherein performing speech recognition comprises associating a score with each of the multiple candidates.

24. The method of claim 23, wherein a score for a candidate is based on scores for components of the candidate.

25. The method of claim 24, further comprising, when components of different candidates are homophones, adjusting the scores for the candidates so that the portion of each score attributable to one of the homophones equals the score of the best-scoring one of the homophones.

26. A method of selecting a block of text in computer-implemented speech recognition, comprising:

receiving data representing a spoken selection command, the command including a command identifier and a text block identifier identifying a block of previously-recognized text, wherein at least one word included in the block of text is not included in the text block identifier;

performing speech recognition on the data to identify the command identifier and the text block identifier; and

selecting the block of text corresponding to the text block identifier,

wherein the text block identifier designates a previously-recognized utterance without using words included in the utterance.

27. The method of claim 26, wherein the text block identifier comprises “

that”

.

28. The method of claim 26, wherein the command identifier consists of “

select” and

the text block identifier consists of “

that”

so that the spoken selection command consists of “

select that”

.

29. Computer software, residing on a computer-readable storage medium, comprising instructions for causing a computer to:

receive data representing a spoken command, the command including a command identifier and a designation of at least one previously-spoken word;

perform speech recognition on the data to identify the command identifier and the designation; and

establish an action position measured from a position of the at least one previously-spoken word based on the command identifier;

wherein the designation comprises the at least one previously-spoken word.

30. A speech recognition system, comprising:

an input device for receiving speech signals; and

a processor configured to;

receive data representing a spoken command, the command including a command identifier and a designation of at least one previously-spoken word;

perform speech recognition on the data to identify the command identifier and the designation; and

establish an action position measured from a position of the at least one previously-spoken word based on the command identifier;

wherein the designation comprises the at least one previously-spoken word.

31. Computer software, residing on a computer-readable storage medium, comprising instructions for causing a computer to:

receive data representing a spoken selection command, the command including a command identifier and a text block identifier identifying a block of previously-recognized text, wherein at least one word included in the block of text is not included in the text block identifier;

perform speech recognition on the data to identify the command identifier and the text block identifier; and

select the block of text corresponding to the text block identifiers;

wherein the text block identifier includes at least a first previously-recognized word of the block of text and at least a last previously-recognized word of the block of text.

32. A speech recognition system, comprising:

an input device for receiving speech signals; and

a processor configured to;

receive data representing a spoken selection command, the command including a command identifier and a text block identifier identifying a block of previously-recognized text, wherein at least one word included in the block of text is not included in the text block identifier;

perform speech recognition on the data to identify the command identifier and the text block identifier; and

select the block of text corresponding to the text block identifiers wherein the text block identifier includes at least a first previously-recognized word of the block of text and at least a last previously-recognized word of the block of text.

33. A computer-based method for use in working with text, comprising:

receiving a command including an utterance designating a portion of the text, performing speech recognition on the utterance to identify the portion of the text, and establishing an action position in the text;

wherein;

the action position is measured from a position of the identified portion of the text and is determined by the command, and

the utterance comprises the portion of text.

34. A method of manipulating an action position in computer-implemented speech recognition, comprising:

receiving data representing a spoken command, the command including a command identifier and a designation of at least one previously-spoken word;

performing speech recognition on the data to identify the command identifier and the designation;

establishing a position measured from the at least one previously-spoken word based on the command identifier; and

performing an action at the established position, the action including inserting at least one word;

wherein the designation is independent of a number of words between the at least one previously-spoken word and the established position.

35. The method of claim 34, wherein the designation comprises a designation term that represents a multi-word utterance.

36. The method of claim 35, wherein the designation term comprises “

that” and

the multi-word utterance comprises an utterance spoken and recognized immediately prior to the spoken command.

37. Computer software, residing on a computer-readable storage medium, comprising instructions for causing a computer to:

receive data representing a spoken command, the command including a command identifier and a designation of at least one previously-spoken word;

perform speech recognition on the data to identify the command identifier and the designation;

establish a position measured from the at least one previously-spoken word based on the command identifier; and

perform an action at the established position, the action including inserting at least one word;

wherein the designation is independent of a number of words between the at least one previously-spoken word and the established position.

38. A speech recognition system, comprising:

an input device for receiving speech signals; and

a processor configured to;

receive data representing a spoken command, the command including a command identifier and a designation of at least one previously-spoken word;

perform speech recognition on the data to identify the command identifier and the designation;

establish a position measured from the at least one previously-spoken word based on the command identifier; and

perform an action at the established position, the action including inserting at least one word;

wherein the designation is independent of a number of words between the at least one previously-spoken word and the established position.

39. A computer-based method for use in working with text, comprising:

receiving a command including an utterance designating a portion of the text,

performing speech recognition on the utterance to identify the portion of the text, and

establishing a position in the text, the position being measured from the identified portion of the text and determined by the command;

performing an action at the established position, the action including inserting at least one word;

wherein the utterance designating the portion of text is independent of a number of words between the portion of text and the established position.