Method and apparatus for improving the transcription accuracy of speech recognition software

• US 20040254791A1
• Filed: 03/01/2004
• Published: 12/16/2004
• Est. Priority Date: 03/01/2003
• Status: Active Grant

• Current Assignee: Frederick J. Bender, Robert E. Coifman
• Original Assignee: Frederick J. Bender, Robert E. Coifman
• Inventors: Coifman, Robert E., Bender, Frederick J.
• Granted Patent: US 7,426,468 B2
• Time in Patent Office: Days
• US Class Current: 704/246
• CPC Class Codes: G10L 15/07 to the speaker
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 15/32 Multiple recognisers used in sequence or in parallel; Score combination systems therefor, e.g. voting systems
• CPC Class Codes: G10L 2015/228 of application context

Abstract

The present invention involves the dynamic loading and unloading of relatively small text-string vocabularies within a speech recognition system. In one embodiment, sub-databases of high likelihood text strings are created and prioritized such that those text strings are made available within definable portions of computer-transcribed dictations as a first-pass vocabulary for text matches. Failing a match within the first-pass vocabulary, the voice recognition software attempts to match the speech input to text strings within a more general vocabulary. In another embodiment, the first-pass text string vocabularies are organized and prioritized and loaded in relation to specific fields within an electronic form, specific users of the system and/or other general context-based, interrelationships of the data that provide a higher probability of text string matches then those otherwise provided by commercially available speech recognition systems and their general vocabulary databases.

Claims

1. A method for improving the accuracy of a computerized, speech recognition system, said speech recognition system including a base vocabulary, the method comprising:

loading a specified vocabulary into computer storage, said specified vocabulary associated with a specific context;

accepting a user'"'"'s voice input into said speech recognition system;

evaluating said user'"'"'s voice input with data values from said specified vocabulary according to an evaluation criterion;

selecting a particular data value as an input into a computerized form field if said evaluation criterion is met; and

if said user'"'"'s voice input does not meet said evaluation criterion, selecting a data value from said base vocabulary as an input into said computerized form field.

2. The method of claim 1 further comprising evaluating said user'"'"'s voice input with data values from said base vocabulary according to a base evaluation criterion if said user'"'"'s voice input does not meet said evaluation criterion.

3. The method of claim 1 wherein said evaluation criterion is a use weighting associated with said data values.

4. The method of claim 1 wherein said step of evaluating further includes the step of applying a matching heuristic against a known threshold.

5. The method of claim 3 wherein said step of applying a matching heuristic further includes a step of comparing said user'"'"'s voice input to a threshold probability of matching an acoustic model derived from said specified vocabulary.

6. The method of claim 1 wherein said context is associated with a topical subject.

7. The speech recognition system of claim 1 wherein said context is associated with a specific user.

8. The speech recognition system of claim 1 wherein said context is associated with said field.

9. A method for improving the accuracy of a computerized, speech recognition system comprising:

loading a first specified vocabulary into computer storage, said first specified vocabulary associated with a first computerized form field;

accepting a users voice input into said speech recognition system;

evaluating said user'"'"'s voice input with data values from said first specified vocabulary according to an evaluation criterion;

selecting a particular data value as input into said first computerized form field if said user'"'"'s voice input meets said evaluation criterion;

loading a second specified vocabulary into computer storage, said second specified vocabulary associated with a second computerized form field;

accepting a user'"'"'s voice input into said speech recognition system;

evaluating said user'"'"'s voice input with against data values from said specified vocabulary according to an evaluation criterion; and

selecting a particular data value as input into a second computerized form field if said users voice input meets said evaluation criterion.

10. The method of claim 9 wherein said evaluation criterion for said steps of evaluating said first and said second specified vocabularies are the same.

11. The method of claim 9 wherein said evaluation criterion for said steps of evaluating said first and said second specified vocabularies are different criterion.

12. The method of claim 9 wherein said first and second computerized form fields are associated with different fields of a computerized medical form.

13. A method for improving the accuracy of a computerized, speech recognition system comprising:

loading a first specified vocabulary into computer storage, said first specified vocabulary associated with a first user of said speech recognition system;

accepting said first user'"'"'s voice input into said speech recognition system;

evaluating said first user'"'"'s voice input with data values from said first specified vocabulary according to an evaluation criterion;

selecting a particular data value as an input into a computerized form field if said first user'"'"'s voice input meets said evaluation criterion;

loading a second specified vocabulary into computer storage, said second specified vocabulary associated with a second user of said speech recognition system;

accepting a second user'"'"'s voice input into said speech recognition system;

evaluating said second user'"'"'s voice input with data values from said specified vocabulary according to an evaluation criterion; and

selecting a particular data value as an input into said computerized form field if said second user'"'"'s voice input meets said evaluation criterion.

14. The method of claim 13 wherein said first and second users of said speech recognition system are different doctors and said computerized form fields are associated with a field within a computerized medical form.

15. A method for improving the accuracy of a computerized, speech recognition system comprising:

loading a first specified vocabulary into computer storage, said first specified vocabulary associated with a first context used within said speech recognition system;

accepting a user'"'"'s voice input into said speech recognition system;

evaluating said user'"'"'s voice input with data values from said first specified vocabulary according to an evaluation criterion;

selecting a particular data value as an input into a computerized form field if said user'"'"'s voice input meets said evaluation criterion;

loading a second specified vocabulary into computer storage, said second specified vocabulary associated with a second context used within said speech recognition system;

accepting said user'"'"'s voice input into said speech recognition system;

evaluating said user'"'"'s voice input with data values from said specified vocabulary according to an evaluation criterion; and

selecting a particular data value as an input into said computerized form field if said user'"'"'s voice input meets said evaluation criterion.

16. The method of claim 15 wherein said first context is a patient'"'"'s age and said second context is a patient diagnosis of said patient.

17. A computerized speech recognition system comprising:

a computerized form including at least one computerized form field;

a first vocabulary database containing data entries for said computerized form field, said first vocabulary associated with a specific criterion;

a second vocabulary database containing data entries for said data field; and

an input for accepting a user'"'"'s vocal input, said vocal input being compared to said first vocabulary as a first pass in selecting an input for said computerized form field, and said vocal input being compared to said second vocabulary as a second pass in selecting an input for said computerized form field.

18. The speech recognition system of claim 17 wherein said criterion is a topical context.

19. The speech recognition system of claim 17 wherein said criterion is associated with a specific user of said speech recognition system.

20. The speech recognition system of claim 17 wherein said criterion is associated with said field.

21. The speech recognition system of claim 17 wherein said first vocabulary database is a subset of said second vocabulary database.

22. A database of data values for use in a computerized speech recognition system comprising:

a first vocabulary database containing data entries for a computerized form including at least one computerized form field, said first vocabulary associated with a specific criterion; and

a second vocabulary database containing data entries for said data field.

23. The speech recognition system of claim 22 wherein said criterion is a topical context.

24. The speech recognition system of claim 22 wherein said criterion is associated with a specific user of said speech recognition system.

25. The speech recognition system of claim 22 wherein said criterion is associated with said field.