Speech recognition correction for devices having limited or no display

• US 7,200,555 B1
• Filed: 07/05/2000
• Issued: 04/03/2007
• Est. Priority Date: 07/05/2000
• Status: Expired due to Term

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: International Business Machines Corporation
• Inventors: Lewis, James R., Ortega, Kerry A., Wang, Huifang, Van Buskirk, Ronald E., Ballard, Barbara E.
• Primary Examiner(s): Opsasnick; Michael N.
• Application Number:
  US09/610,061
• Time in Patent Office: 2,463 Days
• Field of Search: 704/270, 704/235, 704/276, 704/275
• US Class Current: 704/235
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

A novel apparatus and method for correcting speech recognized text in a predominantly speech-only environment for use with a device having only a limited or no display device available. The method is preferably implemented by a machine readable storage mechanism having stored thereon a computer program, the method comprising the following steps. First, audio speech input can be received and speech-to-text converted to speech recognized text. Second, a first speech correction command for performing a correction operation on speech recognized text stored in a text buffer can be detected in the speech recognized text. Third, if a speech correction command is not detected in the speech recognized text, the speech recognized text can be added to the text buffer. Fourth, if a speech command is detected in the speech recognized text, the detected correction speech command can be performed on speech recognized text stored in the text buffer.

Claims

1. A speech correction method for correcting misrecognized text in a speech recognition application comprising the steps of:

receiving audio speech input and speech-to-text converting said received audio speech input to speech recognized text;

audibly confirming said speech-to-text conversion of said speech recognized text, wherein audibly confirming comprises audibly playing back of said speech recognized text using a speech synthesis engine so that it can be determined if said recorded speech recognized text had been misrecognized in said speech-to-text converting step, determining whether a word of said speech recognized text is a homonym, and if so audibly spelling the homonym using said speech synthesis engine;

detecting in said speech recognized text a first speech correction command for performing a correction operation on speech recognized text stored in a text buffer;

if a first speech correction command is not detected in said speech recognized text, adding said speech recognized text to said text buffer; and

,

if a first speech correction command is detected in said speech recognized text, performing said detected speech correction command on speech recognized text stored in said text buffer.

2. The method according to claim 1, further comprising the step of:

responsive to detecting said first speech correction command in said speech recognized text, said first speech correction command indicating a preference to terminate said speech correction method, determining if said speech recognized text stored in said text buffer had been spelled out;

adding said speech recognized text determined to have been spelled out to a speech recognition vocabulary of speech recognizable words; and

,

terminating said speech correction method.

3. The method according to claim 1, further comprising the step of:

responsive to detecting said first speech correction command in said speech recognized text, said first speech correction command indicating a preference to correct misrecognized text in said text buffer, audibly playing a list of speech correction candidates, wherein each speech correction candidate in said list is statistically alternative recognized text to said audio speech input;

receiving a selection of one of said speech correction candidates in said list; and

,

replacing said misrecognized text in said text buffer with said selected speech correction candidate.

4. A speech correction method for correcting misrecognized text in a speech recognition application, the method comprising the steps of:

receiving audio speech input and speech-to-text converting said received audio speech input to speech recognized text;

detecting in said speech recognized text a first speech correction command for performing a correction operation on speech recognized text stored in a text buffer;

if a first speech correction command is not detected in said speech recognized text, adding said speech recognized text to said text buffer; and

,

if a first speech correction command is detected in said speech recognized text, performing said detected speech correction command on speech recognized text stored in said text buffer;

responsive to detecting said first speech correction command in said speech recognized text, said first speech correction command indicating a preference to correct misrecognized text in said text buffer, audibly playing a list of speech correction candidates, wherein each speech correction candidate in said list is statistically alternative recognized text to said audio speech input;

receiving a second speech correction command, said second speech correction command indicating both preferred replacement text and a preference to replace said misrecognized text with said preferred replacement text in said text buffer; and

,

responsive to receiving said second speech correction command, replacing said misrecognized text in said text buffer with said preferred replacement text.

5. A speech correction method for correcting misrecognized text in a speech recognition application comprising the steps of:

receiving audio speech input and speech-to-text converting said received audio speech input to speech recognized text;

detecting in said speech recognized text a first speech correction command for performing a correction operation on speech recognized text stored in a text buffer;

if a first speech correction command is not detected in said speech recognized text, adding said speech recognized text to said text buffer;

if a first speech correction command is detected in said speech recognized text, performing said detected speech correction command on speech recognized text stored in said text buffer;

responsive to detecting said first speech correction command in said speech recognized text, said first speech correction command indicating a preference to correct misrecognized text in said text buffer, audibly playing a list of speech correction candidates, wherein each speech correction candidate in said list is statistically alternative recognized text to said audio speech input;

receiving a second speech correction command, said second speech correction command indicating a preference to replace said misrecognized text in said text buffer with spelled-out replacement text;

responsive to receiving said second speech correction command, accepting audibly spelled-out replacement text, said audibly spelled-out replacement text comprising a series of spoken alphanumeric characters;

speech-to-text converting said series of spoken alphanumeric characters, each speech-to-text converted alphanumeric character stored in a temporary buffer, and combining said speech-to-text converted alphanumeric characters into spelled out replacement text; and

,

replacing said misrecognized text in said text buffer with said spelled out replacement text.

6. The method according to claim 5, further comprising the steps of:

detecting in said audibly spelled-out replacement text, a third speech correction command, said third speech correction command indicating a preference to delete a particular alphanumeric character stored in said temporary buffer; and

,

responsive to detecting said third speech correction command, deleting said particular alphanumeric character from said temporary buffer.

7. The method according to claim 5, further comprising the steps of:

detecting in said audibly spelled-out replacement text, a third speech correction command, said third speech correction command indicating both a preferred replacement alphanumeric character and a preference to replace a particular alphanumeric character with said preferred replacement alphanumeric character in said temporary buffer; and

,

responsive to detecting said third speech correction command, replacing said particular alphanumeric character with said preferred alphanumeric character in said temporary buffer.

8. The method according to claim 5, wherein said accepting step further comprises the step of:

prior to accepting audibly spelled out replacement text, playing a pre-stored set of instructions for providing said spelled out replacement text.

9. A machine readable storage, having stored thereon a computer program having a plurality of code sections for performing a speech correction method for correcting misrecognized text in a speech recognition application, said code sections executable by a machine for causing a machine to perform the steps of:

receiving audio speech input and speech-to-text converting said received audio speech input to speech recognized text;

audibly confirming said speech-to-text conversion of said speech recognized text, wherein audibly confirming comprises playing back of said speech recognized text using a speech synthesis engine so that it can be determined if said recorded speech recognized text had been misrecognized in said speech-to-text converting step, determining whether a word of said speech recognized text is a homonym, and if so audibly spelling the homonym using said speech synthesis engine;

detecting in said speech recognized text a first speech correction command for performing a correction operation on speech recognized text stored in a text buffer;

if a first speech correction command is not detected in said speech recognized text, adding said speech recognized text to said text buffer; and

,

if a first speech correction command is detected in said speech recognized text, performing said detected speech correction command on speech recognized text stored in said text buffer.

10. The machine readable storage according to claim 9, further comprising the step of:

responsive to detecting said first speech correction command in said speech recognized text, said first speech correction command indicating a preference to terminate said speech correction method, determining if said speech recognized text stored in said text buffer had been spelled out;

adding said speech recognized text determined to have been spelled out to a speech recognition vocabulary of speech recognizable words; and

,

terminating said speech correction method.

11. The machine readable storage according to claim 9, further comprising the step of:

responsive to detecting said first speech correction command in said speech recognized text, said first speech correction command indicating a preference to correct misrecognized text in said text buffer, audibly playing a list of speech correction candidates, wherein each speech correction candidate in said list is statistically alternative recognized text to said audio speech input;

receiving a selection of one of said speech correction candidates in said list; and

,

replacing said misrecognized text in said text buffer with said selected speech correction candidate.

12. A machine readable storage, having stored thereon a computer program having a plurality of code sections for performing a speech correction method for correcting misrecognized text in a speech recognition application, said code sections executable by a machine for causing a machine to perform the steps of:

receiving audio speech input and speech-to-text converting said received audio speech input to speech recognized text;

detecting in said speech recognized text a first speech correction command for performing a correction operation on speech recognized text stored in a text buffer;

if a first speech correction command is not detected in said speech recognized text, adding said speech recognized text to said text buffer;

if a first speech correction command is detected in said speech recognized text, performing said detected speech correction command on speech recognized text stored in said text buffer;

responsive to detecting said first speech correction command in said speech recognized text, said first speech correction command indicating a preference to correct misrecognized text in said text buffer, audibly playing a list of speech correction candidates, wherein each speech correction candidate in said list is statistically alternative recognized text to said audio speech input;

receiving a second speech correction command, said second speech correction command indicating both preferred replacement text and a preference to replace said misrecognized text with said preferred replacement text in said text buffer; and

,

responsive to receiving said second speech correction command, replacing said misrecognized text in said text buffer with said preferred replacement text.

13. A machine readable storage, having stored thereon a computer program having a plurality of code sections for performing a speech correction method for correcting misrecognized text in a speech recognition application, said code sections executable by a machine for causing a machine to perform the steps of:

receiving audio speech input and speech-to-text converting said received audio speech input to speech recognized text;

detecting in said speech recognized text a first speech correction command for performing a correction operation on speech recognized text stored in a text buffer;

if a first speech correction command is not detected in said speech recognized text, adding said speech recognized text to said text buffer;

if a first speech correction command is detected in said speech recognized text, performing said detected speech correction command on speech recognized text stored in said text buffer;

responsive to detecting said first speech correction command in said speech recognized text, said first speech correction command indicating a preference to correct misrecognized text in said text buffer, audibly playing a list of speech correction candidates, wherein each speech correction candidate in said list is statistically alternative recognized text to said audio speech input;

receiving a second speech correction command, said second speech correction command indicating a preference to replace said misrecognized text in said text buffer with spelled-out replacement text;

responsive to receiving said second speech correction command, accepting audibly spelled-out replacement text, said audibly spelled-out replacement text comprising a series of spoken alphanumeric characters;

speech-to-text converting said series of spoken alphanumeric characters, each speech-to-text converted alphanumeric character stored in a temporary buffer, and combining said speech-to-text converted alphanumeric characters into spelled out replacement text; and

,

replacing said misrecognized text in said text buffer with said spelled out replacement text.

14. The machine readable storage according to claim 13, further comprising the steps of:

detecting in said audibly spelled-out replacement text, a third speech correction command, said third speech correction command indicating a preference to delete a particular alphanumeric character stored in said temporary buffer; and

,

responsive to detecting said third speech correction command, deleting said particular alphanumeric character from said temporary buffer.

15. The machine readable storage according to claim 13, further comprising the steps of:

detecting in said audibly spelled-out replacement text, a third speech correction command, said third speech correction command indicating both a preferred replacement alphanumeric character and a preference to replace a particular alphanumeric character with said preferred replacement alphanumeric character in said temporary buffer; and

,

responsive to detecting said third speech correction command, replacing said particular alphanumeric character with said preferred alphanumeric character in said temporary buffer.

16. The machine readable storage according to claim 13, wherein said accepting step further comprises the step of:

prior to accepting audibly spelled out replacement text, playing a pre-stored set of instructions for providing said spelled out replacement text.