System, method and computer program product for large-scale street name speech recognition

• US 20020133344A1
• Filed: 01/24/2001
• Published: 09/19/2002
• Est. Priority Date: 01/24/2001
• Status: Abandoned Application

• Current Assignee: Bevocal Incorporated (Microsoft Corporation)
• Original Assignee: Bevocal Incorporated (Microsoft Corporation)
• Inventors: Damiba, Bertrand A.
• Application Number:
  US09/770,750
• Publication Number:
  US 20020133344A1
• Time in Patent Office: Days
• US Class Current: 704/247
• CPC Class Codes: G10L 15/1815 Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning

Abstract

A system, method and computer program product are provided for speech recognition. During operation, a database of words are maintained. Initially, a probability is assigned to each of the words which indicates a prevalency of use of the word. Further, an utterance is received for speech recognition purposes. Such utterance is matched with one of the words in the database based on least in part on the probability.

Claims

1. A speech recognition method, comprising the steps of:

(a) maintaining a database of words;

(b) assigning a probability to each of the words which indicates a prevalency of use of the word;

(c) receiving an utterance for speech recognition purposes; and

(d) matching the utterance with one of the words in the database based on least in part on the probability.

2. The method as recited in claim 1, wherein the words include names of streets.

3. The method as recited in claim 2, wherein the probability is determined using statistical data corresponding to use of the streets.

4. The method as recited in claim 3, wherein the statistical data includes traffic counts.

5. The method as recited in claim 3, wherein the statistical data includes traffic along the streets and along intersecting streets.

6. The method as recited in claim 1, wherein the speech recognition is carried out over a network.

7. A computer program product, comprising:

(a) computer code for maintaining a database of words;

(b) computer code for assigning a probability to each of the words which indicates a prevalency of use of the word;

(c) computer code for receiving an utterance for speech recognition purposes; and

(d) computer code for matching the utterance with one of the words in the database based on least in part on the probability.

8. The computer program product as recited in claim 7, wherein the words include names of streets.

9. The computer program product as recited in claim 8, wherein the probability is determined using statistical data corresponding to use of the streets.

10. The computer program product as recited in claim 9, wherein the statistical data includes traffic counts.

11. The computer program product as recited in claim 9, wherein the statistical data includes traffic along the streets and along intersecting streets.

12. The computer program product as recited in claim 7, wherein the speech recognition is carried out over a network.

14. The system as recited in claim 13, wherein the words include names of streets.

15. The system as recited in claim 14, wherein the probability is determined using statistical data corresponding to use of the streets.

16. The system as recited in claim 15, wherein the statistical data includes traffic counts.

17. The system as recited in claim 15, wherein the statistical data includes traffic along the streets and along intersecting streets.

18. The system as recited in claim 13, wherein the speech recognition is carried out over a network.

13. A system, comprising:

(a) logic for maintaining a database of words;

(b) logic for assigning a probability to each of the words which indicates a prevalency of use of the word;

(c) logic for receiving an utterance for speech recognition purposes; and

(d) logic for matching the utterance with one of the words in the database based on least in part on the probability.