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

• US 20110153620A1
• Filed: 09/27/2010
• Published: 06/23/2011
• Est. Priority Date: 03/01/2003
• Status: Active Grant

• Current Assignee: Robert E. Coifman
• Original Assignee: Robert E. Coifman
• Inventors: Coifman, Robert E.
• Granted Patent: US 10,733,976 B2
• Time in Patent Office: Days
• US Class Current: 707/748
• CPC Class Codes: G10L 15/065 Adaptation

Abstract

A virtual vocabulary database is provided for use with a with a particular user database as part of a speech recognition system. Vocabulary elements within the virtual database are imported from the user database and are tagged to include numerical data corresponding to the historical use of the vocabulary element within the user database. For each speech input, potential vocabulary element matches from the speech recognition system are provided to the virtual database software which creates virtual sub-vocabularies from the criteria according to predefined criteria templates. The software then applies vocabulary element weighting adjustments according to the virtual sub-vocabulary weightings and applies the adjustment to the default weighting provided by the speech recognition system. The modified weightings are returned with the associated vocabulary elements to the speech engine for selection of an appropriate match to the input speech.

Claims

1-19. -19. (canceled)

20. A records database for use within a speech recognition system comprising:

a plurality of dictated data records, each dictated data record containing a plurality of data fields, said data fields containing dictated vocabulary elements previously dictated into said speech recognition system, at least one of said dictated vocabulary elements having an associated tag, said tag including weighting data indicating a frequency of use of said dictated vocabulary element within said speech recognition system, said weighting data being updated following the dictation of said associated dictated vocabulary element by users of said speech recognition system.

21. The records database of claim 20 further comprising:

a plurality of general data records, each general data record containing a plurality of data fields, said data fields containing general vocabulary elements provided from a general vocabulary and not previously dictated into said speech recognition system, said general vocabulary elements being updated to include a tag including weighting data indicating a frequency of use, said updated general vocabulary element becoming a dictated vocabulary element within said dictated data records once said vocabulary element is dictated into said records database.

22. The records database of claim 20 wherein said vocabulary element is a single word text string.

23. The records database of claim 20 wherein said vocabulary element is a multiple word text string.

24. The records database of claim 20 wherein said vocabulary element is a phoneme.

25. The records database of claim 20 wherein said plurality of dictated data records are stored on a database located on the internet and said dictated vocabulary elements are generated at a dictation office connected to the internet, said weighting data of said tag associated with said dictated vocabulary element being updated when dictated at said dictation office.

26. The records database of claim 25 wherein said records database is an electronic medical records database and said dictated data records are patient data records.

27. The records database of claim 26 wherein said data fields include a medical code.

28. The records database of claim 20 wherein a first portion of said plurality of dictated data records are stored on a database located on the internet and a second portion of said plurality of dictated data records are stored in a dictation office connected to the internet, said dictated vocabulary elements within both said first and second portions being generated at said dictation office, said weighting data of said tags associated with said dictated vocabulary elements of both said first and second portions of said data records being updated when dictated at said dictation office.

29. The records database of claim 28 wherein said database is an electronic medical records database, said dictated data records are patient data records, said first portion of said dictated data records including medical codes and said second portion of said dictated data records include patient data.