CN-122020623-A - Encryption verification method and system for face recognition

CN122020623ACN 122020623 ACN122020623 ACN 122020623ACN-122020623-A

Abstract

The invention is applicable to the field of face recognition and provides an encryption verification method and system for face recognition, wherein the method comprises the following steps of creating a facial feature template of a user and generating a first action sequence for normal verification and a second action sequence for abnormal condition verification for the facial feature template; the method comprises the steps of collecting a user face video stream, detecting and identifying a face action sequence which is automatically executed by a user in real time and extracting face biological feature data, matching the similarity between the face action sequence and an action sequence in a registration stage, comparing the extracted face biological feature data with a face feature template, executing verification passing response when the face action sequence is matched with a first action sequence, and triggering preset implicit safety response when executing verification passing response when the face action sequence is matched with a second action sequence. The invention introduces a user-defined and multipurpose dual action sequence mechanism, and solves the problems of immobilization of living body detection actions, easy observation and imitation and incapability of coping with stress scenes.

Inventors

TAO JIE
DONG LIJUN
ZHAO YANG
SHEN XIAOBO
LU YONGJIN

Assignees

江苏锡铁慧通科技有限公司
无锡地铁集团有限公司

Dates

Publication Date: 20260512
Application Date: 20260130

Claims (10)

1. An encryption verification method for face recognition, the method comprising the steps of: In the registration stage, creating a facial feature template of a user and generating a first action sequence for normal verification and a second action sequence for abnormal condition verification for the facial feature template, wherein the first action sequence and the second action sequence comprise a plurality of facial actions and are different in sequence; in the verification stage, collecting a user face video stream, detecting and identifying a face action sequence which is automatically executed by a user in real time, and extracting face biological characteristic data; Matching the similarity of the facial action sequence and the action sequence of the registration stage, and comparing the extracted facial biological feature data with a facial feature template; when matching the first action sequence, a verification passing response is executed, and when matching the second action sequence, a preset implicit security response is triggered while the verification passing response is executed.
2. The encryption verification method for face recognition according to claim 1, wherein the step of creating and generating a first action sequence for normal verification and a second action sequence for abnormal condition verification for the facial feature template of the user specifically comprises: establishing a facial feature template of a user based on a process of acquiring user information by a terminal device, wherein images of all facial actions of the user can be acquired in the process; Establishing a preset action library comprising a plurality of action units according to the images of all facial actions, and pushing a user-defined action sequence interface capable of being selected to a user; Generating a first action sequence and a second action sequence according to a user-defined selection result; and encrypting and storing the first action sequence, the second action sequence and the facial feature template.
3. The encryption and authentication method for face recognition according to claim 2, wherein the step of storing the first action sequence, the second action sequence and the facial feature template in an encryption association manner specifically comprises: respectively performing action unit decomposition and feature extraction on the first action sequence and the second action sequence to generate digital fingerprints of the action sequences; Extracting feature point distribution data from a facial feature template by using a key derivation function based on facial features of a user to generate a sequence encryption key; symmetrically encrypting the digital fingerprints of the action sequence based on a sequence encryption key, and carrying out association binding on the encrypted action fingerprints and the facial feature template through an irreversible hash chain structure; The sequence encryption key is split into a plurality of key components which are respectively bound and stored with different facial features of the user, so that the reconstruction of the complete key needs to verify a plurality of facial features simultaneously.
4. The method for encrypting and verifying face recognition according to claim 1, wherein the step of matching the facial motion sequence with the motion sequence of the registration stage in similarity and comparing the extracted facial biometric data with the facial feature template comprises: Comparing the type and sequence of the facial action sequence with the registered first action sequence or second action sequence, and calculating the sequence consistency score of the facial action sequence; performing time sequence dynamic alignment on the sequences subjected to the sequence comparison, analyzing the start and stop time and time interval of action execution, and calculating the time sequence matching score of the sequences; Performing similarity calculation on the extracted facial biological feature data and a facial feature template to obtain a biological feature matching score; Based on a preset decision rule, a matching result is generated according to the sequence consistency score, the time regular matching score and the biological characteristic matching score.
5. The encryption and authentication method for face recognition according to claim 3, wherein the sequence encryption key, the digital fingerprint of the action sequence, and the hash value of the facial feature template are stored in a security chip, the security chip supports a cryptographic algorithm capable of generating and managing a session key for data transmission, and the security chip negotiates the session key with the authentication server through an asymmetric encryption algorithm for encrypting and transmitting facial biometric data and the action sequence data.
6. The encryption verification method for face recognition according to claim 1, further comprising: When multiple verification failures or abnormal login positions are detected, generating a random dynamic action challenge instruction, wherein the random dynamic action challenge instruction is used for displaying and executing a challenge action sequence with a certain specific sequence on a user interface, the number of the challenge action sequence is the same as that of facial actions contained in the first action sequence or the second action sequence, and the challenge action sequence is executed in a hysteresis way; collecting a current action sequence of a user, and splitting the current action sequence into two auxiliary action sequences according to a time sequence; Matching the former auxiliary action sequence with a preset first action sequence or second action sequence, and matching the latter auxiliary action sequence with a challenge action sequence; when both match, then it is determined that the action sequence match was successful.
7. An encryption verification system for face recognition, the system comprising: The registration creation module is used for creating a facial feature template of a user and generating a first action sequence for normal verification and a second action sequence for abnormal condition verification for the facial feature template of the user in a registration stage, wherein the first action sequence and the second action sequence comprise a plurality of facial actions and are different in sequence; The verification acquisition module is used for acquiring a user face video stream, detecting and identifying a face action sequence which is automatically executed by a user in real time and extracting face biological characteristic data in a verification stage; the matching and comparing module is used for matching the similarity between the facial action sequence and the action sequence in the registration stage and comparing the extracted facial biological feature data with a facial feature template; And the response execution module is used for executing the verification passing response when being matched with the first action sequence, and triggering a preset hidden safety response when executing the verification passing response when being matched with the second action sequence.
8. The encryption validation system for face recognition of claim 7, wherein the registration creation module comprises: The template establishing unit is used for establishing a facial feature template of the user based on a process of acquiring user information by the terminal equipment, wherein images of all facial actions of the user can be acquired in the process; the action library establishing unit is used for establishing a preset action library comprising a plurality of action units according to the images of all facial actions and pushing a user-defined action sequence interface capable of being selected to a user; the sequence generating unit is used for generating a first action sequence and a second action sequence according to the user-defined selection result of the user; and the encryption storage unit is used for carrying out encryption association storage on the first action sequence, the second action sequence and the facial feature template.
9. The encryption verification system for face recognition according to claim 7, wherein the matching comparison module includes: a sequence comparison unit, configured to compare the type and sequence of the facial action sequence with the registered first action sequence or second action sequence, and calculate a sequence consistency score thereof; the time sequence alignment unit is used for dynamically aligning the time sequence of the sequence comparison, analyzing the start and stop time and time interval of the execution of the action and calculating the time regularity matching score of the time sequence; the feature comparison unit is used for carrying out similarity calculation on the extracted facial biological feature data and the facial feature template to obtain a biological feature matching score; the result generation unit is used for generating a matching result according to the sequence consistency score, the time regular matching score and the biological characteristic matching score based on a preset decision rule.
10. The encryption validation system for face recognition according to claim 7, wherein the system further comprises a dynamic validation module comprising: The instruction generating unit is used for generating a random dynamic action challenge instruction when detecting multiple verification failures or abnormal login positions, and displaying and executing a challenge action sequence with a certain specific sequence on a user interface, wherein the number of the challenge action sequence is the same as that of facial actions contained in the first action sequence or the second action sequence, and the challenge action sequence is executed in a hysteresis way; the sequence splitting unit is used for collecting the current action sequence of the user and splitting the current action sequence into two auxiliary action sequences according to the time sequence; The double matching unit is used for matching the former auxiliary action sequence with a preset first action sequence or second action sequence and matching the latter auxiliary action sequence with a challenge action sequence; and the result judging unit is used for judging that the action sequence matching is successful when the two are matched.

Description

Encryption verification method and system for face recognition Technical Field The invention relates to the field of face recognition, in particular to an encryption verification method and system for face recognition. Background Along with the rapid development of the biological recognition technology, the face recognition is widely applied to the fields of identity verification, payment safety, access control and the like due to the characteristics of non-contact, convenience and high efficiency. The traditional face recognition system mainly relies on single facial feature comparison, and identity authentication is completed by collecting facial images of a user and matching the facial images with a pre-stored template. However, the identification mode based on the static features has obvious potential safety hazards and is easily threatened by attacks presented by photos, videos, 3D masks and the like. To enhance security, the prior art has introduced a biopsy mechanism, typically requiring the user to perform specific action instructions, such as blinking, opening the mouth, shaking the head, etc. But such action instructions are generally uniformly generated and prompted by the system, and all users execute the same action sequence, and lack personalization. An attacker can easily obtain the action requirements by looking at it and bypass the detection by recording playback or making a dynamic mask. More importantly, when the user is stressed to perform face recognition, the existing system cannot distinguish the voluntary state from the stressed state, which may lead to the user to finish identity verification under stress, and cause property or security loss. In summary, the conventional face recognition technology has the problems that the living body detection action is immobilized and is easy to observe and imitate. Therefore, a need exists for a face recognition encryption verification method that can compromise security, privacy and user experience, which can be used conveniently in normal situations and can provide protection in special situations. Disclosure of Invention Aiming at the defects existing in the prior art, the invention aims to provide an encryption verification method and system for face recognition so as to solve the problems existing in the background art. The invention is realized in that an encryption verification method for face recognition comprises the following steps: In the registration stage, creating a facial feature template of a user and generating a first action sequence for normal verification and a second action sequence for abnormal condition verification for the facial feature template, wherein the first action sequence and the second action sequence comprise a plurality of facial actions and are different in sequence; in the verification stage, collecting a user face video stream, detecting and identifying a face action sequence which is automatically executed by a user in real time, and extracting face biological characteristic data; Matching the similarity of the facial action sequence and the action sequence of the registration stage, and comparing the extracted facial biological feature data with a facial feature template; when matching the first action sequence, a verification passing response is executed, and when matching the second action sequence, a preset implicit security response is triggered while the verification passing response is executed. The invention further provides that the step of creating the facial feature template of the user and generating a first action sequence for normal verification and a second action sequence for abnormal condition verification for the facial feature template specifically comprises the following steps: establishing a facial feature template of a user based on a process of acquiring user information by a terminal device, wherein images of all facial actions of the user can be acquired in the process; Establishing a preset action library comprising a plurality of action units according to the images of all facial actions, and pushing a user-defined action sequence interface capable of being selected to a user; Generating a first action sequence and a second action sequence according to a user-defined selection result; and encrypting and storing the first action sequence, the second action sequence and the facial feature template. As a further scheme of the invention: the step of encrypting and storing the first action sequence, the second action sequence and the facial feature template specifically comprises the following steps: respectively performing action unit decomposition and feature extraction on the first action sequence and the second action sequence to generate digital fingerprints of the action sequences; Extracting feature point distribution data from a facial feature template by using a key derivation function based on facial features of a user to generate a sequence encryption key; symmetrically enc