Detecting speech recognition errors in an embedded speech recognition system

• US 6,754,627 B2
• Filed: 03/01/2001
• Issued: 06/22/2004
• Est. Priority Date: 03/01/2001
• Status: Expired due to Term

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: International Business Machines Corporation
• Inventors: Woodward, Steven G.
• Primary Examiner(s): Dorvil, Richemond
• Assistant Examiner(s): Nolan, Daniel A.
• Application Number:
  US09/798,658
• Publication Number:
  US 20020123892A1
• Time in Patent Office: 1,209 Days
• Field of Search: 704/285, 704/270, 704/270.1, 704/275, 704/244, 704/257, 704/251, 704/260, 704/243, 382/100
• US Class Current: 704/235
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2015/0631 Creating reference templates; Clustering
• CPC Class Codes: G10L 2015/225 Feedback of the input speech

Abstract

A method for processing a misrecognition error in an embedded speech recognition system during a speech recognition session can include the step of speech-to-text converting audio input in the embedded speech recognition system based on an active language model. The speech-to-text conversion can produce speech recognized text that can be presented through a user interface. A user-initiated misrecognition error notification can be detected. The audio input and a reference to the active language model can be provided to a speech recognition system training process associated with the embedded speech recognition system.

Claims

1. An embedded speech recognition system comprising:

an embedded computer system;

a speech recognition system configured for operation if said embedded computer system;

a speech-enabled application for processing text converted in said speech recognition system; and

,

misrecognition error logic for notifying said speech recognition system when a misrecognition error has occurred, wherein said misrecognition logic is disposed within said embedded computer system.

2. The embedded speech recognition system of claim 1, further comprising an activatable error notification button coupled to said embedded computer system, said button triggering said misrecognition error logic when activated, wherein said error notification button is an input mechanism explicitly included within said embedded speech recognition system in order to perform said triggering.

3. The embedded speech recognition system of claim 1, wherein said embedded computer system comprises:

a central processing unit (CPU);

memory;

audio circuitry; and

,

an audio input device.

4. The embedded speech recognition system of claim 3, further comprising at least one speech recognition language model wherein said speech recognition language model includes at least one speech command, and wherein said misrecognition logic is triggered responsive to the detection of said speech command within a speech input received via said audio input device.

5. A method for processing a misrecognition error in an embedded speech recognition system during a speech recognition session, comprising:

speech-to-text converting audio input in the embedded speech recognition system based on an active language model, said speech-to-text conversion producing speech recognized text;

presenting said speech recognized text through a user interface;

detecting a user-initiated misrecognition error notification; and

,

responsive to detecting said error notification, providing said audio input and a reference to said active language model to a speech recognition system training process associated with the embedded speech recognition system.

6. The method of claim 5, wherein said detecting step comprises receiving a hardware-generated notification caused by the activation of an error notification button, wherein said error notification button is an input mechanism explicitly included within said embedded speech recognition system in order to generate said notification.

7. The method of claim 5, wherein said detecting step comprises receiving a software-generated notification caused by the receipt of a error notification speech command, wherein said speech command is matched against a previously established error command stored within said embedded speech recognition system.

8. The method of claim 5, wherein said providing step comprises:

storing said audio input;

storing a reference to said active language model; and

,

providing said stored audio input and reference to said training process subsequent to the speech recognition session.

9. A machine readable storage, having stored thereon a computer program for processing a misrecognition error in an embedded speech recognition system during a speech recognition session, said computer program having a plurality of code sections executable by a machine for causing the machine to perform the steps of:

speech-to-text converting audio input in the embedded speech recognition system based on an active language model, said speech-to-text conversion producing speech recognized text;

presenting said speech recognized text through a user interface;

detecting a user-initiated misrecognition error notification; and

,

responsive to detecting said error notification, providing said audio input and a reference to said active language model to a speech recognition system training process associated with the embedded speech recognition system.

10. The machine readable storage of claim 9, wherein said detecting step comprises receiving a hardware-generated notification caused by the activation of an error notification button, wherein said error notification button is an input mechanism explicitly included within said embedded speech recognition system in order to generate said notification.

11. The machine readable storage of claim 9, wherein said detecting step comprises receiving a software-generated notification caused by the receipt of a error notification speech command, wherein said speech command is matched against a previously established error command stored within said embedded speech recognition system.

12. The machine readable storage of claim 9, wherein said providing step comprises:

storing said audio input;

storing a reference to said active language model; and

,

providing said stored audio input and reference to said training process subsequent to the speech recognition session.