SPEECH RECOGNITION SYSTEM AND DATA UPDATING METHOD

• US 20100185446A1
• Filed: 01/20/2010
• Published: 07/22/2010
• Est. Priority Date: 01/21/2009
• Status: Active Grant

• Current Assignee: Clarion Co., Ltd. (Faurecia SA)
• Original Assignee: Clarion Co., Ltd. (Faurecia SA)
• Inventors: Takahashi, Hisashi, Asahara, Akinori, Kokubo, Hiroaki, HOMMA, Takeshi
• Granted Patent: US 8,538,759 B2
• Time in Patent Office: Days
• US Class Current: 704/251
• CPC Class Codes: G01C 21/3608 using speech input, e.g. using speech recognition
• CPC Class Codes: G01C 21/3859 Differential updating map data
• CPC Class Codes: G09B 19/04 Speaking with audible presentation of the material to be studied G09B5/04
• CPC Class Codes: G09B 19/06 Foreign languages with audible presentation of material to be studied G09B5/04
• CPC Class Codes: G09B 29/106 using electronic means
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
• CPC Class Codes: G10L 2015/228 of application context

Abstract

It is provided a speech recognition system installed in a terminal coupled to a server via a network. The terminal holds map data including a landmark. The speech recognition system manages recognition data including a word corresponding to a name of the landmark, and sends update area information and updated time to the server. The server generates, when recognition data of the area of the update area information sent from the terminal has been changed after updated time, difference data between latest recognition data and recognition data of the update area information at a time of the updated time, and sends the generated difference data and map data of the update area information to the terminal. The terminal updates the map data based on the map data sent from the server. The speech recognition system updates the recognition data managed by the terminal based on the difference data.

Claims

1. A speech recognition system installed in a terminal coupled to a server via a network, wherein:

the terminal holds map data including a landmark;

the speech recognition system manages recognition data including a word corresponding to a name of the landmark included in the held map data, and sends update area information indicating an area of the map data to be updated and update data on the area indicated by the update area information to the server;

the server is configured to;

generate, in a case of which recognition data of the area indicated by the update area information sent from the terminal has been changed after a time indicated by the update data sent from the terminal, difference data between latest recognition data and recognition data corresponding to the area indicated by the update area information at a time indicated by the update data; and

send the generated difference data to the terminal along with map data on the area indicated by the update area information;

the terminal updates the map data held in the terminal based on the map data sent from the server; and

the speech recognition system updates the recognition data managed by the terminal based on the difference data sent from the server.

2. The speech recognition system according to claim 1, wherein:

the recognition data includes confusion information including, for each word, a confusion word having a tendency to cause a recognition error with the each word, and a confusion score which represents the tendency of the confusion word to cause recognition error;

the server sends the difference data including the confusion information to the terminal; and

the speech recognition system updates the confusion information included in the recognition data held by the terminal based on the confusion information sent from the server.

3. The speech recognition system according to claim 2, wherein, in a case of which a word included in the confusion information sent from the server is included in confusion information for a different area, the speech recognition system is configured to update confusion information for the different area which is corresponding to the word.

4. The speech recognition system according to claim 2, which is configured to:

refer to the confusion information based on a recognition result of a speech uttered by a user; and

correct the recognition result based on the referred confusion information.

5. The speech recognition system according to claim 4, which is configured to:

refer to confusion information based on a recognition result of a speech uttered by a user;

add, in a case of which the referred confusion information includes a word having a tendency to cause a recognition error, the word having a tendency to cause a recognition error to the recognition result, and

determine a score of the word having a tendency to cause a recognition error based on the confusion score.

6. The speech recognition system according to claim 5, which is configured to change the recognition result to be output based on the determined score of the word.

7. The speech recognition system according to claim 2, wherein:

the server extracts confusion information on a word having a tendency to cause a recognition error with a word included in the difference data, and sends the extracted confusion information as confusion subject information to the terminal; and

the speech recognition system updates the confusion information managed by the terminal based on the confusion subject information sent from the server.

8. The speech recognition system according to claim 2, wherein the speech recognition system is configured to:

extract confusion information on a word having a tendency to cause a recognition error with a word included in the confusion information sent from the server; and

set the confusion score of the extracted confusion information for confusion information on a case of which the word having a tendency to cause a recognition error is to be recognized correctly.

9. A data updating method for a terminal coupled to a server via a network,

the terminal including a speech recognition system and holding map data including a landmark,

the speech recognition system managing recognition data including a word corresponding to a name of the landmark included in the held map data, the data update method comprising the steps of:

sending, by the terminal, update area information indicating an area of the map data to be updated, and update data on the area to be updated to the server;

generating, by the server, in a case of which recognition data of the area indicated by the update area information sent from the terminal has been changed after a time point indicated by the update data sent from the terminal, difference data between latest recognition data and recognition data corresponding to the area indicated by the update area information at a time indicated by the update data;

sending, by the server, the generated difference data to the terminal along with map data on the area indicated by the update area information;

updating, by the terminal, the map data held in the terminal based on the map data sent from the server; and

updating, by the terminal, the managed recognition data based on the difference data sent from the server.

10. The data updating method according to claim 9, wherein:

the recognition data includes confusion information including, for each word, a confusion word having a tendency to cause a recognition error with the each word, and a confusion score which represents the tendency of the confusion word to cause recognition error;

the server sends the difference data including the confusion information to the terminal; and

the method further comprises the step of updating, by the terminal, the confusion information included in the recognition data held by the terminal based on the confusion information sent from the server.

11. The data updating method according to claim 10, further comprising the step of updating, by the terminal, in a case of which a word included in the confusion information sent from the server is included in confusion information for a different area, confusion information for the different area corresponding to the word.

12. The data updating method according to claim 10, further comprising the steps of:

referring to, by the terminal, the confusion information based on a recognition result of a speech uttered by a user; and

correcting, by the terminal, the recognition results based on the referred confusion information.

13. The data updating method according to claim 12, further comprising the steps of:

referring to, by the terminal, the confusion information based on a recognition result of a speech uttered by a user;

adding, by the terminal, in a case of which the referred confusion information includes a word having a tendency to cause a recognition error, the word having a tendency to cause a recognition error to the recognition result, and

determining, by the terminal, a score of the word having a tendency to cause a recognition error based on the confusion score.

14. The data updating method according to claim 13, further comprising the step of changing, by the terminal, the recognition result to be output based on the score of the word determined based on the confusion score.

15. The data updating method according to claim 10, further comprising the steps of:

extracting, by the server, confusion information on a word having a tendency to cause a recognition error with a word included in the difference data, and sending the extracted confusion information as confusion subject information to the terminal; and

updating, by the terminal, the managed confusion information based on the confusion subject information sent from the server.

16. The data updating method according to claim 10, further comprising the steps of:

extracting, by the terminal, confusion information on a word having a tendency to cause a recognition error with a word included in the confusion information sent from the server; and

setting, by the terminal, the confusion score of the extracted confusion information to confusion information on a case of which the word having a tendency to cause a recognition error is to be recognized correctly.

17. A speech recognition system, comprising:

speech recognition unit for recognizing a speech by using recognition data;

recognition data storage unit for storing the recognition data; and

recognition data updating unit for updating the recognition data stored in the recognition data storage unit.

18. The speech recognition system according to claim 17, further comprising:

information storage unit for storing data corresponding to the speech to be recognized by the speech recognition unit; and

database updating unit for updating the data stored in the information storage unit,

wherein the recognition data updating unit updates the recognition data relating to the data updated by the database updating unit.

19. The speech recognition system according to claim 17, further comprising receiver for receiving update data from an information center,

wherein the recognition data updating unit updates the recognition data by using the update data received by the receiver.

20. The speech recognition system according to claim 17, further comprising data input unit for inputting update data recorded in a storage medium,

wherein the recognition data updating unit updates the recognition data by using the update data input from the data input means.

21. The speech recognition system according to claim 18, wherein the data stored in the information storage unit is map data used by a navigation device coupled to the speech recognition system.

22. The speech recognition system according to claim 18, wherein:

the recognition data updating unit includes update recognition data selection unit for selecting recognition data which need to be updated form recognition data relating to the data updated by the database updating unit based on an update content of the data; and

the recognition data updating unit updates the recognition data selected by the update recognition data selection unit.