Method and apparatus for excluding text phrases during re-dictation in a speech recognition system

• US 6,374,214 B1
• Filed: 06/24/1999
• Issued: 04/16/2002
• Est. Priority Date: 06/24/1999
• Status: Expired due to Term

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: International Business Machines Corporation
• Inventors: Smith, Maria E., Friedland, Steven J.
• Primary Examiner(s): Dorvil, Richemond
• Application Number:
  US09/339,421
• Time in Patent Office: 1,027 Days
• Field of Search: 704/235, 704/270, 704/272, 704/275
• US Class Current: 704/235
• 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 and system for correcting misrecognized phrases in a computer speech recognition system prevents prior misrecognized text phrases from reoccurring during re-dictation. In particular, the present invention provides a method for correcting a user identified misrecognized text at a location in a dictated electronic document. The method includes: receiving dictation from a user; processing the user input to identify a plurality of alternate text selections; excluding the misrecognized text from the plurality of alternate text selections; and replacing it with a replacement text. These steps are repeated each time the user selects and re-dictates text at the location of the misrecognized text.

Claims

1. In a computer system for speech recognition, a method for correcting a user identified misrecognized text having a first location in a dictated electronic document comprising the steps of:

(a) receiving a user input in the form of a spoken utterance corresponding to an intended phrase;

(b) processing said user input to identify a plurality of alternate text selections which are determined statistically as the most likely text recognitions corresponding to said spoken utterance;

(c) excluding said misrecognized text from said plurality of alternate text selections and replacing said misrecognized text with a replacement text selected from the remaining alternate text selections and which is the most likely text recognition;

(d) storing said misrecognized text and a location of said misrecognized text in a memory location for excluded text; and

(e) repeating steps (a)-(d) when said replacement text is misrecognized text, thereby excluding subsequent ones of said misrecognized text from said alternate text selections.

2. The method of claim 1, further comprising the steps of:

receiving a second user input identifying a second misrecognized text at a second location in said document; and

,

clearing said memory location for excluded text.

3. The method of claim 1, further comprising the steps of:

identifying a set of acoustic characteristics of said spoken utterance;

processing said spoken utterance to determine if it has a similar set of acoustic characteristics to those of said excluded text spoken utterance; and

clearing said excluded text from said memory location if said set of acoustic characteristics of said spoken utterance are substantially dissimilar from a set of acoustic characteristics for said excluded text spoken utterance.

4. The method of claim 1, further comprising the steps of:

counting each excluded text stored in said memory location; and

removing the earliest stored excluded text from said memory location when a number of excluded text exceeds a predetermined value.

5. The method of claim 4, wherein said predetermined value is two.

6. A computer speech recognition system for correcting a user identified misrecognized text having a first location in a dictated electronic document, comprising:

input means for receiving a user input in the form of a spoken utterance corresponding to an intended phrase;

processor means for processing said user input to identify a plurality of alternate text selections which are determined statistically as the most likely text recognitions corresponding to said spoken utterance;

wherein said processing means excludes said misrecognized text from said plurality of alternate text selections and replaces said misrecognized text with a replacement text selected from the remaining alternate text selections and which is the most likely text recognition; and

memory means for storing said misrecognized text, a location of said misrecognized text, and subsequent misrecognized text at said location for excluding subsequent ones of said misrecognized text from said alternate text selections in a memory location for excluded text.

7. The system of claim 6, wherein a second user input identifying a second misrecognized phrase at a second location in said document is received by said input means and said memory location for excluded text is cleared.

8. The system of claim 6, further comprising:

identification means for identifying a set of acoustic characteristics of said spoken utterance;

wherein said processor means processes said spoken utterance to determine if it has a similar set of acoustic characteristics to those of said excluded text spoken utterance, and said excluded text is cleared from said memory location if said set of acoustic characteristics of said spoken utterance are substantially dissimilar from a set of acoustic characteristics for said excluded text spoken utterance.

9. The system of claim 6, further comprising a counting means for counting each excluded text stored in said memory location, wherein said earliest stored excluded text is removed from said memory location when a number of excluded text exceeds a predetermined value.

10. The system of claim 9, wherein said predetermined value is two.

11. A machine readable storage, having stored thereon a computer program having a plurality of code sections executable by a machine for causing the machine to correct a user identified misrecognized text having a first location in a dictated electronic document and perform the steps of:

(a) receiving a user input in the form of a spoken utterance corresponding to an intended phrase;

(b) processing said user input to identify a plurality of alternate text selections which are determined statistically as the most likely text recognitions corresponding to said spoken utterance;

(c) excluding said misrecognized text from said plurality of alternate text selections and replacing said misrecognized text with a replacement text selected from the remaining alternate text selection and which is the most likely text recognition;

(d) storing said misrecognized text and a location of said misrecognized text in a memory location for excluded text; and

(e) repeating steps (a)-(d) when said replacement text is misrecognized text, thereby excluding subsequent ones of said misrecognized text from said alternate text selections.