Voiceprint verification method and device

• US 10,276,168 B2
• Filed: 12/12/2016
• Issued: 04/30/2019
• Est. Priority Date: 07/31/2014
• Status: Active Grant

• Current Assignee: Tencent Technology Company Limited (Tencent Holdings Limited)
• Original Assignee: Tencent Technology Company Limited (Tencent Holdings Limited)
• Inventors: Sun, Xiao, Wu, Zongzhuo, Long, Qiang, Li, Ling, Cen, Lifang, Zhang, Xiang, Huang, Runjia, Wang, Hui
• Primary Examiner(s): Lerner, Martin
• Application Number:
  US15/375,673
• Publication Number:
  US 20170092276A1
• Time in Patent Office: 869 Days
• Field of Search: 704233, 704236, 704243, 704244, 704246, 704273
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
• CPC Class Codes: G10L 17/02 Preprocessing operations, e.g. segment selection; Pattern representation or modelling, e.g. based on linear discriminant analysis [LDA] or principal components; Feature selection or extraction
• CPC Class Codes: G10L 17/08 Use of distortion metrics or a particular distance between probe pattern and reference templates
• CPC Class Codes: G10L 17/22 Interactive procedures; Man-machine interfaces
• CPC Class Codes: H04L 63/0861 using biometrical features, e.g. fingerprint, retina-scan cryptographic mechanisms or cryptographic arrangements for entity authentication using biological data H04L9/3231
• CPC Class Codes: H04L 9/3231 Biological data, e.g. fingerprint, voice or retina network architectures or network communication protocols for supporting authentication of entities using biometrical features in a packet data network H04L63/0861

Abstract

Embodiments of the present invention provide a voiceprint verification method and device. The voiceprint verification method comprises receiving verification voice data relating to a verification phrase; generating a verification voiceprint on the basis of said verification voice data; determining whether a similarity value between the verification voiceprint and a reference voiceprint conforms to a preset similarity value; and, if the similarity value between the verification voiceprint and a reference voiceprint conforms to a preset similarity value, then determining there is a match between the verification voiceprint and the reference voiceprint.

Claims

1. A method for voiceprint verification, comprising:

receiving, by a processor, verification voice data associated with a verification corpus;

generating, by the processor, a verification voiceprint based on the verification voice data;

calculating a score of a similarity between the verification voiceprint and a pre-stored reference voiceprint according to a characteristic value of each of at least two characteristics of the verification voiceprint and the reference voiceprint and a weight of each of the at least two characteristics, wherein the at least two characteristics comprise at least two of a wavelength, a frequency, an intensity and a rhythm and correspond to at least two different weights;

determining, by the processor, whether the score of the similarity between the verification voiceprint and the reference voiceprint reaches a preset similarity score;

determining, by the processor, that the verification voiceprint is matched with the reference voiceprint, in a case that the score of the similarity between the verification voiceprint and the reference voiceprint reaches the preset similarity score;

storing the verification voice data and attribute information about the verification voice data, in a case that the verification voiceprint is determined to match with the reference voiceprint, the attribute information including a duration of the verification voice data and the score of the similarity between the verification voiceprint and the reference voiceprint;

selecting apiece of the stored verification voice data based on the attribute information; and

updating the reference voiceprint based on registration voice data and the selected piece of the stored verification voice data, the reference voiceprint being generated based on the registration voice data.

2. The method according to claim 1, wherein

the receiving the verification voice data associated with the verification corpus comprises:

receiving the verification voice data inputted by a user based on the verification corpus which is obtained by a client.

3. The method according to claim 2, wherein in a case that the verification voiceprint is matched with the reference voiceprint, identity authentication for the user is determined to be passed.

4. The method according to claim 1, further comprising:

receiving a service operation request carrying a service type,

wherein different service types correspond to different preset similarity scores, and

wherein the determining whether the score of the similarity between the verification voiceprint and the reference voiceprint reaches the preset similarity score comprises;

determining whether the score of the similarity between the verification voiceprint and the reference voiceprint reaches the preset similarity score corresponding to the service type.

5. The method according to claim 1, further comprising:

obtaining the registration voice data associated with a registration corpus; and

generating the reference voiceprint based on the registration voice data.

6. The method according to claim 5, further comprising:

detecting whether an ambient voice volume meets a condition for obtaining the registration voice data based on an ambient voice volume threshold.

7. The method according to claim 5, wherein

the verification corpus comprises a character string identical with at least some characters in the registration corpus, a character string consisting of at least some characters in the registration corpus in a reverse order, a character string consisting of at least some characters in the registration corpus out of order, or a character string formed by repeating at least one character in the registration corpus.

8. The method according to claim 5, wherein

the verification corpus comprises a character string including at least some characters in the registration corpus in a reverse order.

9. The method according to claim 1, further comprising:

obtaining user interest information;

obtaining a registration corpus related to the user interest information;

obtaining the registration voice data associated with the registration corpus for a preset number of times;

determining whether a score of a similarity among the registration voice data reaches a voiceprint registration similarity score; and

generating the reference voiceprint based on the registration voice data, in a case that the score of the similarity among the registration voice data reaches the voiceprint registration similarity score.

10. The method according to claim 1, further comprising:

receiving a service operation request carrying a service type transmitted by a client;

obtaining user information, and generating an electronic note based on the user information; and

transmitting a verification instruction carrying the verification corpus and the generated electronic note to the client based on the service operation request;

wherein the receiving the verification voice data associated with the verification corpus comprises;

receiving the electronic note submitted by the client, and the verification voice data inputted by a current user operating the client based on the verification corpus, which is obtained by the client; and

wherein the generating the verification voiceprint based on the verification voice data comprises;

determining whether the electronic note submitted by the client is consistent with the generated electronic note; and

generating the verification voiceprint based on the verification voice data, in a case that the electronic note submitted by the client is consistent with the generated electronic note.

11. An apparatus for voiceprint verification, comprising:

a memory; and

one or more processors,

wherein program instructions are stored in the memory, and the program instructions, when executed by the one or more processors, cause the apparatus to perform the following operations;

receiving verification voice data associated with a verification corpus;

generating a verification voiceprint based on the verification voice data;

calculating a score of a similarity between the verification voiceprint and a pre-stored reference voiceprint according to a characteristic value of each of at least two characteristics of the verification voiceprint and the reference voiceprint and a weight of each of the at least two characteristics, wherein the at least two characteristics comprise at least two of a wavelength, a frequency, an intensity and a rhythm and correspond to at least two different weights;

determining whether the score of the similarity between the verification voiceprint and the reference voiceprint reaches a preset similarity score;

determining that the verification voiceprint is matched with the reference voiceprint, in a case that the score of the similarity between the verification voiceprint and the reference voiceprint reaches the preset similarity score;

storing the verification voice data and attribute information about the verification voice data, in a case that the verification voiceprint is determined to match with the reference voiceprint, the attribute information including a duration of the verification voice data and the score of the similarity between the verification voiceprint and the reference voiceprint;

selecting a piece of the stored verification voice data based on the attribute information; and

updating the reference voiceprint based on registration voice data and the selected piece of the stored verification voice data, the reference voiceprint being generated based on the registration voice data.

12. The apparatus according to claim 11, wherein

the receiving the verification voice data associated with the verification corpus comprises:

receiving the verification voice data inputted by a user based on the verification corpus which is obtained by a client.

13. The apparatus according to claim 12, wherein in a case that the verification voiceprint is matched with the reference voiceprint, identity authentication for the user is determined to be passed.

14. The apparatus according to claim 11, wherein the operations further comprise receiving a service operation request carrying a service type,

wherein different service types correspond to different preset similarity scores, and

wherein the determining whether the score of the similarity between the verification voiceprint and the reference voiceprint reaches the preset similarity score comprises:

determining whether the score of the similarity between the verification voiceprint and the reference voiceprint reaches the preset similarity score corresponding to the service type.

15. The apparatus according to claim 11, wherein the operations further comprise:

obtaining the registration voice data associated with a registration corpus; and

generating the reference voiceprint based on the registration voice data.

16. The apparatus according to claim 15, wherein

the verification corpus comprises a character string including at least some characters in the registration corpus in a reverse order.

17. The apparatus according to claim 15, wherein the operations further comprise:

detecting whether an ambient voice volume meets a condition for obtaining the registration voice data based on an ambient voice volume threshold.

18. The apparatus according to claim 15, wherein

the verification corpus comprises a character string identical with at least some characters in the registration corpus, a character string consisting of at least some characters in the registration corpus in a reverse order, a character string consisting of at least some characters in the registration corpus out of order, or a character string formed by repeating at least one character in the registration corpus.

19. The apparatus according to claim 11, wherein the operations further comprise:

obtaining user interest information;

obtaining a registration corpus related to the user interest information;

obtaining the registration voice data associated with the registration corpus for a preset number of times;

determining whether a score of a similarity among the registration voice data reaches a voiceprint registration similarity score; and

generating the reference voiceprint based on the registration voice data, in a case that the score of the similarity among the registration voice data reaches the voiceprint registration similarity score.

20. The apparatus according to claim 11, wherein the operations further comprise:

receiving a service operation request carrying a service type transmitted by a client;

obtaining user information, and generating an electronic note based on the user information; and

transmitting a verification instruction carrying the verification corpus and the generated electronic note to the client based on the service operation request;

wherein the receiving the verification voice data associated with the verification corpus comprises;

receiving the electronic note submitted by the client, and the verification voice data inputted by a current user operating the client based on the verification corpus, which is obtained by the client; and

wherein the generating the verification voiceprint based on the verification voice data comprises;

determining whether the electronic note submitted by the client is consistent with the generated electronic note; and

generating the verification voiceprint based on the verification voice data, in a case that the electronic note submitted by the client is consistent with the generated electronic note.