Word recognition system using language context at current cursor position to affect recognition probabilities

• US 5,920,836 A
• Filed: 06/26/1997
• Issued: 07/06/1999
• Est. Priority Date: 11/13/1992
• Status: Expired due to Term

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: Dragon Systems, Inc. (Microsoft Corporation)
• Inventors: Gould, Joel M., McGrath, Frank J.
• Primary Examiner(s): Hudspeth, David R.
• Assistant Examiner(s): Chawan, Vijay B.
• Application Number:
  US08/882,813
• Time in Patent Office: 740 Days
• Field of Search: 704/240, 704/243, 704/245, 704/251, 704/253, 704/255, 704/256, 704/275, 704/241
• US Class Current: 704/251
• CPC Class Codes: G09B 19/04 Speaking with audible presentation of the material to be studied G09B5/04
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 15/18 using natural language modelling
• 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 15/285 Memory allocation or algorithm optimisation to reduce hardware requirements
• CPC Class Codes: G10L 2015/0631 Creating reference templates; Clustering
• CPC Class Codes: G10L 2015/0638 Interactive procedures
• CPC Class Codes: G10L 2015/228 of application context

Abstract

A computerized system performs word recognition. It does pattern matching to select which vocabulary words appears, according to the pattern matching, to most probably correspond to word signals it seeks to recognize. It produces an output corresponding to the vocabulary words selected at a movable cursor position in a body of text. The system improves its recognition by obtaining information about the linguistic context of the current cursor position in the body of text and using that information to vary the probability of which one or more words is selected by its pattern matching. The pattern matching can be speech recognition performed on signals representing the sound of spoken words. In some embodiments, the word recognition is executed on a computer system capable of running multiple programs at once; the recognized word is output to another program for insertion at a cursor position in that other program; and the linguistic context information is obtained from data structures created by the other program.

Claims

1. A computerized method of performing word recognition comprising the steps of:

receiving user generated word signals representing words to be recognized;

performing pattern matching upon the word signals to select which one or more of a plurality of vocabulary words appears, according to said pattern matching, to most probably correspond to each such word signal;

producing an output, at a movable cursor position in a body of text, corresponding to the one or more vocabulary words selected by said pattern matching for each of said word signals;

obtaining information about the linguistic context of the current cursor position in said body of text and using said information to define a current language context and using said information to define a current language context; and

varying the probability of which one or more words is selected by said pattern matching as appearing to most probably corresponding to a given word signals as a function of said current language context.

2. A computerized method as in claim 1 wherein:

said word signals are acoustic signals representing the sound of spoken words; and

said pattern matching performs speech recognition on said acoustic word signals.

3. A computerized method as in claim 1 wherein:

said method is executed on a computer system capable of running a plurality of active computer programs at one time;

said step of producing an output includes supplying the spelling of vocabulary words selected by said pattern matching to another program running on said computer system for insertion at a cursor position into a body of text represented by that other program; and

said step of obtaining information about the linguistic context of the current cursor position includes obtaining such information from data structures created by said other program.

4. A computer program stored in machine readable memory for performing word recognition comprising the following program instructions:

instructions for receiving user generated word signals representing words to be recognized;

pattern matching instructions for performing pattern matching upon the word signals to select which one or more of a plurality of vocabulary words appears, according to said pattern matching, to most probably correspond to each such word signal

output instructions for producing an output, at a movable cursor position in a body of text, corresponding to the one or more vocabulary words selected by said pattern matching for each of said word signals;

context detecting program instructions for obtaining information about the linguistic context of the current cursor position in said body of text and using said information to define a current language context; and

probability altering instructions for varying the probability of which one or more words is selected by said pattern matching as appearing to most probably corresponding to a given word signals as a function of said current language context.

5. A computer program as in claim 4 wherein:

said word signals are acoustic signals representing the sound of spoken words; and

said pattern matching instructions include instructions for performing speech recognition on said acoustic word signals.

6. A computer program as in claim 4 wherein:

said output instructions includes instructions which supply the spelling of vocabulary words selected by said pattern matching instructions to another program running on said computer system for insertion at a cursor position into a body of text represented by that other program; and

said context detecting instructions include instructions for obtaining information about the textual context of the current cursor position in said other program.

7. A computer system capable of performing word recognition comprising:

means for receiving user generated word signals representing words to be recognized;

means for performing pattern matching upon the word signals to select which one or more of a plurality of vocabulary words appears, according to said pattern matching, to most probably correspond to each such word signal

means for producing an output, at a movable cursor position in a body of text, corresponding to the one or more vocabulary words selected by said pattern matching for each of said word signals;

means for obtaining information about the linguistic context of the current cursor position in said body of text and using said information to define a current language context; and

means for varying the probability of which one or more words is selected by said pattern matching as appearing to most probably corresponding to a given word signals as a function of said current language context.

8. A computer system as in claim 7 wherein:

said word signals are acoustic signals representing the sound of spoken words; and

said means for performing pattern matching include means for performing speech recognition on said acoustic word signals.

9. A computer program as in claim 7 wherein:

said output instructions includes instructions which supply the spelling of vocabulary words selected by said pattern matching instructions to another program running on said computer system for insertion at a cursor position into a body of text represented by that other program; and

said context detecting instructions include instructions for obtaining information about the textual context of the current cursor position in said other program.