Architecture for DSR client and server development platform

• US 7,062,444 B2
• Filed: 01/24/2002
• Issued: 06/13/2006
• Est. Priority Date: 01/24/2002
• Status: Expired due to Fees

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: He, Liang, Xie, ChuanQuan, Cui, Song
• Primary Examiner(s): ABEBE, DANIEL DEMELASH
• Application Number:
  US10/057,705
• Publication Number:
  US 20030139930A1
• Time in Patent Office: 1,601 Days
• Field of Search: 704/270, 704/275, 379/88.03, 345/727
• US Class Current: 704/275
• 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

A comprehensive, practical and efficient architecture for a DSR client and server development platform is disclosed. A system comprising: DSR client module for capturing speech, extracting speech features, interpreting markup content and displaying content; DSR gateway module coupled for communication with the DSR client module, the DSR gateway module for receiving a markup document, interpreting tag elements of the markup document, and dynamically generating grammar from the markup document, and controlling display content navigation by speech recognition; and DSR document server coupled for communication with the DSR gateway module, the DSR document server for processing requests from the DSR gateway module and for producing a markup document in response.

Claims

1. A system comprising:

DSR client module for capturing speech, extracting speech features, sending the speech features, interpreting markup content and displaying content;

DSR gateway module coupled for communication with the DSR client module, the DSR gateway module for receiving the speech features and a DSR markup document, interpreting tag elements of the DSR markup document, and dynamically generating grammar from the DSR markup document, and controlling display content navigation of the DSR markup document by distributed speech recognition, wherein the DSR client module sends speech features to the DSR gateway module via a network; and

DSR document server coupled for communication with the DSR gateway module, the DSR document server for processing requests from the DSR gateway module and for producing the markup document in response.

2. The system of claim 1, wherein said DSR client module includes at least one of an audio capturer, a feature extractor, a data wrapper and an interpreter, each of which performs a part of the function of the DSR client module.

3. The system of claim 2, wherein said audio capturer records speech, samples speech signals, and performs voice active detection or end-point detection.

4. The system of claim 2, wherein said feature extractor performs MFCC and vector quantization.

5. The system of claim 2, wherein said data wrapper performs processing of a connection request, processes events for synchronization, receives display content, and transmits speech feature data.

6. The system of claim 2, wherein said interpreter allocates tasks to the audio capturer, speech feature extractor, and data wrapper, and interprets markup document content.

7. The system of claim 1, wherein said DSR gateway module includes at least one of a server browser, a DNS server, and a utility platform, each of which performs a part of the function of DSR gateway module.

8. The system of claim 7, wherein said server browser includes at least one of a data wrapper, an HTTP wrapper, and an interpreter, each of which performs a part of the function of said server browser.

9. The system of claim 8, wherein said data wrapper parses the markup document, determines the syntax of the markup document, and consolidates the information for the server browser, and assigns tasks to the utility platform.

10. The system of claim 8, wherein said HTTP wrapper processes a markup document request and document transmission.

11. The system of claim 7, wherein said utility platform is controlled by said server browser and performs various system jobs.

12. The system of claim 11, wherein said system jobs include speech recognition, TTS conversion, dynamic grammar building, display content generation, and transmission and workload balance control.

13. The system of claim 1, wherein events are used for synchronization between said DSR client module and said DSR gateway module.

14. The system of claim 13, wherein said events include system synchronization events and content synchronization events.

15. A method comprising:

a DSR client performing front-end speech processing;

the DSR client sending speech feature data to a DSR gateway via a network;

the DSR client receiving via the network a first event from the DSR gateway with identification information for a first component if a distributed speech recognition result performed by the DSR gateway indicates that the DSR client is to display the first component of a current DSRML document;

the DSR client displaying the first component of the current document;

the DSR client receiving via the network a second event from the DSR gateway if the distributed speech recognition result is non-decipherable; and

the DSR client receiving via the network the display content from the DSR gateway.

16. The method of claim 15 further comprising the DSR client adjusting transmission control conditions according to transmission parameters.

17. A method comprising:

a DSR gateway receiving speech feature data from a DSR client via a network;

the DSR gateway performing distributed speech recognition of the speech feature data from the DSR client;

the DSR gateway sending via the network a first event with identification information for a first component to the DSR client if a distributed speech recognition result indicates that the DSR client is to display the first component of the current DSRML document;

the DSR gateway sending via the network a second event to the DSR client if the distributed speech recognition result is non-decipherable; and

the DSR gateway sending via the network a DSRML request to a DSR document server if the distributed speech recognition result indicates that the DSR client needs a new document.

18. The method of claim 17 further comprising the DSR gateway adjusting transmission control conditions according to transmission parameters.

19. The method of claim 18, wherein adjusting transmission control conditions according to transmission parameters includes adjusting bandwidth.

20. The method of claim 17 further comprising:

the DSR gateway parsing the DSRML document;

the DSR gateway compiling all grammars that for speech recognition engine;

the DSR gateway generating display content for the DSR client; and

the DSR gateway sending via the network the display content to the DSR client, wherein sending the display content includes adjusting transmission control conditions according to transmission requirements.

21. A machine readable medium having stored thereon executable code which causes a machine to perform the method comprising:

performing front-end speech processing;

sending speech feature data to a DSR gateway via a network;

receiving via the network a first event from the DSR gateway with identification information for a first component if a distributed speech recognition result indicates that the DSR client is to display the first component of the current DSRML document;

displaying the first component of the current document;

receiving via the network a second event from the DSR gateway if the distributed speech recognition result is non-decipherable; and

receiving the display content from the DSR gateway.

22. The machine readable medium of claim 21, wherein the executable code further includes the:

adjusting transmission control conditions according to transmission parameters.

23. A machine readable medium having stored thereon executable code which causes a machine to perform the method comprising:

receiving speech feature data from a DSR client via a network;

performing distributed speech recognition of the speech feature data from the DSR client;

sending via the network a first event with identification information for a first component to the DSR client if a distributed speech recognition result indicates that the DSR client is to display the first component of the current DSRML document;

sending via the network a second event to the DSR client if the distributed speech recognition result is non-decipherable; and

sending via the network a DSRML request to a DSR document server if the distributed speech recognition result indicates that the DSR client needs a new document.

24. The machine readable medium of claim 23, wherein the executable code further includes:

adjusting transmission control conditions according to transmission parameters.

25. The machine readable medium of claim 24, wherein adjusting transmission control conditions according to transmission parameters includes adjusting bandwidth.

26. The machine readable medium of claim 23, wherein the executable code further includes:

parsing the DSRML document;

compiling all grammars that for speech recognition engine;

generating display content for the DSR client; and

sending via the network the display content to the DSR client, wherein sending the display content includes adjusting transmission control conditions according to transmission requirements.