Security prevention management method and system based on face recognition

The method triangulates face key points into triangular areas, calculating sensitivity and similarity to direction changes, improving face recognition accuracy and efficiency by adapting to facial direction changes.

CN118762392BActive Publication Date: 2024-12-03ZHONGDING INFORMATION TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411246862.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-06
Publication Date
2024-12-03
Estimated Expiration
2044-09-06

AI Technical Summary

Technical Problem

Face recognition verification systems are highly dependent on frontal images, leading to reduced accuracy due to changes in user facial direction.

Method used

A method that performs face key point detection, triangulates key points into triangular areas, calculates sensitivity and similarity of these areas to direction changes and user images, and uses representative triangular areas for verification.

Benefits of technology

Improves face recognition accuracy and efficiency by adapting to facial direction changes, reducing errors, and enhancing the security of access control systems.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118762392B_ABST
    Figure CN118762392B_ABST
Patent Text Reader

Abstract

The present invention belongs to the field of image recognition technology, and specifically relates to a security management method and system based on face recognition, the method comprising: performing face key point detection and triangulation processing on face images of users in different directions, obtaining multiple triangular areas composed of key points, calculating the sensitivity of each triangular area of ​​the user to direction changes according to the internal angle distribution range of each triangular area of ​​the user, determining the sensitivity of the triangular area to face images of different users according to the similarity of the internal angle distribution range of the triangular area of ​​the user and other users, and then determining the resolution value of each triangular area, taking the key points corresponding to the multiple triangular areas with the largest resolution value as representative key points, and performing face recognition verification according to the second representative triangular area composed of representative key points in the verification image of the user to be verified. The present invention improves the efficiency, safety and accuracy of the face recognition access control system.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of image recognition technology, and more specifically, to a security prevention management method and system based on face recognition. Background Art

[0002] The smart home system integration project uses integrated wiring technology, network communication technology, security technology, automatic control technology, and audio and video technology to integrate facilities related to home life, build an efficient management system for residential facilities and family schedule affairs, improve home safety, convenience, comfort, and artistry, and achieve an environmentally friendly and energy-saving living environment; the face recognition access control system is one of the security technologies.

[0003] In the related technology, for example, the Chinese patent document with the authorization announcement number CN109145737B discloses a fast face recognition method, device, electronic device and storage medium, including: obtaining the location information of the sample set and the input face image; using the gradient direction distribution characteristics of the key point neighborhood pixels to obtain the descriptor information of the sample set and the input face image; calculating the key point energy value of the sample set and the input face image: traversing all the key points of the input face image through all the key points of the sample set respectively, and recording all the matching key points obtained as target key points; determining the target sample according to the target key point; and finding the best matching sample with the input face image from the target sample according to the Euclidean distance. The invention can reduce the time complexity of the face matching process, shorten the time spent on face recognition, and achieve the purpose of fast face recognition.

[0004] Face recognition verification in related technologies is highly dependent on the frontal image of the face, which causes changes in the user's facial direction to seriously affect the accuracy of face recognition verification. Summary of the invention

[0005] In order to solve the technical problem in the above-mentioned related technologies that face recognition verification is highly dependent on the frontal image of the face, resulting in changes in the user's facial direction seriously affecting the accuracy of face recognition verification, the present invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides a security management method based on face recognition, including: performing face key point detection on a user's face images in different directions, recording the face image of the user in a preset direction as a frontal image; performing triangulation processing on all key points in the frontal image to obtain a plurality of triangular areas composed of key points; dividing the face image of the user in other directions into a plurality of triangular areas according to the key points that constitute the triangular areas; weighting the statistical results of the sizes of the inner angles of the triangular areas in the face images of the user in all directions according to the size of the direction, obtaining the probability that the inner angle of the triangular area of ​​the user belongs to each angle, and all probabilities constitute the distribution range of the inner angle of the triangular area of ​​the user; calculating the sensitivity of each triangular area of ​​the user to direction changes according to the distribution range of the inner angle of each triangular area of ​​the user; determining the sensitivity of the triangular area to the face images of different users according to the similarity of the inner angle distribution range of the triangular areas of the user and other users; determining the resolution value of each triangular area, the resolution value is negatively correlated with the sensitivity of the triangular area to direction changes, and positively correlated with the sensitivity of the triangular area to the face images of different users; the triangular area with the largest resolution value is selected as the resolution value. The key points corresponding to the triangular area are taken as representative key points. It represents a preset number, and performs face recognition verification according to the second representative triangle area composed of representative key points in the verification image of the user to be verified.

[0007] The present invention triangulates the key points of the user's face image in different directions, connects the key points of the face image into multiple triangular areas, and constructs a more comprehensive and diversified facial geometric feature model by capturing the subtle structure and shape changes of the user's face; subsequently, the resolution value of each triangular area is calculated according to the sensitivity of the triangular area to the direction change and the sensitivity of the triangular area to the face images of different users, so that the first representative triangular area finally obtained according to the resolution value has a lower sensitivity to the direction change and a higher sensitivity to different users, that is, the first representative triangular area finally obtained can distinguish the face images of different users, can perform feature comparison and matching more quickly, accelerates the matching process of facial features during face recognition verification, and improves the efficiency of the face recognition access control system; at the same time, it can adapt to the change of the user's face direction during face recognition, reduces the face recognition errors caused by the direction change, and improves the safety and accuracy of the face recognition access control system.

[0008] Preferably, the probability that the inner angle of the triangular area of ​​the user belongs to each angle satisfies the expression: ; In the formula, Indicates the user's The triangular area of The interior angles belong to the angle The probability of represents the total number of all directions, Indicates The size of the direction, Indicates that the user is in Direction The triangular area of The size of the interior angles, Indicates the angle, The value of is [1,178], represents the judgment function, Used to judge Is it equal to the angle , in response to ,but ; In response to ,but .

[0009] The present invention calculates the probability that the inner angle of the user's triangular area belongs to each angle, captures the subtle structure and shape changes of the user's face, obtains the distribution range of the inner angle of the user's triangular area, and constructs a more comprehensive and diversified facial geometry feature model.

[0010] Preferably, the direction refers to the angle between the normal vector of the plane where the face is located and the normal vector of the plane where the camera is located when the user is collecting the face image. The size of the direction , Take all integers in the range [1,18].

[0011] By collecting facial images of each user in different directions, face recognition errors caused by direction changes can be significantly reduced, thereby improving the security and accuracy of the face recognition access control system.

[0012] Preferably, the sensitivity of each triangular area of ​​the user to the direction change satisfies the expression: ; In the formula, Indicates the user's The sensitivity of the triangle area to changes in direction, , Respectively represent the user's The triangular area of The maximum angle whose frequency is not 0 and the minimum angle whose frequency is not 0 in the distribution range of the interior angles are: Indicates the user's The triangular area of The interior angles belong to the angle The probability of Represents the base 2 logarithmic function.

[0013] By calculating the sensitivity of each triangular area of ​​the user to direction changes, the first representative triangular area finally obtained can adapt to the changes in the user's facial direction during face recognition, reducing face recognition errors caused by direction changes and improving the security and accuracy of the face recognition access control system.

[0014] Preferably, the similarity of the inner angle distribution range of the triangle area between the user and other users satisfies the expression: ; In the formula, Indicates the user's The similarity of the distribution range of the internal angles of the triangular areas, Represents the number of all users, Indicates the user's The triangular area of The interior angles belong to the angle The probability of Indicates that the user other than this user User's The triangular area of The interior angles belong to the angle The probability of represents the logarithmic function with base 2, Represents an exponential function with a natural constant as its base.

[0015] Preferably, determining the sensitivity of the triangular area to the face images of different users includes: comparing the sensitivity of all users to the face images of other users. The similarity of the internal angle distribution ranges of the triangle areas is sorted in descending order, and the sorted The similarity is recorded as the first value, and the difference between 1 and the first value is recorded as the second value. The sensitivity of the triangular area to the face images of different users, Indicates rounding down. Indicates the number of all users.

[0016] By calculating the sensitivity of the triangular area to the facial images of different users, the first representative triangular area finally obtained can distinguish the facial images of different users, and can perform feature comparison and matching faster, which accelerates the facial feature matching process during face recognition verification and improves the efficiency of the face recognition access control system.

[0017] Preferably, the resolution value of each triangular area satisfies the expression: ; In the formula, Indicates The resolution value of the triangle area for different users, Indicates the user's The sensitivity of the triangle area to changes in direction, Represents all users The average sensitivity of the triangular area to the change of direction, Indicates The sensitivity of the triangular area to the face images of different users.

[0018] Preferably, the face recognition verification process includes: triangular areas are recorded as first representative triangular areas; each user is regarded as a registered user, and the possibility that the user to be verified belongs to each registered user is calculated according to the size of each inner angle of the second representative triangular area and the distribution range of the inner angle of the first representative triangular area of ​​each registered user; in response to the possibility that the user to be verified belongs to each registered user is not less than a first preset threshold, the user to be verified passes the face recognition verification; in response to the possibility that the user to be verified belongs to each registered user is less than the first preset threshold, but there is a possibility that is not less than the second preset threshold, the management personnel is notified to perform personal information identification verification on the user to be verified; in response to the possibility that the user to be verified belongs to each registered user is less than the second preset threshold, the user to be verified fails the face recognition verification, and the first preset threshold is greater than the second preset threshold.

[0019] Preferably, the possibility that the user to be verified belongs to each registered user satisfies the expression: ; In the formula, Indicates that the user to be verified belongs to The possibility of having registered users, Indicates of registered users The first represents the distribution range of the inner angle of the triangle area. The interior angle is equal to The probability of Indicates the first The second represents the area of ​​the triangle The size of the interior angle.

[0020] In a second aspect, the present invention provides a security prevention management system based on face recognition, which adopts the following technical solutions:

[0021] The security and prevention management system based on face recognition includes: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the security and prevention management method based on face recognition is implemented.

[0022] By adopting the above technical solution, the above-mentioned security prevention management method based on face recognition is generated into a computer program and stored in a memory so as to be loaded and executed by a processor, thereby making a terminal device based on the memory and the processor for easy use.

[0023] The beneficial effects of the present invention are:

[0024] The present invention triangulates the key points of the user's face image in different directions, connects the key points of the face image into multiple triangular areas, and constructs a more comprehensive and diversified facial geometric feature model by capturing the subtle structure and shape changes of the user's face; subsequently, the resolution value of each triangular area is calculated according to the sensitivity of the triangular area to the direction change and the sensitivity of the triangular area to the face images of different users, so that the first representative triangular area finally obtained according to the resolution value has a lower sensitivity to the direction change and a higher sensitivity to different users, that is, the first representative triangular area finally obtained can distinguish the face images of different users, can perform feature comparison and matching more quickly, accelerates the matching process of facial features during face recognition verification, and improves the efficiency of the face recognition access control system; at the same time, it can adapt to the change of the user's face direction during face recognition, reduces the face recognition errors caused by the direction change, and improves the safety and accuracy of the face recognition access control system. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0026] Figure 1 is a flow chart schematically illustrating a security prevention management method based on face recognition in the present invention;

[0027] Figure 2 Schematically illustrates 68 key points in a frontal image of a user. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0029] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0030] The embodiment of the present invention discloses a security prevention management method based on face recognition, referring to Figure 1 , comprising steps S1 to S7:

[0031] S1. Collect facial images of each user in different directions.

[0032] It should be noted that each person's facial features appear differently at different angles. For example, the difference in features between the side face and the front face may affect the face recognition effect. By collecting facial images of users from different directions, the face recognition access control system can establish a more comprehensive and diverse facial feature model. Therefore, when actually performing face recognition verification, the system can more accurately identify the user even if the user's facial angle changes, thereby improving the recognition accuracy and robustness of the face recognition access control system.

[0033] Specifically, for each user, a high-definition camera is used to collect facial images of the user in different directions, where the direction refers to the angle between the normal vector of the plane where the face is located and the normal vector of the plane where the high-definition camera is located when the user is collecting facial images; wherein, The size of the direction , Take all integers in the range [1,18].

[0034] Furthermore, the face image of the user in a preset direction is recorded as a frontal image, where the preset direction refers to a direction corresponding to an angle of 0°.

[0035] It should be noted that by collecting facial images of each user in different directions, face recognition errors caused by direction changes can be significantly reduced, thereby improving the security and accuracy of the face recognition access control system.

[0036] S2. Perform facial key point detection and triangulation processing on the user's facial images in different directions to obtain multiple triangular regions in the user's facial images in all directions.

[0037] Specifically, facial landmark detection is performed on the user's face images in different directions through the face key point detector (shape_predictor_68_face_landmarks.dat) of Python's Dlib library to obtain 68 key points in the user's face images in different directions.

[0038] For example, the 68 key points in the frontal image of the user are as follows: Figure 2 shown.

[0039] Among them, Python's Dlib library is a C++ open source toolkit that contains machine learning algorithms, and the face key point detector (shape_predictor_68_face_landmarks.dat) is a module in this open source toolkit.

[0040] It should be noted that the key points obtained in different face images represent the facial features of different users and can be used to distinguish different faces.

[0041] Furthermore, all key points in the frontal image are triangulated using the Delaunay triangulation algorithm to obtain a plurality of triangular areas composed of the key points, wherein the vertices of each triangular area are the key points in the frontal image, each triangular area has three internal angles, and the sum of the sizes of the three internal angles is equal to 180°; based on the key points that constitute the triangular area, the user's facial image in other directions is divided into a plurality of triangular areas.

[0042] It should be noted that the Delaunay triangulation algorithm is a well-known technology and will not be described in detail here.

[0043] It should be noted that triangulating the user's face images in all directions improves the extraction accuracy and matching robustness of facial features, and helps to enhance the accuracy and reliability of the face recognition system.

[0044] It should be further explained that the present invention calculates the probability that the inner angle of the user's triangular area belongs to each angle, captures the subtle structure and shape changes of the user's face, obtains the distribution range of the inner angle of the user's triangular area, and constructs a more comprehensive and diverse facial geometry feature model.

[0045] S3. Obtain the inner angle distribution range of each triangular area of ​​the user.

[0046] Specifically, according to the size of the direction, the statistical results of the sizes of the inner angles of the triangular areas of the user's face images in all directions are weighted to obtain the probability that the inner angle of the triangular area of ​​the user belongs to each angle, and the probability that the inner angle of the triangular area of ​​the user belongs to each angle satisfies the expression:

[0047] ;

[0048] In the formula, Indicates the user's The triangular area of The interior angles belong to the angle The probability of represents the total number of all directions, Indicates The size of the direction, Indicates that the user is in Direction The triangular area of The size of the interior angles, Indicates the angle, The value of is [1,178], represents the judgment function, Used to judge Is it equal to the angle , in response to ,but ; In response to ,but .

[0049] It should be noted that when the user is collecting facial images, the closer the user's face is facing the high-definition camera, the angle between the normal vector of the plane where the face is located and the normal vector of the plane where the camera is located, that is, the direction corresponding to the collected facial image, is closer to 0°; when the user is performing facial recognition verification, the user consciously faces the camera, mainly because doing so helps improve the accuracy of the facial recognition system. Facing the camera can ensure that the key points of facial features are presented at the best angle in the camera's field of view as much as possible, which enables the system to capture and analyze facial features more clearly and accurately; in addition, this can avoid recognition errors or unrecognition caused by angle deviation, and improve the efficiency and success rate of the verification process; therefore, the closer the direction is to 0°, that is, the smaller the direction, the more likely it is to occur in the actual facial recognition verification process, so when counting the size of the inner angle of the triangular area in the user's facial image in all directions, the statistical results are weighted by the size of the direction. The smaller the size of the direction, the higher the weight. The bigger.

[0050] Furthermore, the user's The triangular area of The probability that the inner angle belongs to all angles constitutes the user's The triangular area of The distribution range of the inner angles; the user’s The distribution range of all inner angles of the triangle area constitutes the user's The distribution range of the interior angles of a triangular area.

[0051] S4. Calculate the sensitivity of each triangular area of ​​the user to the direction change according to the distribution range of the inner angles of each triangular area of ​​the user.

[0052] Specifically, according to the distribution range of the inner angles of each triangular area of ​​the user, the sensitivity of each triangular area of ​​the user to the direction change is calculated; then the sensitivity of each triangular area of ​​the user to the direction change satisfies the expression:

[0053] ;

[0054] In the formula, Indicates the user's The sensitivity of the triangle area to changes in direction, , Respectively represent the user's The triangular area of The maximum angle whose frequency is not 0 and the minimum angle whose frequency is not 0 in the distribution range of the interior angles are: Indicates the user's The triangular area of The interior angles belong to the angle The probability of Represents the base 2 logarithmic function.

[0055] It should be noted that Represents the user's The triangular area of The information entropy of the distribution range of the inner angles. The larger the value, the higher the information entropy of the user's first The triangular area of The more chaotic the distribution of the inner angles is, the more likely the user is to have The triangular area of The greater the difference in the sizes of the inner angles, the greater the The sensitivity of the triangle area to changes in direction The bigger; Represents the user's The triangular area of The size of the concentrated area in the distribution range of the inner angles. The larger the value, the greater the user's The triangular area of The less concentrated the distribution of the inner angles is, the more likely the user is to have The triangular area of The greater the difference in the sizes of the inner angles, the greater the The sensitivity of the triangle area to changes in direction The bigger.

[0056] It should be further explained that by calculating the sensitivity of each triangular area of ​​the user to direction changes, the first representative triangular area finally obtained can adapt to the changes in the user's facial direction during face recognition, reducing face recognition errors caused by direction changes and improving the security and accuracy of the face recognition access control system.

[0057] S5. Determine the sensitivity of the triangular area to the facial images of different users based on the similarity between the inner angle distribution ranges of the triangular area of ​​the user and other users.

[0058] Specifically, according to the inner angle distribution ranges of the triangular areas of the user and the inner angle distribution ranges of the triangular areas of other users, the similarity of the inner angle distribution ranges of the triangular areas of the user and other users is calculated, and the similarity of the inner angle distribution ranges of the triangular areas of the user and other users satisfies the expression:

[0059] ;

[0060] In the formula, Indicates the user's The similarity of the distribution range of the internal angles of the triangular areas, Represents the number of all users, Indicates the user's The triangular area of The interior angles belong to the angle The probability of Indicates that the user other than this user User's The triangular area of The interior angles belong to the angle The probability of represents the logarithmic function with base 2, Represents an exponential function with a natural constant as its base.

[0061] It should be noted that Represents the user's The triangular area of The distribution range of the interior angles is the same as that of the User's The triangular area of The greater the value, the more overlapped the distribution range of the inner angles of this user is with other users. Similarity of the distribution range of the interior angles of the triangular regions The bigger.

[0062] Furthermore, for all users and other users The similarity of the internal angle distribution ranges of the triangle areas is sorted in descending order, and the sorted The similarity is recorded as the first value, and the difference between 1 and the first value is recorded as the second value. The sensitivity of the triangular area to the face images of different users, Indicates rounding down. Indicates the number of all users.

[0063] It should be noted that the user and other users The greater the similarity of the inner angle distribution range of the triangle area, the greater the similarity of the user to other users. The smaller the difference in the distribution range of the inner angles of the triangle areas, the smaller the similarity is. The greater the similarity, the closer the user is to other users. The smaller the difference in the distribution range of the internal angles of the triangle areas, the The smaller the sensitivity of the triangular area to the face images of different users.

[0064] It should be further explained that by calculating the sensitivity of the triangular area to the facial images of different users, the first representative triangular area finally obtained can distinguish the facial images of different users, and can perform feature comparison and matching faster, thereby accelerating the matching process of facial features during face recognition verification and improving the efficiency of the face recognition access control system.

[0065] S6. Determine the resolution value of each triangular area.

[0066] It should be noted that the present invention expects that the first representative triangle area finally obtained according to the resolution value will have the ability to distinguish the facial images of different users, and at the same time have the ability to cope with the changes in facial direction when the user performs facial recognition verification. Therefore, it is expected that the first representative triangle area finally obtained according to the resolution value will be less sensitive to direction changes, and at the same time more sensitive to the facial images of different users. Therefore, the resolution value is negatively correlated with the sensitivity of the triangle area to direction changes, and positively correlated with the sensitivity of the triangle area to different users.

[0067] Specifically, the resolution value of each triangular area satisfies the expression:

[0068] ;

[0069] In the formula, Indicates The resolution value of the triangle area for different users, Indicates the user's The sensitivity of the triangle area to changes in direction, Represents all users The average sensitivity of the triangular area to the change of direction, Indicates The sensitivity of the triangular area to the face images of different users.

[0070] It should be noted that The sensitivity of the triangular area to the face images of different users The value range of is [0,1], so The distribution obeys the gamma transformation, The sensitivity of the triangular area to the face images of different users The larger the number, the more The average sensitivity of the triangle area to the change of direction The smaller the The resolution value of the triangle area for different users The bigger.

[0071] S7. Obtain representative key points according to the resolution value, and perform face recognition verification according to a second representative triangle area formed by the representative key points in the verification image of the user to be verified.

[0072] Specifically, the maximum resolution value The triangle area is recorded as the first representative triangle area. represents a preset number; takes the key point corresponding to the first representative triangle area as the representative key point; takes the inner angle distribution range of the first representative triangle area of ​​each user as the verification information of each user; takes each user as a registered user, and stores the verification information of all registered users in the face recognition access control system.

[0073] It is particularly important to note that in addition to the verification information of registered users, the facial recognition access control system also stores the personal information of registered users, including but not limited to name, frontal image, address, and personal identification document information (such as ID number).

[0074] The specific value of the preset quantity can be set according to the actual application scenario and requirements, and the value range of the preset quantity is , Represents the total number of all triangular areas in the front image. The present invention sets the preset number to .

[0075] Furthermore, when the user to be verified performs face recognition verification through the face recognition access control system, a face image of the user to be verified is collected through a high-definition camera and recorded as a verification image of the user to be verified; face key point detection is performed on the verification image of the user to be verified to obtain all key points of the verification image of the user to be verified; a second representative triangle area is formed according to representative key points among all key points of the verification image of the user to be verified; and face recognition verification is performed according to the second representative triangle area, including:

[0076] 1. Calculate the possibility that the user to be verified belongs to each registered user according to the size of each inner angle of the second representative triangle area and the verification information of each registered user;

[0077] 2. In response to the possibility that the to-be-verified user belongs to each of the registered users, there is a possibility that is not less than a first preset threshold, indicating that the to-be-verified user has a high degree of match with at least one registered user in the system, and the identity of the to-be-verified user is considered to have been confirmed, and the to-be-verified user passes the face recognition verification;

[0078] 3. In response to the possibility that the to-be-verified user belongs to each registered user being less than the first preset threshold, but there is a possibility that is not less than the second preset threshold, indicating that the to-be-verified user has a certain degree of match with some registered users in the system, but the degree of match is not enough to directly confirm its identity, the administrator is notified to perform personal information identification and verification on the to-be-verified user, and the administrator checks the information provided by the to-be-verified user with the personal information of the registered users stored in the face recognition access control system;

[0079] 4. In response to the possibility that the user to be verified belongs to each registered user being less than the second preset threshold, it means that the matching degree between the user to be verified and all registered users in the system is insufficient, and the user to be verified fails the face recognition verification.

[0080] Among them, the first preset threshold is greater than the second preset threshold, the specific values ​​of the first preset threshold and the second preset threshold can be set according to actual application scenarios and requirements, and the value range of the first preset threshold and the second preset threshold is [0,1]. The present invention sets the first preset threshold to 0.8 and the second preset threshold to 0.5.

[0081] It should be noted that by setting different thresholds, the uncertainties that may arise in the identification process can be effectively dealt with, providing a multi-level verification solution.

[0082] The possibility that the user to be verified belongs to each registered user satisfies the expression:

[0083] ;

[0084] In the formula, Indicates that the user to be verified belongs to The possibility of having registered users, Indicates of registered users The first represents the distribution range of the inner angle of the triangle area. The interior angle is equal to The probability of Indicates the first The second represents the area of ​​the triangle The size of the interior angles, Indicates the number of presets.

[0085] It should be noted that the present invention triangulates the key points of the user's facial image in different directions, connects the key points of the facial image into multiple triangular areas, and constructs a more comprehensive and diversified facial geometric feature model by capturing the subtle structure and shape changes of the user's face; subsequently, the resolution value of each triangular area is calculated according to the sensitivity of the triangular area to direction changes and the sensitivity of the triangular area to the facial images of different users, so that the first representative triangular area finally obtained according to the resolution value is less sensitive to direction changes and more sensitive to different users. In other words, the first representative triangular area finally obtained can distinguish the facial images of different users, and can perform feature comparison and matching faster, which accelerates the matching process of facial features during face recognition verification and improves the efficiency of the face recognition access control system; at the same time, it can adapt to the changes in the facial direction of the user during face recognition, reduce face recognition errors caused by direction changes, and improve the security and accuracy of the face recognition access control system.

[0086] An embodiment of the present invention also discloses a security and prevention management system based on face recognition, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a security and prevention management method based on face recognition according to the present invention is implemented.

[0087] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, and their configuration and functions are known in the art, so they will not be described in detail here.

[0088] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0089] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A security prevention management method based on face recognition, characterized in that: include: Perform facial key point detection on the user's facial images in different directions, and record the user's facial images in a preset direction as frontal images; Perform triangulation processing on all key points in the front image to obtain multiple triangular areas composed of the key points; divide the face image of the user in other directions into multiple triangular areas according to the key points that constitute the triangular areas; According to the size of the direction, the statistical results of the size of the inner angle of the triangular area of ​​the user's face image in all directions are weighted to obtain the probability that the inner angle of the triangular area of ​​the user belongs to each angle, satisfying the expression: ; In the formula, Indicates the user's The triangular area of The interior angles belong to the angle The probability of represents the total number of all directions, Indicates The size of the direction, Indicates that the user is in Direction The triangular area of The size of the interior angles, Indicates the angle, The value of is [1,178], represents the judgment function, Used to judge Is it equal to the angle , in response to ,but ; In response to ,but , all probabilities constitute the inner angle distribution range of the triangular area of ​​the user; Calculate the sensitivity of each triangular area of ​​the user to the direction change according to the distribution range of the inner angles of each triangular area of ​​the user; determine the sensitivity of the triangular area to the facial images of different users according to the similarity between the distribution range of the inner angles of the triangular areas of the user and other users; Determine the resolution value of each triangular area, wherein the resolution value is negatively correlated with the sensitivity of the triangular area to direction changes and positively correlated with the sensitivity of the triangular area to face images of different users; The triangle area is recorded as the first representative triangle area. Indicates the number of presets; The key points corresponding to the first representative triangle area are used as representative key points; the inner angle distribution range of the first representative triangle area of ​​each user is used as the verification information of each user; each user is used as a registered user, and the verification information of all registered users is stored in the face recognition access control system, and the face key point detection is performed on the verification image of the user to be verified, and all the key points of the verification image of the user to be verified are obtained. The second representative triangle area is composed of representative key points in all the key points in the verification image of the user to be verified, and the face recognition verification is performed, including calculating the possibility that the user to be verified belongs to each registered user according to the size of each inner angle of the second representative triangle area and the inner angle distribution range of the first representative triangle area of ​​each registered user to satisfy the expression: ; In the formula, Indicates that the user to be verified belongs to The possibility of having registered users, Indicates of registered users The first represents the distribution range of the inner angle of the triangle area. The interior angle is equal to The probability of Indicates the first The second represents the area of ​​the triangle the size of the inner angle; in response to the possibility that the user to be verified belongs to each registered user, there is a possibility that is not less than a first preset threshold, the user to be verified passes the face recognition verification; In response to the possibility that the user to be verified belongs to each registered user being less than the first preset threshold, but there is a possibility that is not less than the second preset threshold, notifying the administrator to perform personal information identification verification on the user to be verified; In response to the possibility that the user to be verified belongs to each registered user is less than the second preset threshold, the user to be verified fails the face recognition verification, and the first preset threshold is greater than the second preset threshold.

2. The security prevention management method based on face recognition according to claim 1 is characterized in that: The direction refers to the angle between the normal vector of the plane where the face is located and the normal vector of the plane where the camera is located when the user is collecting the face image. The size of the direction , Take all integers in the range [1,18].

3. The security prevention management method based on face recognition according to claim 1 is characterized in that: The sensitivity of each triangle area of ​​the user to the direction change satisfies the expression: ; In the formula, Indicates the user's The sensitivity of the triangle area to changes in direction, , Respectively represent the user's The triangular area of The maximum angle whose frequency is not 0 and the minimum angle whose frequency is not 0 in the distribution range of the interior angles are: Indicates the user's The triangular area of The interior angles belong to the angle The probability of Represents the base 2 logarithmic function.

4. The security management method based on face recognition according to claim 1 is characterized in that: The similarity between the inner angle distribution range of the triangle area of ​​the user and other users satisfies the expression: ; In the formula, Indicates the user's The similarity of the distribution range of the internal angles of the triangular areas, Represents the number of all users, Indicates the user's The triangular area of The interior angles belong to the angle The probability of Indicates that the user other than this user User's The triangular area of The interior angles belong to the angle The probability of represents the logarithmic function with base 2, Represents an exponential function with a natural constant as its base.

5. The security prevention management method based on face recognition according to claim 1 is characterized in that: The step of determining the sensitivity of the triangular area to the facial images of different users includes: For all users and other users The similarity of the internal angle distribution ranges of the triangle areas is sorted in descending order, and the sorted The similarity is recorded as the first value, and the difference between 1 and the first value is recorded as the second value. The sensitivity of the triangular area to the face images of different users, Indicates rounding down. Indicates the number of all users.

6. The security prevention management method based on face recognition according to claim 1 is characterized in that: The resolution value of each triangular area satisfies the expression: ; In the formula, Indicates The resolution value of the triangle area for different users, Indicates the user's The sensitivity of the triangle area to changes in direction, Represents all users The average sensitivity of the triangular area to the change of direction, Indicates The sensitivity of the triangular area to the face images of different users.

7. The security management system based on face recognition is characterized by: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the security prevention management method based on face recognition according to any one of claims 1 to 6 is implemented.