Method and system for displaying a variable number of alternative words during speech recognition

• US 6,363,347 B1
• Filed: 02/16/1999
• Issued: 03/26/2002
• Est. Priority Date: 10/31/1996
• Status: Expired due to Term

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Rozak, Michael J.
• Primary Examiner(s): Korzuch, William
• Assistant Examiner(s): Lerner, Martin
• Application Number:
  US09/250,991
• Time in Patent Office: 1,134 Days
• Field of Search: 704/235, 704/244, 704/251, 704/252, 704/255, 704/257, 704/270, 704/275, 704/276, 704/243, 345/341, 345/342, 345/347, 345/118
• US Class Current: 704/244
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

A method and system for editing words that have been misrecognized. The system allows a speaker to specify a number of alternative words to be displayed in a correction window by resizing the correction window. The system also displays the words in the correction window in alphabetical order. A preferred system eliminates the possibility, when a misrecognized word is respoken, that the respoken utterance will be again recognized as the same misrecognized word. The system, when operating with a word processor, allows the speaker to specify the amount of speech that is buffered before transferring to the word processor.

Claims

1. A computer-readable medium containing instructions for causing a computer dictation system, operable in a dictation mode and a non-dictation mode, to display a variable number of words in a correction window, by:

displaying the correction window;

resizing the correction window using a pointing device to point to a border of the correction window while dragging the pointing device;

calculating a number of words that can be displayed in the correction window based on the resizing;

displaying the calculated number of words in the correction window;

receiving an editing command; and

performing a correction step based on the editing command, the correction step being different, depending on whether the computer dictation system is in the dictation mode or the non-dictation mode.

2. The computer-readable medium of claim 1 wherein the editing command is an input from a delete key of a keyboard and wherein when in the dictation mode the correction step is to delete a displayed word and when in the non-dictation mode the correction step is to delete a displayed character.

3. The computer-readable medium of claim 1 wherein the editing command is an input from an arrow key of a keyboard and wherein when in the dictation mode the correction step is to move an insertion point over one displayed word and when in the non-dictation mode the correction step is to move the insertion point over one displayed character.

4. The computer-readable medium of claim 1 wherein the editing command is a mouse click and wherein when in the dictation mode the correction step is to select a displayed word and when in the non-dictation mode the correction step is to set an insertion point.

5. The computer-readable medium of claim 1 including selecting the words to be displayed from a list of alternative words for a word being corrected.

6. The computer-readable medium of claim 1 wherein the words are displayed in alphabetical order.

7. A method in a recognition system for training to recognize spoken utterances as words, the method comprising:

receiving from a speaker a spoken utterance;

recognizing the spoken utterance as a recognized word;

indicating the recognized word to the speaker;

receiving from the speaker a correct word that is the correct recognition of the spoken utterance;

in response to receiving the correct word from the speaker, saving information pertaining to the spoken utterance and the correct word;

after receiving the correct word and saving information;

determining whether training is to occur based on predetermined criteria, the predetermined criteria causing training to be performed when a computer on which the recognition is done is idle; and

in response to a determination that training is to occur, training to recognize the spoken utterance as the correct word when the computer is idle.

8. The method of claim 7 wherein the indicating of the recognized word includes displaying the recognized word and wherein the receiving of the correct word is a non-spoken utterance replacement of the recognized word.

9. The method of claim 8 wherein the non-spoken utterance replacement is typing of a correct word.

10. The method of claim 7 wherein the recognizing is continuous speech recognition.

11. The method of claim 7 wherein the recognizing is discrete speech recognition.

12. The method of claim 7 wherein the predetermined criteria cause training to be performed in a background mode to the recognition.

13. The method of claim 7 wherein the predetermined criteria cause training to be deferred until a predetermined plurality of spoken utterances and correct words have been received.

14. A method in a recognition system for training to recognize spoken utterances as words, the method comprising:

receiving from a speaker a spoken utterance;

recognizing the spoken utterance as a recognized word;

indicating the recognized word to the speaker;

receiving from the speaker a correct word that is the correct recognition of the spoken utterance;

in response to receiving the correct word from the speaker, saving information pertaining to the spoken utterance and the correct word;

after receiving the correct word and saving information;

determining whether training is to occur based on predetermined criteria; and

in response to a determination that training is to occur, training to recognize the spoken utterance as the correct word, wherein the predetermined criteria cause training to be deferred until a predetermined plurality of spoken utterances and correct words have been received.