Portable acoustic interface for remote access to automatic speech/speaker recognition server

• US 6,615,171 B1
• Filed: 08/13/1999
• Issued: 09/02/2003
• Est. Priority Date: 06/11/1997
• Status: Expired due to Term

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: International Business Machines Corporation
• Inventors: Poon, Peter S, Prochilo, Carl, Kanevsky, Dimitri, Maes, Stephane Herman
• Primary Examiner(s): Banks-Harold, Marsha D.
• Assistant Examiner(s): Storm, Donald L.
• Application Number:
  US09/776,939
• Time in Patent Office: 1,481 Days
• Field of Search: 704/200, 704/246-250, 704/270, 704/273, 704/275, 704/231, 704/251, 704/270.1
• US Class Current: 704/246
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications

Abstract

A portable speech signal preprocessing (SSP) device having a microphone for receiving spoken speech and background noise, a digital signal processor (DSP) for processing the received noise into feature vectors, a coupler for coupling to a communication device and for transmission over a communication channel. An automatic speech/speaker recognition (ASSR) server receives over the communication channel the preprocessed speech data and recognizes the spoken speech/speaker. A system having the portable SSP device and the ASSR server can be used to remotely activate, reset, or change PIN codes in smartcards, magnetic cards, or electronic money cards.

Claims

1. A method for processing acoustic data for client/server communication, comprising the steps of:

receiving as input by a client, acoustic signals comprising speech, silence and background noise;

preprocessing said acoustic signals by said client, to generate feature vectors representing said acoustic signals, and sending said feature vectors to an ASSR (automatic speech/speaker recognition) server over a communication channel;

communicating between said client and said ASSR server to determine a channel signature of the communication channel using feature vectors representing said background noise; and

performing speaker identification and verification by said ASSR server using said channel signature and said feature vectors representing said speech, to verify that a user of said client is authorized to access a system or service.

2. The method of claim 1, wherein the communication channel comprises a telephone line.

3. The method of claim 1, wherein the step of communicating between said client and said ASSR server to determine a channel signature, comprises the steps of:

generating characterization signals by the client; and

processing said characterization signals by the ASSR server.

4. The method of claim 1, further comprising the step of deconvolve the features vectors representing said speech using said channel signature.

5. The method of claim 1, further comprising the step of using feature vectors representing said silence to determine the channel signature.

6. The method of claim 1, further comprising the step of encrypting said feature vectors by said client before sending said feature vectors to said ASSR server.

7. The method of claim 1, further comprising the step of compressing said feature vectors by said client before sending said feature vectors to said ASSR server.

8. The method of claim 1, further comprising the step of noise masking the feature vectors by the client before sending said feature vectors to the ASSR server.

9. The method of claim 1, further comprising the step of storing data by the client, wherein the stored data comprises one of a client password, a user ID (identification), a serial number, a PIN (personal identification number), a key, a cookie, a digital certificate, a certificate of authenticity, and combinations thereof, provided by the ASSR server.

10. The method of claim 1, further comprising the step of resetting one of a user and PIN (personal identification number) or both, if said user of said client is verified.

11. The method of claim 1, further comprising the step of changing a smartcard PIN (personal identification number) by forwarding changed smartcard data to said client from said ASSR server.

12. The method of claim 11, wherein said changed smartcard data is encrypted by server and wherein said client decrypts said changed smartcard data for changing card PIN.

13. The method of claim 1, further comprising the step of changing a PIN (personal identification number) of a magnetic card by forwarding changed PIN data to a magnetic card writer and said client from said ASSR server.

14. One or more program storage devices readable by one or more machines, tangibly embodying a program of instructions executable by the one or more machines, to perform the method steps of claim 1.

15. A method for processing acoustic data for client/server communication, comprising the steps of:

storing a plurality of models of user enrollment data in a memory of an automatic speech/speaker recognition (ASSR) server;

receiving as input by a portable client, acoustic signals comprising user speech;

preprocessing said acoustic signals by said portable client, to generate feature vectors representing said acoustic signals, and sending said feature vectors to said ASSR server over a communication channel;

performing speaker identification and verification by said ASSR server using said stored models and said feature vectors representing said user speech, to verify that said user of said client is authorized to access a system or service; and

resetting one of a password and PIN (personal identification number) or both, if said user is verified, wherein the step of resetting a PIN comprises changing a smartcard PIN by forwarding changed smartcard data to said client from said ASSR server.

16. The method of claim 15, wherein the communication channel comprises a telephone line.

17. The method of claim 15, further comprising the step of communicating between said client and said ASSR server to determine a channel signature of the communication channel using feature vectors representing background noise, wherein the channel signature is used for performing speaker identification and verification by said ASSR server.

18. The method of claim 17, wherein the step of communicating between said client and said ASSR server to determine a channel signature, comprises the steps of:

generating characterization signals by the client; and

processing said characterization signals by the ASSR server.

19. The method of claim 17, further comprising the step of using feature vectors representing silence to determine the channel signature.

20. The method of claim 15, further comprising the step of encrypting said feature vectors by said client before sending said feature vectors to said ASSR server.

21. The method of claim 15, further comprising the step of noise masking the feature vectors by the client before sending said feature vectors to the ASSR server.

22. The method of claim 15, wherein said changed smartcard data is encrypted by said ASSR server and wherein said client decrypts said changed smartcard data for changing said smartcard PIN.

23. The method of claim 15, further comprising the step of compressing said feature vectors before sending said feature vectors to said ASSR server.

24. One or more program storage devices readable by one or more machines, tangibly embodying a program of instructions executable by the one or more machines, to perform the method steps of claim 15.

25. A method for processing acoustic data for client/server communication, comprising the steps of:

storing a plurality of models of user enrollment data in a memory of an automatic speech/speaker recognition (ASSR) server;

receiving as input by a portable client, acoustic signals comprising user speech;

preprocessing said acoustic signals by said portable client, to generate feature vectors representing said acoustic signals, and sending said feature vectors to said ASSR server over a communication channel;

performing speaker identification and verification by said ASSR server using said stored models and said feature vectors representing said user speech, to verify that said user of said client is authorized to access a system or service; and

resetting one of a password and PIN (personal identification number) or both, if said user is verified, wherein the step of resetting a PIN comprises changing a magnetic card PIN by forwarding changed PIN data to a magnetic card writer and said client from said ASSR server.