A security management system for facial image storage data

Abstract

The invention discloses a security management system for face image storage data, which belongs to the technical field of face image data security management, and comprises a storage data processing module, a user terminal, and a data verification scheduling module; the storage data processing module is used for processing face image data stored in a door access control database; the user terminal is used by users, and comprises a face data management module and a face verification module; the face data management module is used for users to manage corresponding face information; the data verification scheduling module is used for data verification scheduling during door access control face recognition; the face verification module is used for face verification; through the mutual cooperation among the storage data processing module, the user terminal, and the data verification scheduling module, the user face data stored in the community door access control is securely processed, thereby avoiding leakage of the user face data stored in the existing door access control system; and improving the user's confidence in using face recognition.

Description

Technical Field

[0001] The invention belongs to the technical field of facial image data security management, and in particular is a security management system for facial image storage data. Background Art

[0002] Traditional access control systems require the use of keys or passwords for identity verification, which is cumbersome to operate and easy to forget. Face recognition technology can achieve contactless identity verification and provide a more convenient and quick access management method. Therefore, face recognition is currently used for access control management in many communities; however, in recent years, data leakage and abuse incidents have been common, posing a serious threat to personal privacy and social security. In order to prevent the leakage and abuse of facial image data, many residents are currently unwilling to use the community's face recognition system for face recognition, fearing that facial information will be leaked, resulting in a large obstacle to the application of the face recognition system; especially when a large amount of resident facial information is stored in the access control system, the stored facial data is also easily coveted by bad people, and the property management personnel lack the corresponding professional capabilities; therefore, it is necessary to establish a secure management system to strengthen the protection of facial image data, protect residents' privacy and enhance people's confidence in use; based on this, the present invention provides a secure management system for facial image storage data. Summary of the invention

[0003] In order to solve the problems existing in the above-mentioned solutions, the present invention provides a security management system for storing facial image data.

[0004] The purpose of the present invention can be achieved through the following technical solutions:

[0005] A security management system for facial image storage data, comprising a storage data processing module, a user terminal, and a data verification scheduling module;

[0006] The stored data processing module is used to process the facial image data stored in the access control database, identify the user corresponding to each facial image data stored in the access control database in real time, and send the facial image data to the user terminal of the corresponding user;

[0007] When a user matching feature of a corresponding user is identified in the access control database, the facial image data of the user corresponding to the user matching feature is deleted from the access control database.

[0008] The user terminal is used by users and includes a face data management module and a face verification module;

[0009] The face data management module is used to manage the corresponding face information of the user, determine the target face image of the user, generate the corresponding user matching features based on the target face image, and send the user matching features to the access control database.

[0010] Furthermore, the method for determining the target face image includes:

[0011] Receiving the facial image data sent by the storage data processing module in real time, determining the user facial image according to the received facial image data; marking the determined user facial image as a target facial image;

[0012] When the facial image data sent by the storage data processing module is not received, the user enters the user's facial image by himself. When the user's facial image is successfully entered, the entered user's facial image is marked as a target facial image.

[0013] Furthermore, the method for setting the user matching feature includes:

[0014] Identify a target face image of a user; determine a plurality of feature regions according to the target face image; process each of the feature regions to form a feature unit, and combine each of the feature units into a user matching feature.

[0015] Furthermore, the method for setting the feature area includes:

[0016] Marking each to-be-selected area in the target face image, and determining each corresponding initial combination based on the user matching features already stored in the access control database;

[0017] Performing feature processing on each of the initial combinations to obtain corresponding simulated user matching features;

[0018] Calculating the difference between the simulated user matching feature and the matching features of each user already stored in the access control database;

[0019] Marking the initial combinations whose difference values ​​are all greater than the threshold value X1 as intermediate combinations;

[0020] Identify the number of to-be-selected areas in each of the intermediate combinations, and select the intermediate combination with the least number of to-be-selected areas as the target to-be-selected combination;

[0021] If there are multiple intermediate combinations with the least number of areas to be selected, the intermediate combination with the largest difference value is selected as the target combination to be selected.

[0022] Furthermore, the calculation method of the difference value includes:

[0023] Determine the corresponding feature units and verification units according to the simulated user matching features for calculating the difference values ​​and the stored user matching features;

[0024] Calculate the corresponding element difference according to the feature unit and the verification unit, and mark the element difference as Kt, where t=1, 2, ..., c, and c is a positive integer;

[0025] According to the formula DC = ∑K t Calculate the corresponding unit difference, where DC is the unit difference;

[0026] The difference values ​​of each unit are accumulated to obtain the corresponding difference value.

[0027] Furthermore, the method for processing the feature area includes:

[0028] Acquire a feature region image, perform grayscale processing on the feature region image to form a corresponding feature region grayscale image, and divide the feature region grayscale image into a plurality of element blocks;

[0029] Identify the grayscale value of each pixel in each element block, and form an initial matrix of the element block according to the position of each grayscale value Among them, r ij represents the element value of the i-th column and the j-th row; i = 1, 2, ..., n, n is a positive integer; j = 1, 2, ..., m, m is a positive integer;

[0030] The initial matrix is ​​adjusted so that Then the adjusted initial matrix is ​​marked as the intermediate matrix, and the intermediate matrix is

[0031] The intermediate matrices corresponding to the element blocks are merged to form feature units corresponding to the feature region image.

[0032] The data verification scheduling module is used to perform data verification scheduling during access control face recognition, obtain a face matching image, match the user end of the corresponding user according to the face matching image, send the face matching image to the matched user end, and receive the face verification results sent by the face verification module in each matched user end in real time, and send the face verification results to the access control system.

[0033] Furthermore, the method of matching the corresponding user end according to the face matching image includes:

[0034] Performing feature processing on the face matching image to obtain face matching image features;

[0035] Determine the matching units corresponding to the matching features of each user in the access control database according to the face matching image features;

[0036] Calculate the similarity value between the matching unit and the corresponding feature unit, and mark the user matching features whose similarity values ​​are all greater than the threshold value X2 as matching verification features;

[0037] The user end corresponding to each matching verification feature is identified and marked as the user end of the face matching image matching.

[0038] Furthermore, the similarity value calculation method includes:

[0039] Identify the matching units and the matching intermediate matrices composed of the corresponding feature units and the characteristic intermediate matrix

[0040] According to the formula and formula Calculate the representative value A and representative value Q corresponding to the matching intermediate matrix and the characteristic intermediate matrix respectively;

[0041] Replace the corresponding matching intermediate matrix and the characteristic intermediate matrix according to the representative value A and the representative value Q; obtain the corresponding matching similarity data and characteristic similarity data;

[0042] The corresponding representative values ​​in the matching similarity data and feature similarity data are marked as A s and Q s ; s = 1, 2, ..., v, v is a positive integer;

[0043] According to the similarity formula Calculate the corresponding similarity value; where: XS is the similarity value.

[0044] The face verification module is used to perform face verification, obtain the face matching image received by the user end, identify the user's target face image, verify the face matching image according to the target face image, and obtain the corresponding face verification result; send the face verification result to the data verification scheduling module;

[0045] When the corresponding face verification result is obtained, the face matching image is deleted, and the user cannot view the face matching image through the user terminal.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] Through the cooperation among the storage data processing module, the user terminal and the data verification scheduling module, the user face data stored in the community access control is securely processed, thereby avoiding the leakage of user face data stored in the existing access control system; and after the security processing, the user's face image is no longer stored in the access control system, while avoiding the leakage of the user's face image from the property owner; the verification process is migrated to the user side, thereby improving the user's confidence in using face recognition; and when the face verification result is obtained, the face matching image is deleted, and the user cannot view the face matching image through the user terminal, thereby protecting the personal privacy of the verification personnel; at the same time, because of the user-side matching process, it will effectively prevent some people from cracking the accumulated face information of the verification personnel after registering as a user. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0049] Figure 1 This is a principle block diagram of the present invention. DETAILED DESCRIPTION

[0050] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] like Figure 1 As shown, a security management system for facial image storage data includes a storage data processing module, a user terminal, and a data verification scheduling module;

[0052] The user terminal is used by users and is generally set up in the form of a mini program or an APP. The user refers to the residents in the community; it includes a face data management module and a face verification module;

[0053] The facial data management module is used to manage the corresponding facial information of the user, determine the user's facial image, form the user's target facial image, and generate the corresponding user matching features based on the target facial image, and send the obtained user matching features to the access control database; the access control database is a database for storing facial image data in the access control system.

[0054] The method for determining the user's face image includes:

[0055] Receive and store facial image data sent by the data processing module in real time, and determine the user's facial image based on the received facial image data; this situation is aimed at residents who have previously applied the facial recognition function in the access control system, and intelligently determine the user's facial image based on their previously recorded facial image data, without the user having to record it again; mark the determined user's facial image as the target facial image;

[0056] When the facial image data sent by the storage data processing module is not received, it means that the user has not entered facial information into the access control system before; the user enters the facial image by himself, and after successful entry, the user's facial image is determined and marked as the target facial image.

[0057] The method for setting user matching features includes:

[0058] Identify the target face image of the user;

[0059] Determine a number of feature areas according to the target face image, where the feature area refers to the facial area that can represent the user; determine the various face recognition areas according to the existing face recognition principle, mark them as candidate areas, determine the target candidate combination according to the combination difference of each candidate area, and mark the candidate area in the target candidate combination as a feature area;

[0060] The combination difference refers to analyzing whether the combination of the selected areas can be distinguished from other facial information after the combination of the selected areas. The benchmark is the user matching features stored in the access control database, that is, whether the user matching features corresponding to the combination are different from the user matching features stored in the access control database. When it is judged to be different from other user matching features, the corresponding combination is marked as the initial combination. According to the positions of the selected areas corresponding to the initial combination, the difference value with the stored user matching features is estimated, that is, the estimation is made according to the positions and numbers of the corresponding selected areas, because some stored user matching features may correspond to different numbers and positions of selected areas, and corresponding factors will be considered in the subsequent matching verification. Therefore, in order to reduce the amount of subsequent data analysis, further refinement is carried out, and all initial combinations with difference values ​​greater than the threshold value X1 are marked as intermediate combinations, the number of selected areas in each intermediate combination is identified, and the intermediate combination with the fewest selected areas is selected as the target selected combination; if there are multiple intermediate combinations with the least number of selected areas, the intermediate combination with the largest difference value is selected as the target selected combination;

[0061] Each feature area is processed to form a feature unit, and each feature unit is combined into a user matching feature.

[0062] The difference value is calculated as follows:

[0063] Processing is performed according to the initial combination to form a corresponding simulated user matching feature, that is, processing is performed according to the method of obtaining the user matching feature; identifying the facial position corresponding to each feature unit in the user matching feature;

[0064] Identify the stored user matching features that need to be calculated for difference values, and mark them as verified user matching features;

[0065] Determine, according to the identified position, the verification unit in the verification user matching feature corresponding to each feature unit, that is, the feature unit in the verification user matching feature at the same position as the feature unit; if there is no feature unit at that position in the verification user matching feature, then the position can be a feature unit preset at any corresponding position, such as feature units at each position preset according to various existing face data, feature units at corresponding positions in other stored user matching features, that is, there are multiple possibilities for the position verification unit, but the feature unit corresponding to the smallest difference value finally obtained;

[0066] The unit difference between each feature unit and the corresponding verification unit is calculated, and the element differences are accumulated to form a unit difference, and then the unit differences are accumulated to form a corresponding difference value.

[0067] That is, identify the matrix corresponding to the feature unit and the matrix corresponding to the verification unit, calculate the difference between the corresponding elements, take the absolute value, and mark it as the element difference K t , where t = 1, 2, ..., c, c is a positive integer, the number of elements in the matrix; then according to the formula DC = ∑K t Calculate the corresponding unit difference, where DC is the unit difference, and add up the unit differences to obtain the corresponding difference value; for multiple possible verification units, take the one with the smallest difference value.

[0068] The methods for processing each feature area include:

[0069] Acquire a feature region image, perform grayscale processing on the feature region image to form a corresponding feature region grayscale image, and divide the feature region grayscale image into a number of element blocks of the same size, that is, the number of elements in each element block is the same and the size is the same; calculate according to the actual size of the feature region grayscale image, the number of elements and the preset element block size range, and if there are multiple sizes to choose from, select the largest size as the corresponding size of the element block, wherein the preset element block size range is set by corresponding professionals;

[0070] Identify the grayscale value of each pixel in each element block, and form the initial matrix of the element block according to the position of each grayscale value Among them, r ijrepresents the element value of the i-th column and the j-th row, that is, the gray value of the corresponding position; i = 1, 2, ..., n, n is a positive integer; j = 1, 2, ..., m, m is a positive integer;

[0071] Adjust the initial matrix so that Then mark the adjusted initial matrix as the intermediate matrix ∑r ij is the sum of the element values ​​corresponding to each element in the initial matrix;

[0072] The intermediate matrices corresponding to each element block are merged to form a feature unit corresponding to the feature area image. The feature unit is a large matrix including each element formed by combining the intermediate matrices according to the corresponding positions of the corresponding element library.

[0073] The stored data processing module is used to process the facial image data stored in the access control system, identify the users corresponding to each stored facial image data, and send the corresponding facial image data to the facial data management module of the corresponding user. When it is recognized that the stored facial image data has the user matching features of the corresponding user, the corresponding facial image data stored in the access control database is deleted.

[0074] The data verification scheduling module is used to perform data verification scheduling during access control face recognition, obtain the face image recognized by the access control, mark it as a face matching image, and process the face matching image according to the setting method of the feature area, that is, process all the to-be-selected areas in the face matching image in a corresponding manner to obtain the face matching image features;

[0075] Position and match the user matching features stored in the access control database with the face matching image features, that is, determine the feature units in the face matching image features that need to be matched and analyzed according to the selected area corresponding to the user matching features, mark them as matching units, and form each matching unit;

[0076] Calculate the similarity between each matching unit and each corresponding feature unit in the user matching feature, and mark the user matching features whose similarity values ​​are greater than the threshold X2 as matching verification features;

[0077] The user terminal identifies the user corresponding to each matching verification feature, sends the face matching image to the corresponding user terminal, and receives the face verification results sent by the face verification module in each corresponding user terminal in real time. The specific verification object can be determined based on the obtained face verification results. If all of them fail, the door cannot be opened; the obtained face verification results are sent to the access control system, and the access control system subsequently performs access control based on the face verification results.

[0078] The calculation of similarity values ​​can be performed as follows:

[0079] One is: based on the existing CNN network or DNN network and other neural networks, a corresponding similarity analysis model is established, and a corresponding training set is established manually for training. The training set includes input data and output data. The input data is a combination of two matrices with equal rows and columns set in a simulation, and the output data is a similarity value. The matching unit and the feature unit are analyzed by the similarity analysis model after successful training to obtain the corresponding similarity value. Because the neural network is the existing technology in this field, the specific establishment and training process is not described in detail in the present invention.

[0080] The other is to identify the intermediate matrices composed of the matching unit and the corresponding feature unit, and mark them as matching intermediate matrices and the characteristic intermediate matrix According to the formula and formula Calculate the representative value A and representative value Q corresponding to the matching intermediate matrix and the characteristic intermediate matrix respectively;

[0081] Use the corresponding representative value A and the representative value Q to replace the corresponding matching intermediate matrix and the feature intermediate matrix in the matching unit and the corresponding feature unit;

[0082] Marking the replaced matching units and feature units as matching similarity data and feature similarity data respectively;

[0083] The corresponding representative values ​​in the matching similarity data and feature similarity data are marked as A s and Q s ; s = 1, 2, ..., v, v is a positive integer;

[0084] According to the similarity formula Calculate the corresponding similarity value; where: XS is the similarity value.

[0085] The face verification module is used to perform face verification, obtain the face matching image received by the user end, identify the user's target face image, verify the face matching image according to the obtained target face image, and obtain the corresponding face verification result; and when the corresponding face verification result is obtained, the obtained face matching image is deleted, and the user cannot view the face matching image through the user end; the obtained face verification result is sent to the data verification scheduling module.

[0086] The face matching image verification is performed based on the target face image and is implemented using existing face recognition technology.

[0087] The above formulas are all calculated by removing dimensions and taking numerical values. The formula is a formula that is closest to the actual situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained by simulating a large amount of data.

[0088] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A security management system for storing facial image data, characterized in that: It includes storage data processing module, user terminal and data verification scheduling module; The stored data processing module is used to process the facial image data stored in the access control database, identify the user corresponding to each facial image data stored in the access control database in real time, and send the facial image data to the user terminal of the corresponding user; When a user matching feature of a corresponding user is identified in the access control database, the facial image data of the user corresponding to the user matching feature is deleted from the access control database; The user terminal is used by users and includes a face data management module and a face verification module; The face data management module is used to manage the corresponding face information of the user, determine the target face image of the user, generate the corresponding user matching features based on the target face image, and send the user matching features to the access control database; The data verification scheduling module is used to perform data verification scheduling during access control face recognition, obtain the face image recognized by the access control, mark it as a face matching image, match the user end of the corresponding user according to the face matching image, send the face matching image to the matched user end, and receive the face verification result sent by the face verification module in each matched user end in real time, and send the face verification result to the access control system; The face verification module is used to perform face verification, obtain the face matching image received by the user end, identify the user's target face image, verify the face matching image according to the target face image, and obtain the corresponding face verification result; send the face verification result to the data verification scheduling module; After obtaining the face verification result, the face matching image is deleted, and the user cannot view the face matching image through the user terminal.

2. A security management system for storing facial image data according to claim 1, characterized in that: In the face data management module, the method for determining the target face image includes: Receiving the facial image data sent by the storage data processing module in real time, determining the user facial image according to the received facial image data; marking the determined user facial image as a target facial image; When the facial image data sent by the storage data processing module is not received, the user enters the user's facial image by himself. When the user's facial image is successfully entered, the entered user's facial image is marked as a target facial image.

3. A security management system for storing facial image data according to claim 2, characterized in that: The method for setting user matching features includes: Identify a target face image of a user; determine a plurality of feature regions according to the target face image; process each of the feature regions to form a feature unit, and combine each of the feature units into a user matching feature.

4. A security management system for storing facial image data according to claim 3, characterized in that: The method of setting the feature area includes: Marking each to-be-selected area in the target face image, and determining each corresponding initial combination based on the user matching features already stored in the access control database; Performing feature processing on each of the initial combinations to obtain corresponding simulated user matching features; Calculating the difference between the simulated user matching feature and the matching features of each user already stored in the access control database; Marking the initial combinations whose difference values ​​are all greater than the threshold value X1 as intermediate combinations; Identify the number of to-be-selected areas in each of the intermediate combinations, and select the intermediate combination with the least number of to-be-selected areas as the target to-be-selected combination; If there are multiple intermediate combinations with the least number of areas to be selected, the intermediate combination with the largest difference value is selected as the target combination to be selected.

5. A security management system for storing facial image data according to claim 4, characterized in that: The difference value is calculated by: Determine the corresponding feature units and verification units according to the simulated user matching features for calculating the difference values ​​and the stored user matching features; Calculate the corresponding element difference according to the feature unit and the verification unit, and mark the element difference as Kt, where t=1, 2, ..., c, and c is a positive integer; According to the formula DC = ∑K t Calculate the corresponding unit difference, where DC is the unit difference; The difference values ​​of each unit are accumulated to obtain the corresponding difference value.

6. A security management system for storing facial image data according to claim 5, characterized in that: Methods for processing feature areas include: Acquire a feature region image, perform grayscale processing on the feature region image to form a corresponding feature region grayscale image, and divide the feature region grayscale image into a plurality of element blocks; Identify the grayscale value of each pixel in each element block, and form an initial matrix of the element block according to the position of each grayscale value ; Among them, r ij Represents the element value of the i-th column and j-th row; i=1, 2, ..., n, n is a positive integer; j=1, 2, ..., m, m is a positive integer; The initial matrix is ​​adjusted so that ; then the adjusted initial matrix is ​​marked as the intermediate matrix, and the intermediate matrix is ; The intermediate matrices corresponding to the element blocks are merged to form feature units corresponding to the feature region image.

7. A security management system for storing facial image data according to claim 6, characterized in that: In the data verification scheduling module, the method of matching the corresponding user end according to the face matching image includes: Performing feature processing on the face matching image to obtain face matching image features; Determine the matching units corresponding to the matching features of each user in the access control database according to the face matching image features; Calculate the similarity value between the matching unit and the corresponding feature unit, and mark the user matching features whose similarity values ​​are all greater than the threshold value X2 as matching verification features; The user end corresponding to each matching verification feature is identified and marked as the user end of the face matching image matching.

8. A security management system for storing facial image data according to claim 7, characterized in that: The similarity value calculation method includes: Identify the matching units and the corresponding feature units in the matching intermediate matrix and the characteristic intermediate matrix ; According to the formula and formula Calculate the representative value A and representative value Q corresponding to the matching intermediate matrix and the characteristic intermediate matrix respectively; Replace the corresponding matching intermediate matrix and the characteristic intermediate matrix according to the representative value A and the representative value Q; obtain the corresponding matching similarity data and characteristic similarity data; The corresponding representative values ​​in the matching similarity data and feature similarity data are marked as A s and Q s ; s = 1, 2, ..., v, v is a positive integer; According to the similarity formula Calculate the corresponding similarity value; where: XS is the similarity value.

Images