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

• US 20070038449A1
• Filed: 10/23/2006
• Published: 02/15/2007
• Est. Priority Date: 03/01/2004
• Status: Active Grant

• Current Assignee: Robert Coifman
• Original Assignee: Robert Coifman
• Inventors: Coifman, Robert
• Granted Patent: US 7,805,299 B2
• Time in Patent Office: Days
• US Class Current: 704/243
• 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. A method for using a speech recognition system with a records database, said speech recognition system including a vocabulary database, said vocabulary database having default vocabulary elements used by said speech recognition system and imported vocabulary elements from said records database, said method comprising:

providing a speech input to said speech recognition system;

evaluating said speech input against said vocabulary elements within said vocabulary database to determine a probable match set of vocabulary elements, said probable match set being determined according to default weightings for said vocabulary elements as determined by said speech recognition system;

providing said probable match set to a vocabulary prioritization module for use with said records database;

said user prioritization module assigning use weightings to said vocabulary elements within said records database according to a plurality of use criteria and creating a predefined template according to said plurality of use criteria;

creating a virtual vocabulary from said probable match set according to said predefined use template;

modifying said default weights of said vocabulary elements within said probable match set according to the presence of said vocabulary elements within said virtual vocabulary; and

selecting a vocabulary element as a match for said speech input based on said probable match set having said modified weightings.

2. A method for creating a user database system for use with a speech recognition system, said user database system including a user database having user vocabulary elements, said speech recognition system including a vocabulary database having default vocabulary elements, said method comprising:

importing vocabulary elements from said user database into said speech recognition system;

creating a virtual vocabulary

providing said probable match set to a vocabulary prioritization module for use with said user database;

said user prioritization module assigning use weightings to said vocabulary elements within said user database according to a plurality of use criteria and creating a predefine use template according to said plurality of use criteria;

creating a virtual vocabulary from said probable match set according to said predefined use template;

modifying said default weights of said vocabulary elements within said probable match set according to the presence of said vocabulary elements within said virtual vocabulary; and

selecting a vocabulary element as a match for said speech input based on said probable match set having said modified weightings.

3. A software system for use with a computerized speech recognition system, said speech recognition system including a default vocabulary database, said default vocabulary database having vocabulary elements used by said speech recognition system, said default vocabulary elements having preexisting weightings according to a default metric of said speech recognition system, said system comprising:

an adjunct vocabulary database having adjunct vocabulary elements, said adjunct vocabulary elements having weightings according to a first adjunct metric,

wherein said speech recognition system evaluates a speech input to said system against individual vocabulary elements within said default vocabulary according to said default weightings to create a default match set and against individual vocabulary elements within said adjunct vocabulary according to said first adjunct weightings to create an adjunct match set;

said speech recognition system selecting a vocabulary element as a match for said speech input based on said combined default vocabulary match set and said adjunct vocabulary match set.