System for controlling multiple user application programs by spoken input

• US 6,125,347 A
• Filed: 09/29/1993
• Issued: 09/26/2000
• Est. Priority Date: 09/29/1993
• Status: Expired due to Term

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: L&H Applications USA, Inc
• Inventors: Cote, William F., MacKenty, Edmund R., Rosnow, Harley M., Kurzweil, Raymond C.
• Primary Examiner(s): Hudspeth, David R.
• Assistant Examiner(s): MCFADDEN, SUSAN IRIS
• Application Number:
  US08/128,990
• Time in Patent Office: 2,554 Days
• Field of Search: 395/2.4, 395/2.44, 395/2.6, 395/2.79, 395/2.84, 395/2.85, 381/43, 381/44, 381/48, 345/333
• US Class Current: 704/275
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• 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 2015/0631 Creating reference templates; Clustering
• CPC Class Codes: G10L 2015/228 of application context
• CPC Class Codes: Y10S 707/99945 Object-oriented database structure processing
• CPC Class Codes: Y10S 707/99948 Application of database or data structure, e.g. distributed, multimedia, or image

Abstract

The system described herein facilitates the control, by spoken input, of multiple user application programs running under an operating system providing multi-tasking capability. A speech recognition program runs as one of the multi-tasked programs. The recognition program is operative to provide translation, either text or program commands, of spoken words in accordance with a translation table. Each of the other user application programs which is to be controlled is provided with a respective translation table section. In accordance with the invention, the recognition program includes means for creating a database for each of the other user application programs which is run and stores, in the respective database, a record reflecting each translation which is sent to a running one of the other application programs. These records identify the speech event which was translated and also identifies alternate possible translations of the speech event. Using the databases, the user can undo the effect of a previously sent translation, select an alternate translation and perform other editing tasks.

Claims

1. A computer readable storage medium encoded with instructions, which when loaded into a digital computer utilizing a multi-tasking operating system that allows a plurality of user application processes to run simultaneously, establishes a speech recognition system, the system comprising:

a speech recognizer running as one of the user application processes, the recognizer providing a recognition output responsive to a speech input, the recognizer further including;

a. an application process selector that directs the recognition output to a designated one of the other user application processes;

b. a recognition history database manager;

i. that creates a recognition history database for each of the other user application processes then running, and

ii. that stores in the recognition history database for each respective other user application process, data constituting a history of the recognition output sent to such other user application process; and

c. undo implementation means, responsive to a spoken user command, for undoing the effect of a previous recognition output provided to a user application receiving such previous recognition result, the undo implementation means utilizing data in the corresponding recognition history database.

2. A computer readable storage medium according to claim 1, wherein the data stored in the recognition history database for each respective use application process includes, with respect to any given recognizer output therein stored, a set of possible recognition output alternatives, and wherein the speech recognizer further includes a recognition correction means, responsive to a spoken user command, for directing a selected recognition output alternative from a recognition history database to the corresponding user application process.

3. A computer readable storage medium according to claim 1, wherein the speech recognizer further includes a set of recognition comparison models associated with each respective other user application process.

4. A method of operating a speech recognizer that allows a plurality of user application processes to run simultaneously in a multi-tasking computer operating system, the recognizer running as one of the user application processes and providing a recognition output responsive to a speech input, the method comprising:

a designating one of the other user application process to receive the recognition output;

b. establishing dynamically a recognition history database for each of the other user application processes then running, and storing dynamically, in each such recognition history database, data constituting a history of the recognition output sent to such other user application process; and

c. utilizing data in the corresponding recognition history database, in response to a spoken user command, to undo the effect of a previous recognition output provided to a user application receiving such previous recognition result.

5. A method according to claim 4, wherein the data stored in the recognition history database for each respective user application process includes with respect to any given recognizer output herein stored, a set of possible recognition output alternatives, and wherein the method further includes directing, in response to a spoken user command, a selected recognition output alternative from a recognition history database to the corresponding user application process.

6. A method according to claim 4, wherein the method further includes associating a set of recognition comparison models with each respective other user application process.