Granted Patent B2

US 8,032,380 · App. 12/029,952 · Granted Oct 4, 2011

Abstract

Claims (36)

1. A method of providing access to a service, comprising:

receiving a first speech signal from a user to form a first utterance associated with a personal identification number;

when the personal identification number is valid, prompting, via a processor, the user to speak the personal identification number a second time;

receiving a second speech signal to form a second utterance associated with the personal identification number; and

providing access to the service if a comparison between the first utterance and the second utterance is greater than a threshold that is associated with a similarity between the first utterance and the second utterance.

2. A method as defined in claim 1 , further comprising generating a reference utterance based on the first utterance and the second utterance when the first utterance and the second utterance are sufficiently similar.

3. A method as defined in claim 2 , wherein the reference utterance is a generalization utterance of the first utterance and the second utterance.

4. A method as defined in claim 1 , further comprising storing a first reference utterance based on the first utterance and storing a second reference utterance based on the second utterance.

5. A method as defined in claim 1 , further comprising verifying the user based on the first utterance using a speaker verification system when the user has used the service previously.

6. A method as defined in claim 5 , wherein verifying the user based on the first utterance comprises determining that the user is a male by detecting a preselected digit at a predetermined position of the first utterance.

7. A method as defined in claim 6 , further comprising determining a similarity between a reference utterance and the first utterance using a male variance vector and calculating a weighted Euclidean distance between the reference utterance and the first utterance.

8. A method as defined in claim 6 , further comprising:

determining that the user is a female when the preselected digit is not detected at the predetermined position of the first utterance;

determining a similarity between a reference utterance and the first utterance using a female variance vector; and

calculating a weighted Euclidean distance between the reference utterance and the first utterance.

9. A system of providing access to a service, comprising:

a processor;

a first module configured to control the processor to receive a first speech signal from a user to form a first utterance associated with personal identification data;

a second module configured to control the processor to attempt to verify the user via a speaker verification system using the first utterance;

a third module configured to control the processor, when the user is verified using the speaker verification system, to receive a second speech signal to form a second utterance associated with the personal identification data; and

a fourth module configured to control the processor to provide access to the service based on whether a comparison between the first utterance and the second utterance is greater than a threshold that is associated with a similarity between the first utterance and the second utterance.

10. A system as defined in claim 9 , further comprising a fifth module to control the processor to determine a similarity between the first utterance and the second utterance.

11. A system as defined in claim 10 , further comprising a sixth module configured to control the processor to request the user to speak a third utterance associated with the personal identification data when the first and second utterances are insufficiently similar.

12. A system as defined in claim 11 , further comprising a seventh module configured to control the processor to determine a similarity between the first utterance and the third utterance and generating a reference utterance based on the first and third utterances when the similarity meets a threshold.

13. A system as defined in claim 11 , further comprising a seventh module configured to control the processor to generate a first reference utterance based on the first utterance, a second reference utterance based on the second utterance, and a third reference utterance based on the third utterance.

14. A system as defined in claim 9 , wherein two possible decisions of attempting to verify the user are verified and not verified.

15. A method of providing access to a service, comprising:

receiving a first speech signal from a user to form a first utterance associated with personal identification data;

when the user has not previously accessed the service, receiving a second speech signal from the user to form a second utterance associated with the personal identification data;

storing the first and second utterances as reference utterances; and

providing access to the service based on the personal identification data and if a comparison between a first utterance and a second utterance is greater than a threshold that is associated with a similarity between the first utterance and the second utterance.

16. A method as defined in claim 15 , further comprising verifying the user based on the first utterance using a speaker verification system when the user has previously accessed the service.

17. A method as defined in claim 15 , further comprising allowing the user to access the service when the user is verified.

18. A method as defined in claim 15 , wherein the service is a dial-up service.

19. A method as defined in claim 15 , further comprising connecting to the user in response to the user having initiated communication with the service.

20. A method as defined in claim 15 , wherein the first utterance is representative of a plurality of digits, the method further comprising recognizing the first utterance using speaker independent speaker recognition by recognizing a plurality of digits using speaker independent speaker recognition.

Assignments (8)