Distributed speech recognition system

• US 7,146,321 B2
• Filed: 10/31/2001
• Issued: 12/05/2006
• Est. Priority Date: 10/31/2001
• Status: Expired due to Term

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: Dictaphone Corporation (Microsoft Corporation)
• Inventors: Cyr, James, Green, Channell, Hold, Martin, Kuhnen, Regina, MacGinite, Andrew
• Primary Examiner(s): Hudspeth, David
• Assistant Examiner(s): Sked, Matthew J
• Application Number:
  US09/984,874
• Publication Number:
  US 20030083883A1
• Time in Patent Office: 1,861 Days
• Field of Search: 704/270, 704/270.1, 704/231, 704/251
• US Class Current: 704/270.1
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
• CPC Class Codes: G10L 15/34 Adaptation of a single recogniser for parallel processing, e.g. by use of multiple processors or cloud computing

Abstract

A distributed speech recognition system includes a speech processor linked to a plurality of speech recognition engines. The speech processor includes an input for receiving speech files from a plurality of users and storage means for storing the received speech files until such a time that they are forwarded to a selected speech recognition engine for processing. The speech processor further includes a dispatch system linked to the storage means for controlling the transmission of speech files to the plurality of speech recognition engines in a controlled manner.

Claims

1. A distributed speech recognition system, comprising:

a speech processor linked to a plurality of speech recognition engines, the speech processor comprising;

an input for receiving speech files from a plurality of users and a storage means for storing received speech files until such time that they are forwarded to a selected speech recognition engine for processing; and

a dispatch system linked to the storage means, for controlling the transmission of speech files to the plurality of speech recognition engines in a controlled manner;

a dynamic monitoring agent which monitors the availability and turnaround time of each of the speech recognition engines linked to the speech processor and performs a comparative analysis of their availability and turnaround time for use in forwarding speech files to the plurality of speech recognition engines.

2. The distributed speech recognition system according to claim 1, wherein the dynamic monitoring agent assigns a weighting factor to each of the plurality of speech recognition engines, and the weighting factor is utilized in assigning speech files to the plurality of speech recognition engines.

3. The distributed speech recognition system according to claim 1, wherein the dispatch system is linked to the dynamic monitoring agent for optimizing the controlled transmission of speech files to selected speech recognition engines.

4. The distributed speech recognition system according to claim 1, wherein the dynamic monitoring agent identifies speech recognition engines most proficient with specific vocabularies and instructs the dispatch system to forward similar speech files to those speech recognition engines best suited for processing of the selected speech file.

5. The distributed speech recognition system according to claim 1, wherein each of the plurality of speech recognition engines further comprises a speech engine wrapper facilitating interface between the speech processor and the speech recognition engine.

6. The distributed speech recognition system according to claim 1, wherein the speech processor further comprises a database of user files which are subsequently combined with speech files prior to transmission to preselected speech recognition engines.

7. The distributed speech recognition system according to claim 1, further comprising an audit system associated with the speech processor.

8. The distributed speech recognition system according to claim 1, further comprising a voice processor associated with the speech processor.

9. The distributed speech recognition system according to claim 1, further comprising a text processor associated with the speech processor.

10. A method for implementing a distributed speech recognition system, comprising the following steps:

linking a speech processor to a plurality of speech recognition engines, the speech processor comprising an input for receiving speech files from a plurality of users and a storage means for storing the received speech files until such a time that they are forwarded to a selected speech recognition engine for processing;

providing the speech processor with a dispatch system linked to the storage means, for controlling the transmission of speech files to the plurality of speech recognition engines in a controlled manner;

linking a dynamic monitoring agent to the dispatch system, wherein the dynamic monitoring agent monitors the availability and turnaround time of each of the speech recognition engines linked to the speech processor and performs a comparative analysis of their availability and turnaround time for use in forwarding speech files to the plurality of speech recognition engines; and

distributing speech files to the various speech recognition engines under the control of the dispatch system.

11. The method according to claim 10, wherein the dynamic monitoring agent assigns a weighting factor to each of the plurality of speech recognition engines, and the weighting factor is utilized in assigning speech files to the plurality of speech recognition engines.

12. The method according to claim 10, wherein the dynamic monitoring agent identifies speech recognition engines most proficient with specific vocabularies and instructs the dispatch system to forward similar speech files to those speech recognition engines best suited for processing of the selected speech file.

13. The method according to claim 10, wherein each of the plurality of speech recognition engines comprises a speech recognition engine wrapper facilitating interface between the speech processor and the speech recognition engine.

14. The method according to claim 10, further comprising the step of providing the speech processor with a database of user files which are subsequently combined with speech files prior to transmission to preselected speech recognition engines.

15. The method according to claim 10, further comprising the step of linking an audit system with the speech processor.

16. The method according to claim 10, further comprising the step of linking a voice processor associated with the speech processor.

17. The method according to claim 10, further comprising the step of linking a text processor with the speech processor.