Method and system for correcting misrecognized spoken words or phrases

• US 5,829,000 A
• Filed: 10/31/1996
• Issued: 10/27/1998
• Est. Priority Date: 10/31/1996
• Status: Expired due to Term

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Jiang, Li, Huang, Xuedong D., Hon, Hsiao-Wuen
• Primary Examiner(s): Hudspeth, David R.
• Assistant Examiner(s): Storm, Donald L.
• Application Number:
  US08/741,696
• Time in Patent Office: 726 Days
• Field of Search: 704/244, 704/252, 704/251, 704/255, 704/231, 704/270, 704/200, 704/275, 704/239, 704/240
• US Class Current: 704/252
• 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. This elimination occurs based on the probabilities of alternative words associated with both the misrecognized utterance and the respoken utterance. 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. The system also uses a word correction metaphor or a phrase correction metaphor.

Claims

1. A method in a computer dictation system for recognizing a word, the method comprising:

receiving a first spoken utterance;

identifying first alternative words each having a probability that the first alternative word is a correct recognition of the first spoken utterance;

recognizing the first spoken utterance as one of the first alternative words wherein the recognized first alternative word is a misrecognition of the first spoken utterance;

receiving a second spoken utterance representing a correction of the recognized first alternative word;

identifying second alternative words each having a probability that the second alternative word is a correct recognition of the second spoken utterance; and

recognizing the second spoken utterance as one of the second alternative words other than the recognized first alternative word based on the combined probabilities of the first alternative words and the second alternative words other than the recognized first alternative word.

2. The method of claim 1 wherein the recognized second alternative word is the second alternative word with the highest probability other than the recognized first alternative word.

3. The method of claim 1 wherein the recognized first alternative word is the first alternative word with the highest probability.

4. A method in a computer dictation system for recognizing a word, the method comprising:

receiving a first spoken utterance;

identifying first alternative words each having a probability that the first alternative word is a correct recognition of the first spoken utterance;

recognizing the first spoken utterance as one of the first alternative words wherein the recognized first alternative word is a misrecognition of the first spoken utterance;

receiving a second spoken utterance representing a correction of the recognized first alternative word;

identifying second alternative words each having a probability that the second alternative word is a correct recognition of the second spoken utterance; and

recognizing the second spoken utterance as one of the second alternative words other than the recognized first alternative word based on the combined probabilities of the first alternative words and the second alternative words wherein the combined probability of a word is the probability of that word as a first alternative word times the probability of that word as a second alternative word.

5. A computer dictation system for recognizing a spoken utterance, comprising:

a recognition component that receives a spoken utterance and provides alternative words with probabilities, the probabilities indicating that the alternative word is a correct recognition of the spoken utterance; and

a dictation editing component that

receives a newly spoken utterance representing a correction of a misrecognized word for a previously spoken utterance, the previously spoken utterance having previous alternative words with probabilities,

sends the newly spoken utterance to the recognition component and is provided with new alternative words, each alternative word having a probability that the alternative word is a correct recognition of the newly spoken utterance, and

recognizes the newly spoken utterance as a word other than the misrecognized word based on the combined probabilities of the words in the previous alternative words and the new alternative words other than the misrecognized word.

6. A computer dictation system for recognizing a spoken utterance, comprising:

a recognition component that receives a spoken utterance and provides alternative words with probabilities, the probabilities indicating that the alternative word is a correct recognition of the spoken utterance; and

a dictation editing component that

receives a newly spoken utterance representing a correction of a misrecognized word for a previously spoken utterance, the previously spoken utterance having previous alternative words with probabilities,

sends the newly spoken utterance to the recognition component and is provided with new alternative words, each alternative word having a probability that the alternative word is a correct recognition of the newly spoken utterance, and

recognizes the newly spoken utterance as a word other than the misrecognized word based on the combined probabilities of the words in the previous alternative words and the new alternative words wherein the combined probability of a word is the probability of that word as a previous alternative word times the probability of that word as a new alternative word.

7. A method in a computer dictation system for correcting a misrecognized word, the method comprising:

receiving a newly spoken utterance representing a correction of the misrecognized word for a previously spoken utterance, the previously spoken utterance having previous alternative words with probabilities;

identifying new alternative words with probabilities for the newly spoken utterance;

recognizing the newly spoken utterance as a newly recognized word other than the misrecognized word based on the combined probabilities of the words other than the misrecognized word in the previous alternative words and the new alternative words; and

replacing the misrecognized word with the newly recognized word.

8. The method of claim 7 wherein the recognized word is the word with the highest probability other than the misrecognized word.

9. A computer-readable medium containing instructions in a computer dictation system to correct a misrecognized word, by:

receiving a newly spoken utterance representing a correction of the misrecognized word for a previously spoken utterance;

selecting a newly recognized word other than the misrecognized word using a combination of probabilities associated with alternative words other than the misrecognized word for the previously spoken utterance and the newly spoken utterance; and

replacing the misrecognized word with the newly recognized word.

10. The computer-readable medium of claim 9 wherein the dictation system has a previous list of words with associated probabilities of matching the previously spoken utterance that resulted in the misrecognized word, wherein the recognizing of the newly spoken utterance generates a new list of words with associated probabilities of matching the newly spoken utterance, and wherein the recognized word is the word with the highest combined probability in the previous and new list other than the misrecognized word.

11. A method in a computer dictation system for speech recognition, the method comprising:

receiving a first spoken utterance;

identifying first alternatives, each first alternative having a probability that the first alternative is a correct recognition of the first spoken utterance;

recognizing the first spoken utterance as one of the first alternatives wherein the recognized first alternative is a misrecognition of the first spoken utterance;

receiving a second spoken utterance representing a correction of the recognized first alternative;

identifying second alternatives, each second alternative having a probability that the second alternative is a correct recognition of the second spoken utterance; and

recognizing the second spoken utterance as one of the second alternatives other than the recognized first alternative based on the combined probabilities of the first alternatives and the second alternatives that are other than the recognized first alternative.

12. The method of claim 11 wherein the alternatives are phrases of one or more words.

13. The method of claim 11 wherein the recognized second alternative is the second alternative with the highest probability other than the recognized first alternative.

14. The method of claim 11 wherein the recognized first alternative is the first alternative with the highest probability.

15. A method in a computer dictation system for speech recognition, the method comprising:

receiving a first spoken utterance;

identifying first alternatives, each first alternative having a probability that the first alternative is a correct recognition of the first spoken utterance;

recognizing the first spoken utterance as one of the first alternatives wherein the recognized first alternative is a misrecognition of the first spoken utterance;

receiving a second spoken utterance representing a correction of the recognized first alternative;

identifying second alternatives, each second alternative having a probability that the second alternative is a correct recognition of the second spoken utterance; and

recognizing the second spoken utterance as one of the second alternatives other than the recognized first alternative based on the combined probabilities of the first alternatives and the second alternatives wherein the combined probability of one of the alternatives is the probability of that alternative as a first alternative times the probability of that alternative as a second alternative.