Biometric voice authentication
View Patent ↗A system and method enrolls a speaker with an enrollment utterance and authenticates a user with a biometric analysis of an authentication utterance, without the need for a PIN (Personal Identification Number). During authentication, the system uses the same authentication utterance to identify who a speaker claims to be with speaker recognition, and verify whether is the speaker is actually the claimed person. Thus, it is not necessary for the speaker to identify biometric data using a PIN. The biometric analysis includes a neural tree network to determine unique aspects of the authentication utterances for comparison to the enrollment authentication. The biometric analysis leverages a statistical analysis using Hidden Markov Models to before authorizing the speaker.
1. A voice authentication system to authenticate a speaker from an authorization utterance, comprising:
a voice identification module to determine an identification of the speaker based on a non-biometric analysis of the authorization utterance; and
a voice verification module coupled to the voice identification module, the voice verification module receiving the identification from the voice identification module, the voice verification module verifying the identification based on a biometric analysis of the same authorization utterance against an enrollment utterance associated with the identification, the enrollment utterance stored in the voice authentication system during enrollment of the speaker in the voice authentication system.
2. The system of claim 1 , wherein the biometric analysis is independent of a PIN (Personal Identification Number).
3. The system of claim 1 , wherein the biometric analysis comprises a neural network analysis.
4. The system of claim 1 , wherein the voice verification module further comprises a voice extraction module to identify unique aspects of the authentication and enrollment utterances relative to common aspects of a plurality of sample utterances.
5. The system of claim 1 , wherein the voice identification module determines a password based on the enrollment utterance, and stores the identification using the password as an index.
6. The system of claim 1 , wherein the voice verification module converts the authentication and enrollment utterances to voice prints, and performs verification based on the voice prints.
7. The system of claim 1 , wherein the voice verification module further comprises a speaker verification engine to perform a statistical analysis of the authentication utterance against the enrollment utterance.
8. The system of claim 7 , the voice verification module verifies the identification based on the biometric analysis and the statistical analysis.
9. The system of claim 7 , wherein the statistical analysis comprises one from the group containing: Hidden Markov Model, Gaussian Mixed Mode, Dynamic Time Warping analysis.
10. The system of claim 1 , wherein the voice identification module uses the authorization utterance as an index to the identification.
11. The system of claim 1 , wherein the authentication utterance and the enrollment utterance are language-independent.
12. The system of claim 1 , wherein the voice identification module further comprises a speech recognition engine to perform a text-independent analysis on the authorization utterance.
13. The system of claim 1 , further comprising a voice encoding module to encode the authorization utterance based on a communication channel used to transmit the authorization utterance.
14. A method of voice authentication for authenticating a speaker's identity from an authorization utterance, comprising:
determining an identification of the speaker based on a non-biometric analysis of the authorization utterance; and
verifying the identification of the speaker as determined by the non-biometric analysis, the verification of the identification based on a biometric analysis of the same authorization utterance against an enrollment utterance associated with the identification, the enrollment utterance stored during enrollment of the speaker for the voice authentication.
15. The method of claim 14 , wherein the biometric analysis is independent of a PIN (Personal Identification Number).
16. The method of claim 14 , wherein the biometric analysis comprises a neural network analysis.
17. The method of claim 14 , wherein the biometric analysis comprises identifying unique aspects of the authentication and enrollment utterances relative to common aspects of a plurality of sample utterances.
18. The method of claim 14 , further comprising:
prior to determining the identification, generating a password based on the enrollment utterance; and
storing the identification of the speaker of the enrollment utterance using the password as an index.
19. The method of claim 14 , wherein the verifying the identification comprises converting the authentication and enrollment utterances to voice prints, and performs verification based on the voice prints.
20. The method of claim 14 , wherein the verifying the identification further comprises performing a statistical analysis of the authentication utterance against the enrollment utterance.
21. The method of claim 20 , wherein the verifying the identification comprises verifying the identification based on the biometric analysis and the statistical analysis.
22. The method of claim 20 , wherein the statistical analysis comprises one from the group containing: Hidden Markov Model, Gaussian Mixed Mode, Dynamic Time Warping analysis.
23. The method of claim 14 , wherein the verifying the identification further comprises performing a text-independent analysis on the authorization utterance.
24. The method of claim 23 , wherein the performing the text-independent analysis comprises performing the text-independent analysis by translating the authorization utterance to text and finding matching text associated with the enrollment utterance.
25. The method of claim 14 , further comprising:
encoding the authorization utterance based on a communication channel used to transmit the authorization utterance.
26. A computer program product comprising a computer readable storage medium structured to store instructions executable by a processor in a computing device, the instructions, when executed cause the processor to:
determine an identification of the speaker based on a non-biometric analysis of the authorization utterance; and
verify the identification of the speaker as determined by the non-biometric analysis based on a biometric analysis of the same authorization utterance against an enrollment utterance associated with the identification, the enrollment utterance stored during enrollment of the speaker prior to the verification.
27. The computer program product of claim 26 , wherein the biometric analysis is independent of a PIN (Personal Identification Number).
28. The computer program product of claim 26 , wherein the biometric analysis comprises a neural network analysis.
29. The computer program product of claim 26 , wherein the biometric analysis comprises identifying unique aspects of the authentication and enrollment utterances relative to common aspects of a plurality of sample utterances.
30. The computer program product of claim 26 , further comprising:
prior to determining the identification, generating a password based on the enrollment utterance; and
storing the identification of the speaker of the enrollment utterance using the password as an index.
31. The computer program product of claim 26 , wherein the verifying the identification comprises converting the authentication and enrollment utterances to voice prints, and performs verification based on the voice prints.
32. The computer program product of claim 26 , wherein the verifying the identification further comprises performing a statistical analysis of the authentication utterance against the enrollment utterance.
33. The computer program product of claim 32 , wherein the verifying the identification comprises verifying the identification based on the biometric analysis and the statistical analysis.
34. The computer program product of claim 26 , wherein the verifying the identification further comprises performing a text-independent analysis on the authorization utterance.
35. The computer program product of claim 34 , wherein the performing the text-independent analysis comprises performing the text-independent analysis by translating the authorization utterance to text and finding matching text associated with the enrollment utterance.