Identity authentication method and apparatus

Patent metadata

Field	Value
Publication number	US-10789343-B2
Application number	US-201816192401-A
Country	US
Kind code	B2
Filing date	Nov 15, 2018
Priority date	May 19, 2016
Publication date	Sep 29, 2020
Grant date	Sep 29, 2020

Abstract

An audio/video stream generated by a target object to be authenticated is obtained. The target object is associated with a user. A determination is made whether a lip reading component and voice component in the audio/video stream are consistent. In response to determining that the lip reading component and voice component are consistent, voice recognition is performed on an audio stream in the audio/video stream to obtain voice content. The voice content is used as an object identifier of the target object. A model physiological feature corresponding to the object identifier is obtained from object registration information. Physiological recognition is performed on the audio/video stream to obtain a physiological feature of the target object. The physiological feature of the target object is compared with the model physiological feature to obtain a comparison result. If the comparison result satisfies an authentication condition, the target object is authenticated.

First claim

What is claimed is:

1. A computer-implemented method, comprising:
obtaining an audio/video stream of a user that is to be authenticated;
determining that the user's voice in the audio/video stream matches the user's lips in the audio/video stream;
in response to determining that the user's voice in the audio/video stream matches the user's lips in the audio/video stream, determining, based on performing automated speech recognition on the audio/video stream, a user identifier for the user;
determining, based on performing automated physiological feature extraction on the audio/video stream, a physiological feature of the user;
obtaining, from stored, object registration information, a stored, model physiological feature corresponding to the determined, user identifier;
generating a comparison result based on comparing the physiological feature of the target object that was determined based on performing automated physiological feature extraction on the audio/video stream with the stored, model physiological feature; and
in response to determining that the comparison result satisfies an authentication condition, determining that the user has been authenticated.

2. The method of claim 1 , wherein the physiological feature of the user comprises a facial feature of the user.

3. The method of claim 1 , wherein the comparison result comprises a similarity score.

4. The method of claim 1 , wherein determining that the comparison result satisfies an authentication condition comprises determining that a similarity score exceeds a threshold score.

5. The method of claim 1 , wherein determining that the user's voice in the audio/video stream matches the user's lips comprises:
determining a lip reading syllable in a video image of the audio/video stream at a particular point in time;
determining a voice syllable in audio of the audio/video stream at the particular point in time; and
determining that the lip reading syllable and the voice syllable match.

6. The method of claim 1 , comprising storing the model physiological feature in the object registration information.

7. The method of claim 1 , comprising receiving a request from the user to authenticate, wherein the audio/video stream of the user is obtained in response to the request.

8. A non-transitory, computer-readable medium storing one or more instructions executable by a computer system to perform operations comprising:
obtaining an audio/video stream of a user that is to be authenticated;
determining that the user's voice in the audio/video stream matches the user's lips in the audio/video stream;
in response to determining that the user's voice in the audio/video stream matchesthe user's lips in the audio/video stream, determining, based on performing automated speech recognition on the audio/video stream, a user identifier for the user;
determining, based on performing automated physiological feature extraction on the audio/video stream, a physiological feature of the user;
obtaining, from stored, object registration information, a stored, model physiological feature corresponding to the determined, user identifier;
generating a comparison result based on comparing the physiological feature of the target object that was determined based on performing automated physiological feature extraction on the audio/video stream with the stored, model physiological feature; and
in response to determining that the comparison result satisfies an authentication condition, determining that the user has been authenticated.

9. The medium of claim 8 , wherein the physiological feature of the user comprises a facial feature of the user.

10. The medium of claim 8 , wherein the comparison result comprises a similarity score.

11. The medium of claim 8 , wherein determining that the comparison result satisfies an authentication condition comprises determining that a similarity score exceeds a threshold score.

12. The medium of claim 8 , wherein determining that the user's voice in the audio/video stream matches the user's lips comprises:
determining a lip reading syllable in a video image of the audio/video stream at a particular point in time;
determining a voice syllable in audio of the audio/video stream at the particular point in time; and
determining that the lip reading syllable and the voice syllable match.

13. The medium of claim 8 , wherein the operations comprise storing the model physiological feature in the object registration information.

14. The medium of claim 8 , wherein the operations comprise receiving a request from the user to authenticate, wherein the audio/video stream of the user is obtained in response to the request.

15. A computer-implemented system, comprising:
one or more computers; and
one or more computer memory devices interoperably coupled with the one or more computers and having tangible, non-transitory, machine-readable media storing one or more instructions that, when executed by the one or more computers, perform one or more operations comprising:
obtaining an audio/video stream of a user that is to be authenticated;
determining that the user's voice in the audio/video stream matches the user's lips in the audio/video stream;
in response to determining that the user's voice in the audio/video stream matches the user's lips in the audio/video stream, determining, based on performing automated speech recognition on the audio/video stream, a user identifier for the user;
determining, based on performing automated physiological feature extraction on the audio/video stream, a physiological feature of the user;
obtaining, from stored, object registration information, a stored, model physiological feature corresponding to the determined, user identifier;
generating a comparison result based on comparing the physiological feature of the target object that was determined based on performing automated physiological feature extraction on the audio/video stream with the stored, model physiological feature; and
in response to determining that the comparison result satisfies an authentication condition, determining that the user has been authenticated.

16. The system of claim 15 , wherein the physiological feature of the user comprises a facial feature of the user.

17. The system of claim 15 , wherein the comparison result comprises a similarity score.

18. The system of claim 15 , wherein determining that the comparison result satisfies an authentication condition comprises determining that a similarity score exceeds a threshold score.

19. The system of claim 15 , wherein determining that the user's voice in the audio/video stream matches the user's lips comprises:
determining a lip reading syllable in a video image of the audio/video stream at a particular point in time;
determining a voice syllable in audio of the audio/video stream at the particular point in time; and
determining that the lip reading syllable and the voice syllable match.

20. The system of claim 15 , wherein the operations comprise storing the model physiological feature in the object registration information.

Assignees

• Alibaba Group Holding Ltd

Inventors

• Li Peng
• Sun Yipeng
• Xie Yongxiang
• Li Liang

Classifications

• H04L9/40

  Network security protocols · CPC title

• G06V40/16

  Human faces, e.g. facial parts, sketches or expressions · CPC title

• G06V40/20

  Movements or behaviour, e.g. gesture recognition (recognition of facial expressions G06V40/16) · CPC title

• G06V40/172

  Classification, e.g. identification · CPC title

• G10L15/08

  Speech classification or search · CPC title