A method, apparatus, and storage medium for distributing facial data across different devices.

Abstract

This invention relates to a method, apparatus, and storage medium for distributing facial data across different devices. The method includes: before performing facial verification, a verification device sends its device information and first reference information to a facial database; simultaneously, the verification device sends comparison information to a verification module; the facial database retrieves stored facial data based on the first reference information, and simultaneously generates first verification information based on the device information and facial data, sending it to the verification module; then, it generates complete facial verification data based on the device information and packages it; the verification module verifies the similarity between the comparison information and the first verification information; when the similarity is greater than or equal to a preset similarity threshold, it sends a verification pass result back to the facial database; and the facial database then distributes the packaged facial verification data to the verification device. This invention, through the refinement and separation of facial data distribution steps, provides a face recognition system composed of multiple devices with improved security, timeliness, and accuracy in facial data transmission.

Description

Technical Field

[0001] This invention relates to the field of facial recognition technology, specifically to a method, apparatus, and storage medium for distributing facial data across different devices. Background Technology

[0002] With the continuous development of facial recognition technology, it has been widely used in residential communities, campuses, office parks, and other places with high population density and mobility for access control. However, facial recognition devices are limited by the manufacturing companies, and their underlying algorithms for recognition and verification are not the same. Moreover, facial recognition technology continues to evolve over time. Even facial recognition devices developed by the same manufacturer may have changes in the areas where facial features are extracted, the extraction algorithms, and the verification areas and algorithms, making it difficult to uniformly manage and verify facial data.

[0003] Conventionally, for the deployment of scattered facial recognition devices, a one-to-one approach can be used, with one facial database and one verification device corresponding to each device. This ensures a one-to-one correspondence between facial data collection and verification features, guaranteeing consistency between the collected and verified features. However, with the continuous development of smart cities, the volume of facial data verification and collection is enormous, rendering this scattered deployment approach unsuitable. A unified collection and management of facial data is needed. For example, if a park uses device A for accommodations and device B for offices, the different verification methods at locations A and B will result in different facial recognition and verification methods. Therefore, a separate facial database needs to be built at location A for verification by the devices there, and another database at location B for verification by the devices there. Users also need to upload their facial information twice. Perhaps, in a facial recognition network with a small number of devices, this one-to-one facial verification network setup is acceptable. As the number of different devices and both new and old devices in a facial recognition network continues to increase, a considerable number of facial databases will need to be built. Furthermore, the facial data in these databases is not shared, leading to data redundancy and barriers to facial recognition data collection and management. Clearly, this conventional one-to-one facial recognition access control system is not suitable for situations requiring large volumes of facial data and unified management. Summary of the Invention

[0004] This invention addresses the need for unified collection and management of facial data from multiple devices in existing technologies. It provides a method, apparatus, and storage medium for distributing data across different devices, enabling unified management of facial data and adapting to facial recognition systems that support both new and old verification devices.

[0005] In a first aspect, the present invention provides a method for distributing facial data across different devices, the method comprising:

[0006] Before performing face verification, the verification device sends the device information and the first reference information to the face database, and at the same time, the verification device sends the comparison information to the verification module.

[0007] The face database retrieves stored face data based on the first reference information, and simultaneously generates first verification information based on the device information and face data and sends it to the verification module. Then, it generates complete face verification data based on the device information and packages it.

[0008] The verification module verifies the similarity between the comparison information and the first verification information. When the similarity is greater than or equal to a preset similarity threshold, it sends the verification pass result back to the face database. The face database then sends the packaged face verification data to the verification device.

[0009] Specifically, the concept of this embodiment is to separate the face database and the verification transmission of the verification device through the calibration module, thereby ensuring the centralization, timeliness, and security of the face data in the face database. Simultaneously, a preliminary verification method using first reference information and first calibration information is employed to improve the accuracy and flexibility of the face verification data in the face database.

[0010] It is easy to understand that the device information refers to the device model and device code information of the current verification device. Based on this device information, the face database can call the same face verification and recognition algorithm as the current verification device to process the face data stored in the database, generating face verification data that can be used for recognition by the current verification device. This solves the technical problem of face verification failure caused by inconsistencies in face recognition algorithms between different verification devices or between new and old verification devices.

[0011] The first reference information includes only the baseline feature points extracted from face data retrieval, enabling the face database to retrieve the face data stored in the database based on these baseline feature points and find the face data to be verified.

[0012] The comparison information is feature information that can be used for preliminary comparison of face similarity after simple processing by the current verification device. It is used by the calibration module for preliminary comparison of the similarity of face data extracted from the face database, thereby improving the accuracy of the complete face verification data sent from the face database to the verification device.

[0013] The first verification information is the feature part information that is the same as the feature part information of the comparison information. It is the feature part information that can be used for preliminary comparison of face similarity after simple processing of face data from the face database using the same recognition algorithm as the current verification device.

[0014] The preset similarity threshold is the minimum acceptable similarity for face similarity correction.

[0015] Another conceptual point of this embodiment is that the face recognition verification steps are decomposed through the interaction of different units. In reality, after capturing a face, the verification device needs to parse the face to be verified, generating complete face data to be verified. The face database also needs to retrieve the face data stored in the database based on the face data to be verified, and process the face data to generate face verification data that can be parsed and compared by the current verification device. The face data to be verified parsed by the verification device is then compared with the face verification data processed by the face database according to the verification device's processing program to complete the face verification. However, if communication only occurs after the current verification device and the face database have completely completed the parsing and generation of face data, it will obviously lead to difficulties in meeting the timeliness requirements of verification, resulting in verification delays. Therefore, this embodiment refines and separates the face verification steps through the calibration module, enabling the response speed of a multi-device face verification system to be comparable to that of a one-to-one device system, thereby ensuring the responsiveness of face verification. It is understandable that the face database stores all face data. Only when a verification device generates a verification request will the face database retrieve the face data of the face to be verified according to the request of that verification device and process it according to the corresponding recognition algorithm to generate face verification data. When multiple verification devices request the face database at the same time, the complete verification steps obviously cannot meet the requirements of timely verification. Therefore, in this embodiment, before performing face verification, the verification device first sends the first reference information obtained from preliminary parsing along with the device information to the face database, and sends the comparison information obtained from preliminary parsing to the verification module. While the face database and the verification module receive the information and begin processing their own tasks, the verification device simultaneously performs deep parsing on the face to be recognized, generates complete face data to be recognized, and waits for the face database to send the face verification data to complete the face verification. After receiving the first reference information and device information, the face database first retrieves face data, then calls the recognition algorithm of the requested verification device to perform preliminary processing on the face data. The preliminary processing of the first verification information is sent to the verification module. Simultaneously, while the verification module is performing verification, the face database generates and packages the face verification data according to the recognition algorithm of the requested verification device. Once the verification module returns a verification pass result to the face database based on the comparison information and the first verification information, the packaged face verification data is immediately sent to the verification device. Therefore, by decomposing and parallelizing the verification steps through the verification device, face database, and verification module, the timeliness of face verification in this embodiment is improved.It is understood that the verification device and the calibration module can generate corresponding task processing columns according to the request information of different verification devices, and process the face verification data of multiple faces to be verified at the same time, so that the verification speed of the method provided in this embodiment is comparable to that of one-to-one face verification, and the accuracy of face extraction is improved by the calibration module.

[0016] In another embodiment of the first aspect, the method further includes second reference information;

[0017] If the similarity between the comparison information and the first verification information is less than the preset similarity threshold, a verification failure result is sent to both the face database and the verification device.

[0018] Based on the verification failure result, the verification device will send second reference information to the face database;

[0019] The face database retrieves stored face data again based on the verification failure result using the first reference information and the second reference information. At the same time, it generates second verification information based on the device information and face data and sends it to the verification module. Then, it generates complete face verification data based on the device information and packages it.

[0020] The verification module verifies the similarity between the comparison information and the second verification information. When the similarity is greater than or equal to a preset similarity threshold, it sends the verification result back to the face database. The face database then sends the packaged face verification data to the verification device.

[0021] Specifically, when retrieving and extracting facial data from the facial database using the first reference information provided in this embodiment, since the facial data is retrieved based on representative feature points, when the facial data in the database reaches a certain size, some faces inevitably have high similarity, leading to errors in the facial data retrieved based on the first reference information. Therefore, this embodiment avoids such highly similar facial data.

[0022] The second reference information includes only the baseline feature points extracted from face data retrieval, and these baseline feature points are different from the baseline feature points of the first reference information. This enables the face database to retrieve the face data stored in the database based on the baseline feature points included in the first and second reference information, and find face data that better meets the requirements for verification.

[0023] The second verification information is the feature part information that is the same as the feature part of the comparison information. It is the feature part information that can be used for preliminary comparison of face similarity after simple processing of face data from the face database using the same recognition algorithm as the current verification device.

[0024] It is easy to understand that the above operations can avoid the scenario where the first reference information may become invalid due to the large volume of face data in the face database and the presence of highly similar faces, thereby improving the accuracy of the face data retrieved from the face database.

[0025] In another implementation of the first aspect, the first reference information and the second reference information are both at least two or more feature points of face data, and the feature points contained in the first reference information and the second reference information do not overlap.

[0026] Specifically, both the first reference information and the second reference information are representative feature points on a human face, and the feature points included in the first reference information are different from those included in the second reference information. The first reference information includes a set of feature points essential for distinguishing faces, while the second reference information includes feature points that supplement the set of facial feature points based on the first reference information, further distinguishing faces. Therefore, the use of the first and second reference information improves the accuracy of facial data retrieved from the facial database.

[0027] In another implementation of the first aspect, when the verification device receives a verification failure result, it collects the face data of the failed verification and the face verification data of the verified face, and updates the feature points of the first reference information and the second reference information through machine learning.

[0028] Specifically, when the face data extracted from the face database based on the first reference information fails the verification by the verification module, it indicates that the feature point set provided by the first reference information can no longer meet the practical requirements of the face verification system. In this case, it is necessary to update the feature point set included in the first reference information to improve the accuracy of the face data retrieved from the face database based on the first reference information.

[0029] It is easy to understand that the new set of feature points included in the first reference information is obtained through a feature point training model. The feature point training model is trained based on multiple first reference information samples, corresponding second reference information samples, and corresponding face data. When the first verification information generated from the face data retrieved from the face database using the first reference information fails verification by the verification module, the verification device, upon receiving a verification failure result, will collect the failed face data and the verified face data, and update the feature points of the first and second reference information through machine learning.

[0030] In another implementation of the first aspect, the comparison information includes at least two or more corresponding feature parts of the eyes, nose, ears, face, and mouth, which are composed of multiple feature points;

[0031] Both the first and second proofreading information include multiple feature points of the feature parts contained in the comparison information, which are used to compare similarity with the comparison information.

[0032] Specifically, the verification module primarily verifies the facial data extracted from the facial database against the facial data to be identified based on the feature parts of the face. Since users may experience obstruction due to masks, glasses, or hair when passing through the verification device, to improve the accuracy and timeliness of the verification module, it includes at least two or more feature parts corresponding to the eyes, nose, ears, face, and mouth, composed of multiple feature points. Furthermore, based on the degree of obstruction, the module preferentially selects the unobstructed parts as the feature parts for verification.

[0033] In another implementation of the first aspect, the verification device sends a first time tag when sending device information and first reference information to the face database; and the verification device sends a second time tag when sending comparison information to the verification module.

[0034] Based on the first time tag, when the face verification data delivery time is greater than or equal to the preset first time threshold, the priority of the corresponding processing item is increased;

[0035] Alternatively, based on the second time tag, when the feedback time of the verification result is greater than or equal to the preset second time threshold, the priority of the corresponding processing item is increased.

[0036] Specifically, although the invention refines and separates the face recognition steps through the verification device, face database, and calibration module, improving the timeliness of the multi-device face recognition system during the recognition process, when the number of different devices in the multi-device face recognition system is too large, and multiple devices simultaneously send data download requests to the face database, the face database will be unable to simultaneously satisfy the data download requests from multiple different devices, resulting in data verification latency. Therefore, when the verification device sends device information and first reference information to the face database, it will also send a first time tag to index the face verification time of the current verification device. When the face verification data download time is greater than or equal to a preset first time threshold, it indicates that the face verification latency cannot meet the verification expectations of the current verification device, and the priority of the current processing item is increased. In addition, when the verification device sends comparison information to the calibration module, it will also send a second time tag to index the face verification time of the current verification device. When the feedback time of the calibration pass result is greater than or equal to a preset second time threshold, the priority of the corresponding processing item is increased.

[0037] In another implementation of the first aspect, when the number of verification devices reaches a certain level, the number of retrieval units in the face database and the number of verification modules for verification will be expanded. When the processing of the retrieval units and verification modules is overloaded, multiple retrieval units and verification modules can be enabled to process face data simultaneously, thereby improving the concurrency and synchronous processing capabilities of the multi-device face verification system.

[0038] Secondly, the present invention provides an apparatus for distributing facial data across different devices, the apparatus comprising:

[0039] The transceiver unit is used to receive device information and first reference information sent by the verification device before performing face verification, as well as to receive the verification pass result sent by the verification module; and to send the face data retrieved and stored according to the first reference information to the verification module, while generating the first verification information according to the device information and face data, and sending the packaged face verification data sent by the verification device after receiving the verification pass result.

[0040] The retrieval unit is used to retrieve stored face data from the face database based on first reference information and / or second reference information.

[0041] The generation unit is used to generate first verification information based on device information and first reference information, and to generate complete face verification data based on device information.

[0042] Specifically, the face database, the verification module, and the verification device all include a transceiver unit for sending and receiving information between them; the retrieval unit and the generation unit are located in or connected to the face database and can directly operate on the face data in the face database.

[0043] Therefore, the face database includes all face data from face recognition systems composed of different verification devices, and neither the verification devices nor the verification module have permission to access the face data in the face database, thus improving the centralized management and security of face data. Furthermore, by refining and separating the steps of the verification devices, face database, and verification module, the data extraction, verification, and distribution process during face recognition can be shortened, improving the timeliness of multi-device face recognition systems in responding to verification requests.

[0044] In another embodiment of the second aspect, the retrieval unit and the generation unit can be integrated into a single module, reducing the complexity of the system.

[0045] In another embodiment of the second aspect, the retrieval unit includes at least two or more groups; the proofreading module includes at least two or more groups.

[0046] When the verification device sends the device information and the first reference information to the face database, it will send a first time tag; and when the verification device sends the comparison information to the verification module, it will send a second time tag.

[0047] Based on the first time tag, when the face verification data delivery time is greater than or equal to the preset first time threshold, the priority of the corresponding processing item is increased, and at least two or more of the retrieval units are activated to prioritize the processing item according to the priority.

[0048] Alternatively, based on the second time tag, when the feedback time of the proofreading result is greater than or equal to the preset second time threshold, the priority of the corresponding processing item is increased, and at least two or more of the proofreading modules are activated to prioritize the processing item according to the priority.

[0049] Specifically, by establishing a priority system, the timeliness of sudden multi-person face verification can be improved, ensuring the effectiveness of face verification.

[0050] Thirdly, embodiments of the present invention provide a terminal, which includes a processor, a memory, and a communication interface; the memory stores a computer program; when the processor executes the computer program, the communication interface is used to send and / or receive data, and the terminal can execute the method described in the first aspect or any possible implementation of the first aspect.

[0051] It should be noted that the processor included in the terminal described in the third aspect above can be a processor specifically designed to execute these methods (referred to as a dedicated processor for distinction), or a processor that executes these methods by calling a computer program, such as a general-purpose processor. Optionally, at least one processor may include both dedicated and general-purpose processors.

[0052] Optionally, the computer program described above can be stored in memory. For example, the memory can be a non-transitory memory, such as read-only memory (ROM), which can be integrated with the processor on the same device or disposed on different devices. This embodiment of the invention does not limit the type of memory or the arrangement of the memory and processor.

[0053] In one possible implementation, at least one of the aforementioned memories is located outside the aforementioned terminal.

[0054] In yet another possible implementation, at least one of the aforementioned memories is located within the aforementioned terminal.

[0055] In another possible implementation, a portion of the memory of the at least one memory is located inside the terminal, while another portion of the memory is located outside the terminal.

[0056] In this invention, the processor and memory may also be integrated into a single device, that is, the processor and memory can be integrated together.

[0057] Fourthly, embodiments of the present invention provide a server, which includes a processor, a memory, and a communication interface; the memory stores a computer program; when the processor executes the computer program, the communication interface is used to send and / or receive data, and the server can perform the methods described in the second aspect or any possible implementation of the second aspect.

[0058] It should be noted that the processor included in the server described in the fourth aspect above can be a processor specifically designed to execute these methods (referred to as a dedicated processor for distinction), or a processor that executes these methods by calling computer programs, such as a general-purpose processor. Optionally, at least one processor may include both dedicated and general-purpose processors.

[0059] Optionally, the computer program described above can be stored in memory. For example, the memory can be a non-transitory memory, such as read-only memory (ROM), which can be integrated with the processor on the same device or disposed on different devices. This embodiment of the invention does not limit the type of memory or the arrangement of the memory and processor.

[0060] In one possible implementation, at least one of the aforementioned storage devices is located outside the aforementioned server.

[0061] In yet another possible implementation, at least one of the aforementioned storage devices is located within the aforementioned server.

[0062] In another possible implementation, a portion of the memory of the at least one memory is located within the server, while another portion of the memory is located outside the server.

[0063] In this invention, the processor and memory may also be integrated into a single device, that is, the processor and memory can be integrated together.

[0064] Fifthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program that, when executed on at least one processor, implements the method described in the first aspect or any of the optional solutions of the first aspect.

[0065] In a sixth aspect, the present invention provides a computer program product comprising a computer program that, when run on at least one processor, implements the method described in the first aspect or any of the alternative solutions of the first aspect.

[0066] Optionally, the computer program product can be a software installation package, which can be downloaded and executed on a computing device when the aforementioned method is required. Attached Figure Description

[0067] The present invention will be further described in detail below with reference to the accompanying drawings and preferred embodiments. However, those skilled in the art will understand that these drawings are drawn only for the purpose of explaining the preferred embodiments and therefore should not be construed as limiting the scope of the invention. Furthermore, unless specifically indicated, the drawings are only schematic representations of the composition or structure of the described objects and may contain exaggerated depictions, and the drawings are not necessarily drawn to scale.

[0068] Figure 1 A schematic diagram of the architecture of the method for distributing face data across different devices provided in this embodiment of the invention;

[0069] Figure 2 A flowchart illustrating a method for distributing face data across different devices, provided in an embodiment of the present invention;

[0070] Figure 3 Another flowchart illustrating the method for distributing face data across different devices provided in this embodiment of the invention;

[0071] Figure 4 A schematic diagram of a device for distributing face data across different devices provided in an embodiment of the present invention;

[0072] Figure 5 This invention provides a schematic diagram of a terminal for distributing face data across different devices. Detailed Implementation

[0073] The following is in conjunction with the appendix Figures 1 to 5 The present invention will be described in detail below.

[0074] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0075] Please see Figure 1 As shown, Figure 1 A schematic diagram of the architecture of the method for distributing face data across different devices provided in this embodiment of the invention.

[0076] This embodiment constructs a face recognition network using a face database 101, a verification module 102, and multiple different verification devices 103a, 103b, 103c, and 103d. The face database 101 stores all uploaded face data; the verification module 102 verifies the data sent during face recognition; and the verification devices 103a, 103b, 103c, and 103d collect and verify the data of the face to be verified.

[0077] In fact, the facial recognition system built using this architecture is compatible with almost all manufacturers of verification devices on the market, thereby improving the system's compatibility and applicability. This is because, for smart city recognition systems in parks, campuses, community clusters, or cities with numerous recognition locations, the brands and models of verification devices vary significantly depending on the scenario and the passage of time. Different brands have different verification algorithms, leading to discrepancies in facial verification results. Even within the same brand, there are differences in model and age, further causing recognition discrepancies. These differences result in false positives and incorrect judgments during facial recognition verification. Conventionally, if the facial database, verification module, and verification device are matched one-to-one, the number of databases and verification modules will surge, significantly increasing costs. Furthermore, different databases may contain multiple identical facial data entries, leading to data redundancy. Building a facial recognition system consisting of a facial database and verification devices requires complex processes such as data verification and distribution, resulting in a large data volume and impacting the verification device's latency. Additionally, the different recognition algorithms of each verification device increase the risk of verification failure.

[0078] To address the aforementioned issues, the architecture and method provided in this embodiment effectively solve the problems. The face database 101 primarily stores all uploaded face data and handles simple face search and verification data generation, reducing data processing difficulty and workload, and ensuring that the face database 101's processing capacity can meet the simultaneous verification requests from different verification devices. Furthermore, before initiating a data distribution request, each verification device simultaneously sends its device information, including device model and device coding information, which corresponds to the recognition algorithm and location used by the verification device. This allows the face database 101 to call the corresponding verification device's recognition algorithm to process the face data stored in the face database 101 and generate corresponding face verification data, thus resolving the verification misjudgment problem caused by different verification device recognition algorithms. Simultaneously, the verification module 102 is used for preliminary correction of the face information in the face database and on the verification devices, improving the accuracy of the face database data. Therefore, the above architecture can achieve centralized storage of facial data, making the data concise and secure. On the other hand, the verification and validation processes of facial data are distributed across the verification devices and verification modules, allowing facial verification operations to be processed separately on different devices, thereby ensuring the timeliness and accuracy of facial verification.

[0079] Please see Figure 2 , Figure 2 A flowchart illustrating a method for distributing facial data across different devices, provided in an embodiment of the present invention.

[0080] After acquiring the face to be recognized, the verification device 103 performs preliminary face analysis. In step S210a, it sends device information and first reference information to the face database 101, and simultaneously in step S210b, it sends comparison information to the verification module 102. For detailed explanations of features such as device information and first reference information, please refer to the instruction manual; they will not be repeated here. Steps S210a and S210b can be performed synchronously, thereby reducing the request latency of the verification device. Furthermore, after completing steps S210a and S210b, the verification device 103 continues with deep analysis of the face to be recognized, generating complete face data. This deep analysis operation does not interfere with the operations within the face database 101 and the verification module 102, and can be processed in parallel, thus shortening the face verification time and improving the user experience.

[0081] Furthermore, after receiving the device information and first reference information from the verification device, the face database 101 will execute step S230 to retrieve the stored face data according to the first reference information, and simultaneously generate first verification information based on the device information and face data. Since the first reference information only contains baseline feature points for face data retrieval, its parsing and processing comparison operations are extremely short, allowing for rapid location of the face data stored in the face database 101. The corresponding verification device's recognition algorithm is then invoked according to the device information to perform preliminary processing on the retrieved face data, generating first verification information for verification. Next, the face database 101 will execute step S231 to send the first verification information to the verification module, and execute step S232 to continue invoking the corresponding verification device's recognition algorithm based on the device information to generate complete face verification data and package it. Steps S231 and S232 can be executed simultaneously. After receiving the first verification information sent by the face database 101, the verification module 102 executes step S220 to compare the similarity between the comparison information and the first verification information. Steps S220 and S232 are independent and can be executed in parallel. After the comparison is completed, in step S221, if the similarity is greater than or equal to a preset similarity threshold, the verification pass result is fed back to the face database. After receiving the verification pass result, the face database 101 executes step S234 to send the packaged face verification data to the verification device 103. The verification device 103 can then verify the face data to be recognized that it has parsed with the received face verification data to complete face recognition.

[0082] Because the deep analysis operation of verification device 103, the verification operation of verification module 102, and the retrieval and generation operation of verification device 101 can be performed synchronously, the verification process of verification device 103 can be completed with almost no delay through the above-mentioned face verification execution process, improving the timeliness of recognition of a single face database corresponding to different verification devices. In addition, this embodiment simplifies the retrieval and verification information, shortening the processing time while ensuring the accuracy of face data extraction. Furthermore, the face database 101 only performs face retrieval and generation operations and does not directly call or modify the stored face data, ensuring the security of face data. The verification module only compares the similarity between the comparison information and the first verification information, without processing the data according to the recognition algorithm of the verification device, making its processing logic simple and its risk resistance strong.

[0083] In another possible implementation, the functions of retrieving, generating, and sending / receiving information in the face database 101 can be integrated into a single processing module, which can be independently encapsulated. This allows the face database 101 to communicate only with the independent processing module, further ensuring the security of the face data within the face database 101. This processing module communicates with the verification module and the authentication device.

[0084] In another possible implementation, the above processing module can be integrated with the proofreading module to reduce the complexity of the system.

[0085] Please see Figure 3 As shown, Figure 3 This invention provides another schematic flowchart of a method for distributing face data across different devices.

[0086] Although the first reference information simplifies the processing speed of the initial analysis and face database retrieval by the verification device, as the number of face data in the face database 101 increases, some face data inevitably have extremely high similarity, leading to false judgments by the verification device 103. Therefore, when the similarity between the comparison information and the first verification information is insufficient, the verification module 102 will execute step S222a, which sends a verification failure result to the face database 101 when the similarity is less than a preset similarity threshold, and step S222b, which sends a verification failure result to the verification device 103 when the similarity is less than the preset similarity threshold.

[0087] After the face database 101 accepts the verification failure result in step S235, it waits for the verification device 103 to accept the verification failure result in step S211. Then, it executes step S212 to send the second reference information to the face database. The face database then re-retrieves face data using the first and second reference information. Since the reference feature points included in the first and second reference information are different, similar faces can be effectively identified. Step S236 is then executed to generate the second verification information, followed by step S237 to send the second verification information to the verification module 102. After receiving the second verification information, the verification module 102 executes step S223 to compare the similarity between the comparison information and the second verification information. Then, step S224 is executed; if the similarity is greater than or equal to a preset similarity threshold, a verification pass result is fed back to the face database. After accepting the verification pass result in step S239, the face database 101 executes step S240 to send the repackaged face verification data to the verification device. Thus, the verification device 103 performs face recognition by parsing the completed face data to be verified and the sent face verification data.

[0088] It is easy to understand that the first reference information and the second reference information are merely baseline feature points used to search for face data, while the first verification information and the second verification information are feature parts processed by the recognition algorithm. Therefore, the verification information can be used to determine whether the face to be verified and the face data in the face database belong to the same person, thereby improving the accuracy of face recognition.

[0089] Through the above steps, this embodiment can address the risk of verification failure due to similar face data when the face database 101 contains a large amount of face data, thereby improving the accuracy of recognition and verification of the multi-device face verification system based on a single face database.

[0090] See Figure 4 As shown, Figure 4 A schematic diagram of a device for distributing face data across different devices provided in an embodiment of the present invention.

[0091] Specifically, the device 30 includes at least a transceiver unit 301, a retrieval unit 302, and a generation unit 303.

[0092] The transceiver unit 301 is used to receive device information and first reference information sent by the verification device before performing face verification, as well as to receive the verification pass result sent by the verification module; and to send the face data retrieved and stored according to the first reference information to the verification module, while generating the first verification information according to the device information and face data, and sending the packaged face verification data sent by the verification device after receiving the verification pass result.

[0093] The retrieval unit 302 is used to retrieve stored face data from the face database based on the first reference information and / or the second reference information.

[0094] The generation unit 303 is used to generate first verification information based on device information and first reference information, and to generate complete face verification data based on device information.

[0095] Specifically, the face database, the verification module, and the verification device all include a transceiver unit for sending and receiving information between them; the retrieval unit and the generation unit are located in or connected to the face database and can directly operate on the face data in the face database.

[0096] Therefore, the face database includes all face data from face recognition systems composed of different verification devices, and neither the verification devices nor the verification module have permission to access the face data in the face database, thus improving the centralized management and security of face data. Furthermore, by refining and separating the steps of the verification devices, face database, and verification module, the data extraction, verification, and distribution process during face recognition can be shortened, improving the timeliness of multi-device face recognition systems in responding to verification requests.

[0097] In another possible embodiment, the retrieval unit and the generation unit can be integrated into a single module, reducing the complexity of the system.

[0098] In yet another possible embodiment, the retrieval unit includes at least two or more groups; the verification module includes at least two or more groups.

[0099] When the verification device sends the device information and the first reference information to the face database, it will send a first time tag; and when the verification device sends the comparison information to the verification module, it will send a second time tag.

[0100] Based on the first time tag, when the face verification data delivery time is greater than or equal to the preset first time threshold, the priority of the corresponding processing item is increased, and at least two or more of the retrieval units are activated to prioritize the processing item according to the priority.

[0101] Alternatively, based on the second time tag, when the feedback time of the proofreading result is greater than or equal to the preset second time threshold, the priority of the corresponding processing item is increased, and at least two or more of the proofreading modules are activated to prioritize the processing item according to the priority.

[0102] Specifically, by establishing a priority system, the timeliness of sudden multi-person face verification can be improved, ensuring the effectiveness of face verification.

[0103] Please see Figure 5 As shown, Figure 5 This invention provides a schematic diagram of a terminal for distributing face data across different devices.

[0104] The terminal 40 includes a processor 401, a communication interface 402, and a memory 403. The processor 401, communication interface 402, and memory 403 can be connected via a bus or other means; this embodiment of the invention uses a bus connection as an example.

[0105] The processor 401 is the computing and control core of the terminal 40. It can parse various instructions and data within the terminal 40. For example, the processor 401 can be a Central Processing Unit (CPU), which can transmit various interactive data between internal structures of the terminal 40. The communication interface 402 can optionally include a standard wired interface or a wireless interface (such as Wi-Fi, mobile communication interface, etc.), and can be used to send and receive data under the control of the processor 401. The communication interface 402 can also be used for the transmission and interaction of internal signaling or instructions within the terminal 40. The memory 403 is a memory device in the terminal 40, used to store programs and data. It is understood that the memory 403 here can include the built-in memory of the terminal 40, or it can include the extended memory supported by the terminal 40. The memory 403 provides storage space, which stores the operating system of the terminal 40. The storage space also stores the program code or instructions required by the processor to perform corresponding operations. Optionally, the storage space can also store relevant data generated by the processor after performing the corresponding operation.

[0106] In this embodiment of the invention, the processor 401 runs executable program code in the memory 403 to perform the following operations:

[0107] Before performing face verification, the verification device sends the device information and the first reference information to the face database, and at the same time, the verification device sends the comparison information to the verification module.

[0108] The face database retrieves stored face data based on the first reference information, and simultaneously generates first verification information based on the device information and face data and sends it to the verification module. Then, it generates complete face verification data based on the device information and packages it.

[0109] The verification module verifies the similarity between the comparison information and the first verification information. When the similarity is greater than or equal to a preset similarity threshold, it sends the verification pass result back to the face database. The face database then sends the packaged face verification data to the verification device.

[0110] In one alternative embodiment, the processor 601 is further configured to:

[0111] If the similarity between the comparison information and the first verification information is less than the preset similarity threshold, a verification failure result is sent to both the face database and the verification device.

[0112] Based on the verification failure result, the verification device will send second reference information to the face database;

[0113] The face database retrieves stored face data again based on the verification failure result using the first reference information and the second reference information. At the same time, it generates second verification information based on the device information and face data and sends it to the verification module. Then, it generates complete face verification data based on the device information and packages it.

[0114] The verification module verifies the similarity between the comparison information and the second verification information. When the similarity is greater than or equal to a preset similarity threshold, it sends the verification result back to the face database. The face database then sends the packaged face verification data to the verification device.

[0115] It should be noted that the implementation of each operation can also be referred to accordingly. Figure 2 and Figure 3 The corresponding description of the method embodiment shown on the terminal side.

[0116] The present invention has been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of the invention. The descriptions of the embodiments above are only for the purpose of helping to understand the invention and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to the invention without departing from its principles, and these improvements and modifications also fall within the protection scope of the claims of the present invention.

Claims

1. A method for distributing facial data across different devices, characterized in that, The method includes: S1. Before performing face verification, the verification device sends the device information and the first reference information to the face database, and at the same time, the verification device sends the comparison information to the verification module. S2. The face database retrieves stored face data based on the first reference information to obtain face data retrieval results; S3. Based on the device information and the facial data retrieval results, first generate verification information and send it to the verification module, then generate complete facial verification data and package it. S4. The proofreading module proofreads the similarity between the comparison information and the proofreading information; S5. When the similarity is greater than or equal to the preset similarity threshold, the verification result is fed back to the face database, and the face database sends the packaged face verification data to the verification device. S6. When the similarity is less than the preset similarity threshold, a verification failure result is fed back to both the face database and the verification device. Based on the verification failure result, the verification device sends second reference information to the face database. Based on the verification failure result, the face database re-retrieves the stored face data with the first reference information and the second reference information to obtain a new face data retrieval result. Based on the new face data retrieval result, the above steps S3-S5 are executed. Wherein, the first reference information and the second reference information are at least two representative feature points of the face data obtained by the verification device, and the feature points contained in the first reference information and the second reference information do not overlap. The comparison information includes multiple feature points corresponding to at least two of the feature parts that make up the eyes, nose, ears, face, and mouth. The calibration information includes multiple feature points of the same feature regions as those contained in the comparison information.

2. The method for distributing face data across different devices as described in claim 1, characterized in that, When the verification device receives a verification failure result, it will collect the face data of the failed verification and the face verification data of the successful verification, and update the feature points of the first reference information and the second reference information through machine learning.

3. The method for distributing face data across different devices as described in claim 1, characterized in that, When the verification device sends the device information and the first reference information to the face database, it will send a first time tag; and when the verification device sends the comparison information to the verification module, it will send a second time tag. Based on the first time tag, when the face verification data delivery time is greater than or equal to the preset first time threshold, the priority of the corresponding processing item is increased; Alternatively, based on the second time tag, when the feedback time of the verification result is greater than or equal to the preset second time threshold, the priority of the corresponding processing item is increased.

4. A device for distributing facial data across different devices, characterized in that, The apparatus for implementing the method according to any one of claims 1-3, the apparatus comprising: The transceiver unit is used to receive device information, first reference information, and comparison information sent by the verification device before performing face verification, as well as to receive the verification pass result sent by the verification module; and to send verification information to the verification module, and to send packaged face verification data to the verification device after receiving the verification pass result. The verification module verifies the similarity between the comparison information and the verification information. When the similarity is greater than or equal to a preset similarity threshold, it feeds back the verification pass result to the face database. The retrieval unit is used to retrieve stored face data from the face database based on first reference information, or first reference information and second reference information, and obtain face data retrieval results. The generation unit is used to generate verification information based on device information and facial data retrieval results, and to generate complete facial verification data based on device information and facial data retrieval results.

5. The apparatus for distributing face data across different devices as described in claim 4, characterized in that, The retrieval unit includes at least two groups; the proofreading module includes at least two groups. When the verification device sends the device information and the first reference information to the face database, it will send a first time tag; and when the verification device sends the comparison information to the verification module, it will send a second time tag. Based on the first time tag, when the face verification data delivery time is greater than or equal to the preset first time threshold, the priority of the corresponding processing item is increased, and at least two sets of the retrieval units are activated to prioritize the processing item according to the priority. Alternatively, based on the second time tag, when the feedback time of the proofreading result is greater than or equal to the preset second time threshold, the priority of the corresponding processing item is increased, and at least two sets of the proofreading modules are activated to prioritize the processing item according to the priority.

6. A terminal, characterized in that, The terminal includes at least one processor, a communication interface, and a memory. The communication interface is used to send and / or receive data, the memory is used to store computer programs, and the at least one processor is used to call at least one computer program stored in the memory to implement the method as described in any one of claims 1-3.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when run on a processor, implements the method as described in any one of claims 1-3.

Images