Service processing method based on identity identification and related apparatus

Abstract

A service processing method based on identity identification, which method is executed by an identity identification device, and comprises: in response to an identity identification operation triggered for an object to be subjected to identification, acquiring an identity identification image collected for the object to be subjected to identification, wherein the identity identification image is used for identity identification (S301); and when the identity identification image successfully matches respective candidate images of a plurality of candidate objects, displaying a resource transfer exemption interface, the resource transfer exemption interface including: a resource gain that can be acquired by the object to be subjected to identification and at least one approach to acquiring the resource gain, wherein the at least one approach is used for transferring identity information of the object to be subjected to identification, and the identity information is used for determining an identity identification result (S302).

Description

A business processing method and related apparatus based on identity recognition

[0001] Related applications

[0002] This application claims priority to Chinese patent application filed on October 11, 2024, application number 202411424308.9, entitled "A Business Processing Method and Related Device Based on Identity Recognition", the entire contents of which are incorporated herein by reference. Technical Field

[0003] This application relates to the field of computer technology, and provides a business processing method, apparatus, device, storage medium, and computer program product based on identity recognition. Background Technology

[0004] Identity recognition is the process of identifying the identity information of an object based on its biometric information. Identity recognition includes, but is not limited to, palm verification, facial verification, and fingerprint recognition. In practical use, the identity information of the object to be identified is obtained by acquiring an identity recognition image of the object and comparing it with candidate images of various candidate objects.

[0005] In related technologies, there are cases where an identity recognition image is successfully matched with candidate images of multiple candidate objects. In this case, in order to identify the identity information of the object to be identified, it is necessary to further verify the object to be identified. Verification methods include, but are not limited to, entering the last four digits of a mobile phone number or entering a verification code.

[0006] However, in scenarios requiring rapid passage, when multiple candidate objects are matched, using additional verification to identify the identity of the person to be identified increases processing time and thus affects passage efficiency. For example, in scenarios such as bus and subway turnstiles, the person to be identified needs to pass through the ticket gate quickly, and additional verification would severely impact the passage efficiency of the person to be identified. Summary of the Invention

[0007] This application provides a business processing method, apparatus, device, storage medium, and computer program product based on identity recognition.

[0008] On the one hand, an identity recognition-based business processing method is provided, executed by an identity recognition device, including:

[0009] In response to an identity recognition operation triggered for an object to be identified, an identity recognition image of the object to be identified is acquired, the identity recognition image being used for identity recognition; and

[0010] When the identity recognition image successfully matches the candidate images of multiple candidate objects, a resource transfer exemption interface is displayed. The resource transfer exemption interface includes: the resource gain that the object to be identified can obtain, and at least one way to obtain the resource gain. The at least one way is used to transmit the identity information of the object to be identified, and the identity information is used to determine the identity recognition result.

[0011] On the one hand, an identity recognition-based business processing method is provided, executed by an identity recognition server, including:

[0012] When an identity recognition image of an object to be identified is received from an identity recognition device, the object to be identified is matched with each of the candidate objects based on the identity recognition image and the candidate images of each candidate object.

[0013] If the matching result indicates that the object to be identified successfully matches multiple candidate objects, then the resources to be transferred for the object to be identified are transferred through a set public account, and a resource transfer exemption instruction is sent to the identity recognition device, so that the identity recognition device displays a resource transfer exemption interface. The resource transfer exemption interface includes the resource gains that the object to be identified can obtain, and at least one way to obtain the resource gains; and

[0014] When the identity information of the object to be identified is received through a selected method among the at least one methods, the identity information is used as the recognition result of the identity recognition image.

[0015] On the one hand, an identity recognition-based business processing device is provided, comprising:

[0016] The identification unit is configured to, in response to an identity recognition operation triggered for an object to be identified, acquire an identity recognition image of the object to be identified, the identity recognition image being used for identity recognition; and

[0017] An exemption unit is used to display a resource transfer exemption interface when the identity recognition image successfully matches the candidate images of multiple candidate objects. The resource transfer exemption interface includes: the resource gain that the object to be identified can obtain, and at least one way to obtain the resource gain. The at least one way is used to transmit the identity information of the object to be identified, and the identity information is used to determine the identity recognition result.

[0018] On the one hand, an identity recognition-based business processing device is provided, comprising:

[0019] The identity recognition unit is used to match the object to be recognized with each candidate object based on the identity recognition image and the candidate images of each candidate object when it receives the identity recognition image of the object to be recognized sent by the palm verification device.

[0020] A resource transfer unit is configured to, if the matching result indicates that the object to be identified successfully matches multiple candidate objects, transfer the resources to be transferred for the object to be identified through a pre-defined public account, and send a resource transfer exemption instruction to the palm verification device, so that the palm verification device displays a resource transfer exemption interface. The resource transfer exemption interface includes the resource gains available to the object to be identified, and at least one method for obtaining the resource gains.

[0021] The information acquisition unit is used to use the identity information as the recognition result of the identity recognition image when it receives the identity information transmitted by the object to be identified through a selected path in the at least one approach.

[0022] On one hand, an electronic device is provided, including a processor and a memory, wherein the memory stores a computer program that, when executed by the processor, causes the processor to perform the steps of the above-described method.

[0023] On one hand, a computer-readable storage medium is provided, comprising a computer program that, when run on an electronic device, causes the electronic device to perform the steps of any of the methods described above.

[0024] On one hand, a computer program product is provided, the program product comprising a computer program stored in a computer-readable storage medium, wherein a processor of an electronic device reads from the computer-readable storage medium and executes the computer program, causing the electronic device to perform the steps of any of the methods described above.

[0025] Other features and advantages of this application will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the published drawings without creative effort.

[0027] Figure 1 is a schematic diagram of the application scenario provided in the embodiments of this application;

[0028] Figure 2 is a schematic diagram of the system architecture provided in an embodiment of this application;

[0029] Figure 3 is a flowchart illustrating the identity recognition-based business processing method on the identity recognition device side provided in the embodiments of this application;

[0030] Figure 4A is a schematic diagram of the first resource transfer exemption interface provided in the embodiment of this application;

[0031] Figure 4B is a schematic diagram of the second resource transfer exemption interface provided in the embodiments of this application;

[0032] Figure 4C is a schematic diagram of the third resource transfer exemption interface provided in the embodiments of this application;

[0033] Figure 5 is a logical schematic diagram of the NFC connection process provided in an embodiment of this application;

[0034] Figure 6 is a logical diagram of the information sharing process provided in an embodiment of this application;

[0035] Figure 7 is a flowchart illustrating the identity recognition-based business processing method on the identity recognition server side provided in an embodiment of this application;

[0036] Figure 8 is a logical diagram of the identity recognition process provided in an embodiment of this application;

[0037] Figure 9A is a logical diagram of the feature combination process provided in an embodiment of this application;

[0038] Figure 9B is a logical schematic diagram of the processing procedure of the first combination provided in the embodiment of this application;

[0039] Figure 9C is a logical schematic diagram of the processing procedure for the second combination provided in the embodiments of this application;

[0040] Figure 9D is a logical schematic diagram of the processing procedure of the third combination provided in the embodiment of this application;

[0041] Figure 10 is a logical schematic diagram of the second feature extraction process provided in the embodiments of this application;

[0042] Figure 11 is a logical diagram of the model training process provided in the embodiment of this application;

[0043] Figure 12 is a logical diagram of the business processing procedure in the subway scenario provided in the embodiments of this application;

[0044] Figure 13 is a logical diagram of the business processing in the vending machine scenario provided in the embodiments of this application;

[0045] Figure 14 is a schematic diagram of the composition structure of the identity recognition-based service device provided in an embodiment of this application;

[0046] Figure 15 is a schematic diagram of the composition structure of the identity recognition-based service device provided in the embodiment of this application;

[0047] Figure 16 is a schematic diagram of the composition structure of the electronic device provided in the embodiment of this application. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. Unless otherwise specified, the embodiments and features in the embodiments of this application can be arbitrarily combined with each other. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than that shown here.

[0049] It is understood that user data, such as accounts and palm images, may be involved in the following specific embodiments of this application. When the various embodiments of this application are applied to specific products or technologies, relevant licenses or consents need to be obtained, and the collection, use and processing of relevant data need to comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0050] To facilitate understanding of the technical solutions provided in the embodiments of this application, some key terms used in the embodiments of this application will be explained below:

[0051] Identity recognition refers to the process of identifying the corresponding identity information of an object based on its biometric information. In this application embodiment, it mainly involves identity recognition based on the acquisition of images of the object to be identified, that is, identity recognition is achieved by identifying the biometric features of the object to be identified from the acquired images through image processing methods. For example, biometric features can be palm features, pupil features, facial features, etc.

[0052] Palm verification operation: Palm verification recognition is a technology that exchanges palm media information for identity information. The palm verification operation refers to the action a user takes to provide the necessary information for identity verification when a business transaction requires it, typically involving placing their palm within the capture range of the palm verification device. Generally, palm verification devices may include image acquisition devices to capture palm images for subsequent identity recognition processes. For example, they may include a 3D camera. 3D cameras, similar to traditional cameras, incorporate liveness detection hardware and software, including depth cameras and infrared cameras, to ensure information security.

[0053] Hand image: refers to an image taken to identify the identity of an object. It can be a 2D image taken by a 2D camera or a 3D image taken by a 3D camera.

[0054] Palm features: These refer to image features obtained by extracting features from a palm image. These features can be the palm image itself, such as the numerical information of each color channel or information such as grayscale and brightness. Alternatively, they can be features obtained by processing the palm image through image processing methods, such as binarization. Or, they can be image features obtained by using an artificial neural network model based on deep learning to extract features from the palm image.

[0055] Object to be identified: In this embodiment, the object to be identified refers to the party that triggers the palm verification operation. In actual business scenarios, before performing palm verification, it is often necessary to register an identity on the business platform so that business processing can be carried out with the successfully registered identity. This allows the business platform to deduct fees based on this identity during business processing. Therefore, the identity recognition process can identify the identity of the object to be identified on the business platform based on biometric information.

[0056] The technical concept of the embodiments of this application will be briefly described below.

[0057] In related technologies, there are cases where an identity recognition image is successfully matched with candidate images of multiple candidate objects. In this case, in order to identify the identity information of the object to be identified, it is necessary to further verify the object to be identified. Verification methods include, but are not limited to, entering the last four digits of a mobile phone number or entering a verification code.

[0058] However, in scenarios requiring rapid passage, when multiple candidate objects are matched, using additional verification to identify the identity of the person to be identified increases processing time and thus affects passage efficiency. For example, in scenarios such as bus and subway turnstiles, the person to be identified needs to pass through the ticket gate quickly, and additional verification would severely impact the passage efficiency of the person to be identified.

[0059] In this embodiment of the application, after the identity recognition operation is triggered by the object to be identified, if the identity recognition image of the object to be identified is successfully matched with the candidate images of multiple candidate objects, the identity recognition device displays a resource transfer exemption interface. The resource transfer exemption interface includes the resource gain that the object to be identified can obtain and at least one way to obtain the resource gain. The at least one way is used to transmit the identity information of the object to be identified, and the identity information is used to determine the identity recognition result.

[0060] On the one hand, compared to additional verification methods such as entering the last four digits of a mobile phone number or a verification code, eliminating the need to transfer resources can ensure the rapid passage of the person to be identified and avoid affecting the passage efficiency. On the other hand, through the design of resource gain, the real identity information of the person to be identified can be obtained. In this way, when performing identity verification again, the real identity information can be used as the palm verification result, thereby improving the accuracy of recognition.

[0061] Next, we will briefly introduce some application scenarios to which the technical solutions of the embodiments of this application are applicable. It should be noted that the application scenarios described below are only for illustrating the embodiments of this application and are not intended to limit the scope. In specific implementation, the technical solutions provided by the embodiments of this application can be flexibly applied according to actual needs.

[0062] The solution provided in this application can be applied to business processing scenarios based on identity recognition, especially scenarios that require rapid business processing, such as public transportation and subway passage.

[0063] Figure 1 shows a schematic diagram of an application scenario provided by an embodiment of this application. In this scenario, an identity recognition device 101 and an identity recognition server 102 may be included.

[0064] The identity recognition device 101 is used to acquire images of the object to be recognized in order to achieve identity recognition of the object. Depending on the identity recognition method used, the identity recognition device 101 can be different devices. For example, when using palm verification recognition, the identity recognition device 101 can be a palm verification device, as specifically shown in Figure 1; or, when using facial verification recognition, the identity recognition device 101 can be a facial verification device; or, when using iris recognition, the identity recognition device 101 can be an iris recognition device. Of course, other image recognition-based devices can also be used, and this embodiment does not impose any limitations on this.

[0065] The identity recognition device 101 may have a related application (APP) installed, such as an identity recognition application. The application involved in the embodiments of this application may be a software client, or a webpage, mini-program, or other client, and there is no limitation on the specific type of client.

[0066] The identity recognition server 102 is the backend server corresponding to the identity recognition application. It is used to realize identity recognition based on the image collected by the identity recognition device 101, return the corresponding identity information to the identity recognition device 101, and also make payment based on the recognized identity.

[0067] The identity recognition server 102 can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms, but it is not limited to these.

[0068] In a real-world scenario, taking the identity recognition device 101 as a palm verification device and the subway payment service as an example, when the person to be identified needs to pass through the subway gate, they can perform a palm verification operation. The palm verification device collects the identity recognition image of the person to be identified and identifies the identity information of the person to be identified based on the identity recognition image. If the identity recognition image of the person to be identified successfully matches the candidate images of multiple candidate objects, then the palm verification device will display a resource transfer exemption interface to the person to be identified, prompting the person to be identified that this trip will be free. The resource transfer exemption interface also includes the resource gains that the person to be identified can obtain, as well as at least one way to obtain the resource gains. Thus, when the person to be identified obtains the resource gains through the selected method of at least one method, the identity information of the person to be identified transmitted through the selected method is obtained, thereby obtaining the recognition result of the identity recognition image. In this embodiment of the application, multiple candidate objects that successfully match the object to be identified can be called highly similar objects. In the presence of highly similar objects, the rapid passage of the object to be identified is ensured by exempting the resources to be transferred. By utilizing resource gains, the real identity information of the object to be identified can be obtained, and the real identity information can be used as the hand verification identification result, thereby improving the identification accuracy.

[0069] It should be noted that the palm verification-based business processing method in this application embodiment can be executed by the identity recognition device 101 or the identity recognition server 102 alone, or by both devices. For example, when executed by the identity recognition device 101 alone, the palm verification-based business processing can be implemented solely by the identity recognition device 101. When executed by both the identity recognition device 101 and the identity recognition server 102, the identity recognition device 101 can collect the identity recognition image of the object to be recognized after the object triggers the palm verification operation, and send the identity recognition image to the identity recognition server 102, which then uses image processing to recognize the identity information of the object to be detected. If the identity recognition image of the object to be recognized successfully matches the candidate images of multiple candidate objects, the identity recognition server 102 notifies the identity recognition device 101 to display the resource transfer exemption interface to prompt the object to be recognized that the trip will be free. The resource transfer exemption interface includes the resource gains that the object to be recognized can obtain, and at least one way to obtain the resource gains. During the process of an object to be identified acquiring resource gains through a selected method from at least one approach, the identity recognition server 102 obtains the identity information of the object to be identified transmitted through the selected method, and then obtains the recognition result of the identity recognition image based on the transmitted identity information. Of course, specific configurations can be made according to the situation in practical applications, and this application does not impose specific limitations here.

[0070] Both the identity recognition device 101 and the identity recognition server 102 may include one or more processors, memory, and interactive I / O interfaces. The memory may also store program instructions required for execution in the palm verification recognition-based business processing method provided in this embodiment. These program instructions, when executed by the processor, can be used to implement the palm verification recognition-based business processing provided in this embodiment.

[0071] In this embodiment, the various devices described above can communicate directly or indirectly through one or more networks. This network can be a wired network or a wireless network; for example, a wireless network can be a mobile cellular network or a Wireless-Fidelity (WIFI) network, or any other possible network. This embodiment does not limit the types of networks used. It should be noted that Figure 1 is merely an example; in reality, the number of terminal devices and servers is not limited, and this embodiment does not impose any specific restrictions.

[0072] Referring to Figure 2, this is the system architecture provided in an embodiment of this application, which is illustrated using palm verification as an identity recognition method. This architecture may include the following three modules:

[0073] (1) Identity recognition front end

[0074] As shown in Figure 2, the identity verification front-end (i.e., the palm verification device) includes a 3D camera and Near Field Communication (NFC). During operation, the palm verification device can utilize the 3D camera for image acquisition and the NFC for data transmission to achieve the function of an identity verification device.

[0075] A palm verification device may include a palm verification and recognition module, a results page module, and a device management module.

[0076] The palm verification and recognition module is used to acquire images of the palm of the object to be identified and perform palm verification and recognition processing. For example, it can use a 3D camera to acquire palm streaming media data and optimize the streaming media data. The optimal palm image is selected by comprehensively evaluating factors such as palm size, angle, image contrast, image brightness, and clarity, and then sent to the application backend service for processing. When the identity recognition device allows, this module can also be used for image preprocessing and palm feature extraction, and the specific configuration can be tailored to the actual situation.

[0077] The results page module displays the resource transfer results, including successful transfer, failed transfer, and unknown result. In this embodiment, when highly similar objects exist, the resource transfer result is unknown. However, to ensure fast passage, a resource transfer exemption interface indicating successful transfer can be displayed to the object to be identified. This interface includes the resource gain available to the object and at least one way to obtain that gain. This not only ensures passage efficiency but also allows the use of resource gains to obtain identity recognition results.

[0078] The equipment management module is used to control the changes in the equipment status of devices such as subway turnstiles after resource transfer. Taking subway turnstiles as an example, the equipment status can be turnstile open or turnstile closed.

[0079] (2) Application backend services

[0080] The application backend services are the backend services corresponding to the palm verification application. The application backend services include palm verification password service, palm verification NFC service, palm-identity matching and association service, palm verification recognition service, algorithm training service, and palm verification device management service.

[0081] The palm verification password service generates a password-formatted message based on information such as the device identifier of the palm verification device, the trigger time of the palm verification operation, and the identity recognition image. Specifically, the generation method involves using a hash function H to generate a unique hash value H(D,T,F) based on the device identifier D of the palm verification device, the trigger time T of the palm verification operation, and the feature vector F of the identity recognition image. This hash value is then converted into a password with a length less than a set threshold, for example, by truncating the first n characters of the hash value as the password, where n is the set length threshold.

[0082] The palm verification NFC service is used to implement NFC functionality. For example, it can be used to transmit the identity information of an object to be identified.

[0083] The palm-identity matching association service is used to store the association between an object and its corresponding palm image. For example, after obtaining the identity information of the object to be identified, the palm-identity matching association service stores the association between the identity information of the object to be identified and the identity recognition image of the object to be identified.

[0084] The palm verification and recognition service is used to implement palm verification and recognition related functions. For example, it can identify the identity information of the object to be identified based on the palm image sent by the front end, and realize the resource transfer process.

[0085] The algorithm training service is used to optimize the identity recognition algorithm based on the identity information of the object to be identified. For example, it can be used to extract features or train the identity recognition model.

[0086] The palm verification device management service is used to manage palm verification devices. For example, remote management of palm verification devices can be achieved through Over-the-Air (OTA) technology.

[0087] (3) Terminal equipment end

[0088] The terminal device includes an operating system, a near-field communication (NFC) module, and applications. The operating system can be, but is not limited to, Android or iOS. The NFC module implements NFC functionality. The applications include a login module for verifying login functionality when the object to be identified logs in. Additionally, the applications can provide a password publishing function for publishing information to be shared. In some implementations, the applications include Moments and Video Accounts, both of which support password publishing.

[0089] The following describes the methods provided by exemplary embodiments of this application in conjunction with the application scenarios and system architecture described above, with reference to the accompanying drawings. It should be noted that the above application scenarios are only shown to facilitate understanding of the spirit and principles of this application, and the embodiments of this application are not limited in any way in this respect.

[0090] Referring to Figure 3, this is a flowchart illustrating the business processing method based on palm verification and recognition provided in an embodiment of this application. This method can be executed by an identity recognition device. The specific implementation flow of this method is as follows:

[0091] S301. In response to an identity recognition operation triggered for an object to be recognized, acquire an identity recognition image of the object to be recognized, and use the identity recognition image for identity recognition.

[0092] In this embodiment, the object to be identified refers to a user who needs to perform business processing. Business processing includes, but is not limited to, services in scenarios such as bus ticket checking and subway ticket checking.

[0093] Identity recognition operations include, but are not limited to, operations that use image acquisition methods for identity recognition, such as palm verification, facial verification, or iris recognition. Taking palm verification as an example, the identity recognition operation is triggered when the person to be identified places their palm in the image acquisition area of ​​the palm verification device.

[0094] In one possible implementation, after the identity recognition device acquires an identity recognition image of the object to be identified, it sends the image to the identity recognition server. The identity recognition server matches the object to be identified with each candidate object based on the identity recognition image. If the matching results indicate that multiple candidate objects successfully match the object to be identified, the server returns a resource transfer exemption instruction to the identity recognition device. Upon receiving the resource transfer exemption instruction, the identity recognition device displays a resource transfer exemption interface to the object to be identified. The matching process between the object to be identified and the candidate objects can also be referred to as the identity recognition process, which is detailed below and will not be repeated here.

[0095] Taking palm verification as an example, the palm verification device can use a camera to collect streaming media data of the palm of the object to be identified. After acquiring the palm streaming media data, the palm verification device can optimize the data by comprehensively evaluating factors such as palm size, angle, image contrast, brightness, and clarity to select the best palm image as the identity recognition image, which is then sent to the identity recognition server for identification. After receiving the identity recognition image and extracting its features, the identity recognition server searches the entire database to identify the identity information of the object to be identified and returns the identity information to the palm verification application. Specifically, the database search can be performed by comparing the palm features of the object to be identified with the candidate palm features of all candidate objects in the database one by one, calculating the feature similarity between them (similarity calculation methods can include Euclidean distance, cosine similarity, etc.). If there are candidate objects whose feature similarity meets the set similarity range, these candidate objects are used as the matching results. For example, if the similarity range is set to be greater than 90%, then when the similarity between the candidate palm feature of a candidate object and the palm feature to be identified is greater than 90%, the candidate object is considered to match the object to be identified.

[0096] In one possible implementation, the identity recognition device can also match the object to be identified with each candidate object based on the identity recognition image, and then display the resource transfer exemption interface to the object to be identified when the matching result indicates that multiple candidate objects are successfully matched with the object to be identified.

[0097] Taking palm verification and recognition as an example, the palm verification device extracts features from the palm image to obtain the palm features of the object to be identified, and matches these palm features with the candidate palm features corresponding to the candidate images of each candidate object. If the matching result indicates that there are multiple candidate palm features that successfully match the palm features of the object to be identified, then the identity recognition image is determined to have successfully matched the candidate images of each of the multiple candidate objects. The identity recognition image can be the optimal palm image selected by the palm verification device from the palm streaming media data, preferably selected through a comprehensive evaluation of factors such as palm size, angle, image contrast, image brightness, and clarity. The matching process between the object to be identified and the candidate objects is detailed below and will not be repeated here.

[0098] S302. When the identity recognition image successfully matches the candidate images of multiple candidate objects, a resource transfer exemption interface is displayed. The resource transfer exemption interface includes: the resource gain that the object to be identified can obtain, and at least one way to obtain the resource gain. The at least one way is used to transfer the identity information of the object to be identified, and the identity information is used to determine the identity recognition result.

[0099] The resource transfer exemption interface refers to the interface that represents the resource transfer result when there are highly similar objects (i.e., the identity recognition image of the object to be identified successfully matches the candidate images of multiple candidate objects). In this embodiment, when there are highly similar objects, the resources to be transferred for the object to be identified will be exempted, i.e., the transfer will be waived. Therefore, the resource transfer result displayed to the object to be identified is that the resource transfer was successful.

[0100] The resource transfer exemption interface may include the resource gains available to the object to be identified, and at least one method for acquiring these resource gains. The method for acquiring resource gains can also be called a resource acquisition method. When the object to be identified acquires resource gains through the provided resource acquisition method, it can send its identity information to the identity verification server via a palm verification device, or directly transmit its identity information to the identity verification server.

[0101] The resource gain available to the object to be identified refers to the additional resources that the object can acquire through at least one provided means. Resource gains include, but are not limited to, one or more of the following: virtual resources and physical resources. Virtual resources refer to resources existing in a digital or computing environment, typically lacking a physical form; examples include game skins, digital collectibles, and game items. Physical resources are typically resources with a physical form, such as cash or small gifts.

[0102] Regardless of the method used to obtain resource gains, the identity information of the object to be identified is obtained with the authorization of the object to be identified. In the embodiments of this application, the identity information can be any information used to distinguish the object, such as the object's ID card number, mobile phone number, or the object's identity identifier in the palm verification application, such as an account identifier (e.g., OpenID).

[0103] In one possible implementation, the resource acquisition method includes, but is not limited to, at least one of the following approaches: acquisition based on short-range communication and acquisition based on message sharing.

[0104] In short-range communication-based acquisition methods, the object to be identified establishes a communication connection with the identity recognition device through a terminal device, thereby transmitting its own identity information to the identity recognition device. The identity recognition device then transmits the object's identity information to an identity recognition server. In this embodiment, the communication connection method used to establish the connection between the terminal device and the identity recognition device includes, but is not limited to, NFC or Bluetooth. Taking NFC as an example, NFC is a short-range, high-frequency wireless communication technology that enables two electronic devices to exchange data within a certain distance. In the NFC-based acquisition method, both the terminal device held by the object to be identified and the identity recognition device are NFC devices. When the two devices are close together, an NFC connection is established between them, and the object's identity information is transmitted to the identity recognition device through this NFC connection.

[0105] In the message-sharing-based identification method, the target entity transmits its identity information to the identity verification server by publishing information to be shared within an application. These applications include, but are not limited to, social media platforms, instant messaging applications, and shopping applications. Before using the application, the target entity registers using its identity information. The application's backend server (referred to as the application server) and the identity verification server support message passing. Thus, when the target entity publishes information to be shared, the application server can transmit the target entity's identity information to the identity verification server. It should be noted that the identity verification server and the application server can be integrated into the same server or they can be two separate servers; there is no limitation on this.

[0106] Depending on the method of resource acquisition, the resource transfer exemption interface may be displayed in several possible ways, including but not limited to the following:

[0107] Case 1: If the resource acquisition method includes short-range communication, then in response to the short-range communication operation triggered by the object to be identified in response to the short-range communication acquisition method, the identity information transmitted by the object to be identified in the process of acquiring resource gains through the short-range communication acquisition method is obtained.

[0108] In some implementations, the resource transfer exemption interface also includes authorization information, which indicates that the object to be identified is permitted to transmit its identity information via short-range communication. In other words, the resource transfer exemption interface includes authorization information in addition to resource gains and resource acquisition methods.

[0109] In one possible implementation, the resource gain displayed in the resource transfer exemption interface can be a specific obtainable resource, or it can be a prompt message indicating an opportunity to obtain a resource gain.

[0110] Referring to Figure 4A, which is a schematic diagram of a resource transfer exemption interface provided in an embodiment of this application. In Figure 4A, the left side is the display screen of the palm verification device, and the right side is the camera of the palm verification device. The display screen shows the resource transfer exemption interface. The resource transfer exemption interface includes the text content "Payment successful", the text content "Congratulations on free payment", the text content "NFC tap, there is a chance to double the reward", and the text content "Using this function requires obtaining identity information, and using it means agreeing to authorization". Among them, the text content "Payment successful" indicates that the resource transfer result is successful. The text content "Congratulations on free payment" is used to indicate that the resource transfer has been exempted in this transaction. The text content "NFC tap, there is a chance to double the reward" indicates that the resource acquisition method is based on NFC. Through NFC, the identified object has the opportunity to obtain double the reward as a resource gain. The text content "Using this function requires obtaining identity information, and using it means agreeing to authorization" is equivalent to authorization information. It should be noted that the resource transfer exemption interface in Figure 4A is for illustrative purposes only. The presentation format of the various contents included in the resource transfer exemption interface is not limited to text. For example, one or more of the contents can also be presented in image form. In addition, the resource transfer results, resource gains, resource acquisition methods, authorization information, etc., can also be distributed in other ways without restriction.

[0111] In the above implementation method, short-range communication is used for information transmission. The data transmission process is convenient and fast, and the identity information of the object to be identified can be obtained quickly, thus improving the efficiency of information acquisition.

[0112] Scenario 2: If the resource acquisition method includes message sharing, then the resource transfer exemption interface will also include information to be shared. In other words, the resource transfer exemption interface will include information to be shared in addition to resource gains and resource acquisition methods.

[0113] In this embodiment of the application, after the object to be identified publishes information to be shared, the identity recognition server can use the identity information associated with the account that published the information to be shared as the identity information of the object to be identified.

[0114] The information to be shared is generated based on at least the device identifier of the palm verification device. Thus, after the information is published, the identity verification server can identify the palm verification device based on its identifier, and then use that device to determine the identity verification image. Finally, the identity information associated with the account that published the information is used as the identity information associated with the identity verification image.

[0115] In some implementations, for identity recognition images with highly similar objects, the identity recognition server can record the association between the palm verification device and the identity recognition image. In this way, the identity recognition image can be determined by using the device identifier of the palm verification device.

[0116] In some implementations, although the probability of encountering highly similar objects is small, in order to avoid the hand verification device being associated with multiple detection images containing highly similar objects, the information to be shared can also be generated based on the device identifier of the hand verification device, combined with information such as the trigger time of the identity recognition operation. By utilizing the trigger time of the identity recognition operation, the identity recognition image can be effectively determined from multiple detection images.

[0117] Referring to Figure 4B, which is a schematic diagram of a resource transfer exemption interface provided in an embodiment of this application. In Figure 4B, the left side is the display screen of the palm verification device, and the right side is the camera of the palm verification device. The display screen shows the resource transfer exemption interface. It includes the text content "Payment successful", the text content "Congratulations on the free order", the content to be shared "Today's joy: Palm verification free order", and the text content "Share the password, with a chance to double the reward". Among them, the text content "Payment successful" represents the resource transfer result of successful resource transfer, the text content "Congratulations on the free order" is used to indicate that the service has been exempted from resource transfer, and the text content "Share the password, with a chance to double the reward" is used to indicate that the resource acquisition method is based on message sharing. By publishing the information to be shared, the identified object has the opportunity to obtain double the reward as a resource gain. It should be noted that the resource transfer exemption interface in Figure 4B is only illustrative, and the presentation form of the various contents included in the resource transfer exemption interface is not limited to text. For example, one or more of the contents can also be presented in the form of images. Furthermore, resource transfer results, resource gains, and resource acquisition methods can also be distributed in other ways without restriction.

[0118] For example, if the resource acquisition methods include message-sharing and NFC-based acquisition, then the resource transfer exemption interface will include: resource gain, resource acquisition method, and information to be shared. In this way, the object to be identified can acquire resource gain through either of the two resource acquisition methods.

[0119] Referring to Figure 4C, which is a schematic diagram of a resource transfer exemption interface provided in an embodiment of this application. In Figure 4C, the left side is the display screen of the palm verification device, and the right side is the camera of the palm verification device. The display screen shows the resource transfer exemption interface. It includes the text content "Payment successful", the text content "Congratulations on the free order", the content to be shared "Today's joy: free order with palm verification", and the text content "Share the password or NFC tap for a chance to double the reward". Among them, the text content "Payment successful" indicates that the resource transfer result is successful. The text content "Congratulations on the free order" is used to indicate that the resource transfer has been exempted in this transaction. The text content "Share the password or NFC tap for a chance to double the reward" indicates that the resource acquisition methods include message-based acquisition and NFC-based acquisition. By publishing the information to be shared or using NFC, the identified object has the opportunity to obtain double the reward as a resource gain. In addition, the resource transfer exemption interface can also include the text "Using this function requires obtaining identity information; use signifies consent to authorization." This text informs the target user that their identity information will be obtained during the resource gain acquisition process. Of course, in practical applications, authorization information can also be obtained in other steps of the identity verification process. For example, the user can be informed of the use of identity information during registration or before palm verification. There are no restrictions on this, and it will not be elaborated further here. It should be noted that the resource transfer exemption interface in Figure 4C is for illustrative purposes only. The presentation format of the various items included in the resource transfer exemption interface is not limited to text; for example, one or more items can be presented as images. Furthermore, the resource transfer result, resource gain, and resource acquisition method can also be presented in other ways, without restriction.

[0120] Furthermore, the object to be identified can trigger a selection operation for at least one selected approach to achieve the transmission of identity information.

[0121] In this embodiment of the application, the selection operation triggered by the object to be identified for a selected path in at least one of the following operations includes at least one of the following operations: a short-range communication operation triggered by the object to be identified for an acquisition method based on short-range communication, and a sharing operation triggered by the object to be identified for a message to be shared. The two methods will be described below.

[0122] (1) Triggering method for short-range communication operation:

[0123] In some implementations, when the identity recognition device receives a short-range communication request from the terminal device of the object to be identified, a connection channel is established between the terminal device of the object to be identified and the palm verification device; the connection channel is used to transmit the identity information of the object to be identified; when the identity information of the object to be identified is obtained through the connection channel, it is determined that the object to be identified has triggered a short-range communication operation for the acquisition method based on short-range communication.

[0124] In this embodiment, the communication connection method used between the terminal device of the object to be identified and the palm verification device is not limited. The connection channel includes, but is not limited to, NFC connection channels, Bluetooth connection channels, etc.

[0125] Taking NFC connection as an example, when the distance between the terminal device and the identification device is within a preset distance range, the terminal device sends an NFC connection request to the identification device. In other words, a preset distance range can be set for the identification device. If the terminal device to be identified is within this preset distance range, it indicates that the terminal device needs to establish an NFC connection with the identification device, thus proceeding with the subsequent NFC connection process. Specifically, this setting can be done by experimentally measuring the signal strength at different distances based on the relationship between the NFC signal strength of the identification device and distance, and selecting a distance range where the signal strength can stably support data transmission as the preset distance range. For example, after multiple experimental measurements, if it is determined that the NFC signal strength is stable and data transmission can be performed normally at a distance of 3-8 centimeters, then the preset distance range is set to 3-8 centimeters.

[0126] When the identification device receives a short-range communication request, it establishes a connection channel between the terminal device of the object to be identified and the palm verification device. It should be noted that during the establishment process, the identification device can negotiate communication parameters with the terminal, such as transmission rate and encryption method. After negotiating the communication parameters, the connection channel is established between the two. Once the connection channel is established, the terminal device and the identification device can exchange data. When the identification information of the object to be identified is obtained through the connection channel, the identification device determines to trigger the short-range communication operation.

[0127] Taking a palm verification recognition scenario as an example, as shown in Figure 5, it is a logical schematic diagram of an NFC connection establishment method provided in this embodiment. The terminal device held by the object to be identified is a mobile phone, and the identity recognition device is a palm verification device. When the mobile phone is close to the palm verification device and the distance between the two is less than 5 centimeters, the mobile phone sends an NFC connection request to the palm verification device. After receiving the NFC connection request, the palm verification device negotiates communication parameters with the mobile phone. The communication parameters include the data transmission rate and encryption method between the two. After negotiating the communication parameters, an NFC connection channel is established between the two. Then, the mobile phone sends the identity information of the object to be identified to the palm verification device through the connection channel.

[0128] (2) Triggering method for sharing operation

[0129] In one possible implementation, after the object to be identified publishes content containing information to be shared, the application server can send the identity information associated with the account that published the information to the identity recognition server. The identity recognition server uses the identity information associated with the account that published the information to be shared as the identity information of the object to be identified. Furthermore, the application server can also send the information to be shared to the identity recognition server. Since the information to be shared is generated based on the device identifier of the identity recognition device and is unique, the identity recognition server can determine the identity recognition device for resource transfer exemption based on the information to be shared, and then determine the identity recognition image sent by the identity recognition device. Further, the identity information and the identity recognition image can be associated.

[0130] In one possible implementation, after receiving the identity information, the identity recognition server sends an information reception instruction to the identity recognition device. Upon receiving the information reception instruction, the identity recognition device determines that the object to be identified has triggered a sharing operation. The information reception instruction is a notification signal sent by the identity recognition server to the identity recognition device after receiving the identity information of the object to be identified. When the identity recognition device receives this instruction, it determines that the object to be identified has triggered a sharing operation, and may subsequently display a gain acquisition interface, indicating that the object to be identified has had its identity information recognized as the result of the identity recognition image.

[0131] In one possible implementation, considering that the object to be identified may delay publishing the information to be shared, and in some scenarios requiring rapid passage, the identity recognition device needs to quickly acquire and detect the image of the next object. Therefore, when the display duration of the resource transfer exemption interface reaches the set display duration, a sharing operation is triggered, and the gain acquisition interface is displayed, thereby quickly acquiring and detecting the image of the next object. The gain acquisition interface is an interactive interface displayed by the identity recognition device after the object to be identified acquires resource gain through a selected method and the identity recognition server confirms receipt of the object's identity information. This interface is mainly used to show the object to be identified the relevant results of acquiring resource gain and to inform the object to be identified through prompts that the identity information transmitted during the acquisition of resource gain has been used as the recognition result of the identity recognition image, serving as information feedback and interactive guidance.

[0132] Taking the palm verification recognition scenario as an example, refer to Figure 6, which is a logical schematic diagram of a message sharing method provided in this embodiment. The target user posts a message with the password "Today's Joy: Palm Verification for Free" in an instant messaging application. After the application server determines that the target user has posted content with the password, it sends the identity information associated with the account that posted the password to the identity recognition server. The identity information associated with the account indicates that the target user who posted the password is account A in the instant messaging application. The identity recognition server uses the identity information associated with the account that posted the information to be shared as the identity information of the target user. Of course, the identity information can also be other information such as a mobile phone number; this is not limited.

[0133] Referring to Figure 7, which is a flowchart illustrating a palm verification and recognition business processing method provided in an embodiment of this application, applied to an identity recognition server. The specific process is as follows:

[0134] S701. When the identity recognition image of the object to be recognized is received from the identity recognition device, the object to be recognized is matched with each candidate object based on the identity recognition image and the candidate images of each candidate object.

[0135] To facilitate the description of the identity verification process, we will use a palm verification device and a palm image as the identity verification image as an example. For other identity verification methods, the following description can also be used as a reference.

[0136] In this embodiment of the application, the identity recognition server can extract features from the palm image to be recognized, obtain the palm features of the object to be recognized, and then match the palm features to be recognized with the candidate palm features corresponding to the candidate images of each candidate object to obtain the matching result.

[0137] In one possible implementation, during the matching process, the identity recognition server calculates the similarity between the candidate palm features corresponding to the candidate images of each candidate object and the palm features to be identified, thereby obtaining the feature similarity between the corresponding candidate object and the object to be identified. If there are multiple candidate objects and their feature similarity with the object to be identified meets the set similarity range, a matching result representing the successful matching between the object to be identified and multiple candidate objects is generated.

[0138] Feature similarity can be achieved using Euclidean distance. The calculation is performed where x and y represent two feature vectors, and n represents the dimension of the feature vectors; cosine similarity can also be used. Calculation, where and These represent two feature vectors respectively; Jaccard similarity can also be used. The calculation is performed, where A and B represent two sets. The candidate hand features corresponding to the candidate images of each candidate object can be pre-calculated and stored, or they can be stored after acquiring the hand image to be identified; there is no limitation on this.

[0139] For example, referring to Figure 8, suppose the candidate objects include A, B, C, D, etc., and the database stores the candidate palm features corresponding to each candidate object. The identity recognition server calculates the feature similarity between each candidate palm feature and the palm feature to be identified. Suppose the similarity range is greater than 98%, and the feature similarity of A, B, and C is greater than 98%. Then, A, B, and C are all candidate objects that are successfully matched with the object to be identified. At this time, a matching result representing the successful matching of the object to be identified with multiple candidate objects is generated.

[0140] It should be noted that, in the embodiments of this application, if there is no candidate object that successfully matches the object to be identified, it indicates that the object to be identified may not have registered its identity, which can help the object to be identified to register its identity; if there is only one candidate object that successfully matches the object to be identified, then the identity information associated with that candidate object can be determined as the identity information of the object to be identified.

[0141] In the above implementation method, identity recognition is performed through image features. Since image features are difficult to copy or tamper with, the accuracy and security of identity recognition can be guaranteed. Furthermore, since feature similarity can measure the degree of similarity between two image features, using feature similarity to determine the matching result can improve the accuracy of the identity recognition result.

[0142] S702. If the matching result indicates that the object to be identified is successfully matched with multiple candidate objects, then the resources to be transferred of the object to be identified are transferred through the set public account, and a resource transfer exemption instruction is sent to the identity recognition device so that the identity recognition device displays the resource transfer exemption interface. The resource transfer exemption interface includes the resource gains that the object to be identified can obtain, and at least one way to obtain the resource gains.

[0143] The specific transfer process can be as follows: First, query the available resource balance B from the set public account. If the balance B is greater than or equal to the resource R to be transferred, deduct the resource R to be transferred from the public account and update the balance of the public account to BR. If the balance B is less than the resource R to be transferred, issue a resource shortage prompt and suspend the transfer operation.

[0144] In this embodiment, the public account is used to perform resource transfers on behalf of the object requiring resource transfer exemption. That is, when resource transfer exemption is performed, the resource transfer is performed from the public account, rather than from the account of the object to be identified.

[0145] In one possible implementation, administrators with business operation permissions can pre-apply for public accounts specific to the business. This way, when highly similar accounts exist, the public account can directly replace the object to be identified for resource transfer. The public account can be configured through an identity verification server, but is not limited to this.

[0146] Highly similar accounts refer to accounts associated with multiple candidate objects in identity recognition scenarios, whose candidate images and the identity recognition images of the objects to be identified reach a certain similarity standard during feature matching, thus successfully matching. These accounts exhibit high similarity in the identity recognition process, making it difficult to directly determine the identity of the object to be identified. Therefore, when processing related business, public accounts are used to prioritize resource transfer to ensure the rapid passage of the object to be identified.

[0147] In one possible implementation, administrators can pre-configure operational strategies for scenarios with highly similar objects. Taking buses and subways as an example, for palm verification recognition scenarios, administrators need to configure the following operational strategies: If the object to be identified performs palm verification and successfully matches multiple candidate objects, then the public account will be used to prioritize payment. If the object to be identified uses other identity verification methods besides palm verification, then verification can still be performed via mobile phone number or QR code scanning to confirm the object's identity information, and payment will be made based on the object's identity information. Furthermore, administrators can configure operational strategies to incentivize the object to provide feedback, thereby further confirming the object's identity information. For example, configuring resource-enhancing features, such as sharing to receive random red envelopes or having the opportunity to double the amount.

[0148] In one possible implementation, if the matching result indicates that the object to be identified successfully matches multiple candidate objects, the identity recognition server can also generate information to be shared and send a resource transfer exemption instruction carrying the information to be shared. Of course, in practical applications, the information to be shared can also be sent separately to the identity recognition device; this is not limited.

[0149] In one possible implementation, the information to be shared is at least a password generated based on the device identifier of the palm verification device, and the password length is less than a set length threshold. The device identifier is used to uniquely identify a palm verification device, and the device identifier can be, but is not limited to, a device serial number (SN). The SN is a string serial number that can uniquely identify a device. Furthermore, the identity verification server can also generate the information to be shared by combining the trigger time of the identity verification operation. The password is, but is not limited to, a combination of one or more of the following: numbers, letters, etc.

[0150] The specific combination method can be as follows: The device identifier D of the palm verification device and the trigger time T of the identity recognition operation are converted into timestamps, and the feature vector F of the identity recognition image are used as inputs. A unique hash value H(D,T,F) is generated through a hash function H. This hash value is then converted into a password with a length less than a set length threshold, for example, by truncating the first n characters of the hash value as the password, where n is the set length threshold. The information to be shared in this way includes both the uniqueness of the device identifier and the timeliness of the trigger time.

[0151] For example, the identity recognition server generates a unique ID based on the device serial number of the palm verification device, the timestamp corresponding to the trigger time of the palm verification operation, and the MD5 value of the palm image to be recognized. Then, it generates a password according to a set length threshold, for example, the generated password is "Today's Joy: Palm Verification Free Order", and then returns the password to the palm verification device.

[0152] By generating password-based information to be shared, the length of the password can be controlled, making it easy for users to remember and publish. This increases the probability of sharing the information and helps optimize subsequent identity recognition algorithms.

[0153] S703. When the identity information of the object to be identified is received through a selected method in at least one of the methods, the identity information is used as the recognition result of the identity recognition image.

[0154] In practical applications, there is a certain probability of encountering highly similar users. When this occurs, verification is typically done via phone number verification or QR code verification. However, in scenarios like public transportation such as buses and subways, verification via phone number or QR code can affect traffic efficiency. In this embodiment, when a highly similar user is identified, a public account is invoked for priority resource transfer to ensure passage. The identity recognition device will also notify the user that resource transfer is waived (i.e., the transaction is free), and will also indicate that the user can obtain resource gains. These resource gains are used to obtain the user's identity information.

[0155] In one possible implementation, if the chosen method is short-range communication-based acquisition, the identity verification server obtains the identity information through the identity verification device. For short-range communication-based acquisition, the identity verification device can directly transmit the acquired identity information to the identity verification server.

[0156] In one possible implementation, if the selected approach is based on message sharing, then when the object to be identified triggers a publishing operation for the information to be shared, the identity server will use the identity information associated with the object publishing the information to be shared as the identity information of the object to be identified.

[0157] An identity verification server can be an application's backend server. When an individual to be identified posts content on a social media application via their mobile phone, the content is uploaded to the identity verification server. Upon receiving the information to be shared, the identity verification server uses the identity information associated with the person who posted the information as the identity information of the individual to be identified. Alternatively, the identity verification server can be a server capable of data interaction with the application's backend server. When an individual to be identified posts content on a social media application via their mobile phone, if the posted content is shareable, the application's backend server uses the identity information associated with the person who posted the shareable content as the identity information of the individual to be identified, and then transmits this identity information to the identity verification server, enabling the identity verification server to obtain the individual's identity information from the shareable information.

[0158] To avoid the recurrence of the problem of highly similar objects, in this embodiment of the application, the identity recognition server can also use the identity information transmitted by the object to be identified to optimize the identity recognition algorithm, thereby improving the accuracy of identity recognition.

[0159] In one possible implementation, the identity recognition server can also select features from each candidate feature dimension based on the identity recognition image and multiple candidate objects to obtain at least one identity recognition feature dimension; when a new palm image of the object to be recognized is received again, the server can use at least one identity recognition feature dimension to extract features from the new palm image to obtain new image features, and use the new image features to obtain the recognition result.

[0160] Feature selection can be implemented in at least one of the following ways, but is not limited to:

[0161] Method 1: The identity recognition server combines each candidate feature dimension to obtain each set of feature dimensions, and performs the following operations for each set of feature dimensions: Combine each candidate feature dimension to obtain each set of feature dimensions, and perform the following operations for each set of feature dimensions: Extract features from the identity recognition image and multiple candidate images according to each set of feature dimensions to obtain corresponding image features, and obtain corresponding predicted matching results based on the obtained image features, and obtain evaluation results corresponding to a set of feature dimensions based on the predicted matching results and the actual matching results; Based on the evaluation results corresponding to each set of feature dimensions, select a set of feature dimensions that meet the set evaluation conditions from each set of feature dimensions, and use the candidate feature dimensions in the selected set of feature dimensions as the identity recognition feature dimensions.

[0162] Let's continue using a hand image as an example for identity recognition. For each set of feature dimensions, feature extraction is performed on the hand image to be recognized and multiple candidate hand images to obtain the corresponding hand features. Based on the obtained hand features, the corresponding predicted matching results are obtained. Based on the predicted matching results and the actual matching results, the evaluation results corresponding to that set of feature dimensions are obtained.

[0163] For feature extraction from palm images, candidate feature dimensions include, but are not limited to, one or more of the following: geometric features representing the overall shape and contour of the palm, such as palm size, finger length, finger spacing, and finger joint features; texture features representing the lines and details of the palm surface, such as palm prints, wrinkles, spots, and texture direction; shape features representing the edges and contours of the palm, such as contour features, curvature variation features, and concavity / convexity features; and depth features extracted from palm images through deep learning, such as convolution features and residual features.

[0164] In this embodiment, the combination of each candidate feature dimension is not limited, and each group of feature dimensions may include all or part of the combination of each candidate feature dimension.

[0165] In this embodiment of the application, the true matching result is determined based on the identity information transmitted by the object to be identified. Based on the identity information transmitted by the object to be identified, the candidate object that truly corresponds to the object to be identified can be determined from among the candidate objects.

[0166] The evaluation result is used to assess the image discrimination capability of combinations of corresponding candidate feature dimensions. For example, when the true matching result and the predicted matching result are the same, the evaluation result indicates that the corresponding combination prediction is correct; when the true matching result and the predicted match are different, the evaluation result indicates that the corresponding combination prediction is incorrect. Evaluation conditions include, but are not limited to, setting the evaluation result to accurate prediction. For example, directly comparing the predicted matching result and the true matching result: if they are the same, the evaluation result E is recorded as 1, indicating accurate prediction; if they are different, E is recorded as 0, indicating incorrect prediction.

[0167] Referring to Figure 9A, assuming the actual matching result is that the image of the hand to be identified matches the candidate hand image 1, the candidate feature dimensions include 10 feature dimensions such as feature 1 (hand size), feature 2 (finger length), and feature 3 (finger spacing). The identity recognition server combines each candidate feature dimension to obtain each set of feature dimensions. Taking 3 sets of feature dimensions as an example, the 3 sets of feature dimensions are combination 1, combination 2, and combination 3. Combination 1 includes candidate feature dimensions such as feature 1, feature 2, and feature 4; combination 2 includes candidate feature dimensions such as feature 2, feature 3, and feature 4; and combination 3 includes candidate feature dimensions such as feature 1, feature 3, and feature 4.

[0168] Referring to Figure 9B, for combination 1, the identity recognition server extracts features from the hand image to be recognized and multiple candidate hand images according to candidate feature dimensions such as feature 1, feature 2, and feature 4, respectively, to obtain the hand features corresponding to the hand image to be recognized and the multiple candidate hand images (i.e., the candidate hand images of A, B, and C). Then, based on the obtained hand features corresponding to the hand image to be recognized and the multiple candidate hand images, the predicted matching result between the hand image to be recognized and the multiple candidate hand images is obtained. The predicted matching result indicates that the hand image to be recognized is successfully matched with candidate hand image 1. Obviously, the predicted matching result is the same as the actual matching result. Therefore, based on the predicted matching result and the actual matching result, the evaluation result corresponding to combination 1 is obtained. The evaluation result corresponding to combination 1 indicates that the prediction is correct.

[0169] Referring to Figure 9C, for combination 2, the identity recognition server extracts features from the hand image to be recognized and multiple candidate hand images according to the candidate feature dimensions such as feature 2, feature 3, and feature 4, respectively, to obtain the corresponding hand features. Based on the obtained hand features, the server obtains the predicted matching result between the hand image to be recognized and multiple candidate hand images. The predicted matching result indicates that the hand image to be recognized is successfully matched with candidate hand image 2. Based on the predicted matching result and combined with the actual matching result, the server obtains the evaluation result corresponding to combination 2. The evaluation result indicates the prediction error.

[0170] Referring to Figure 9D, for combination 3, the identity recognition server extracts features from the hand image to be recognized and multiple candidate hand images according to the candidate feature dimensions such as feature 1, feature 3, and feature 4, respectively, to obtain the corresponding hand features. Based on the obtained hand features, the server obtains the prediction matching result between the hand image to be recognized and multiple candidate hand images. The prediction matching result indicates that the hand image to be recognized is successfully matched with both candidate hand image 1 and candidate hand image 2. Combining the actual matching result, the server obtains the evaluation result corresponding to combination 3. The evaluation result corresponding to combination 3 indicates a prediction error.

[0171] Finally, based on the evaluation results corresponding to the three sets of feature dimensions, the candidate feature dimensions included in combination 1 are used as the identity recognition feature dimensions.

[0172] Feature selection allows us to choose the subset of features from the original feature set that best represent the data characteristics and distinguish different categories. This reduces the dimensionality of features, avoids redundant information, and improves the efficiency and accuracy of the model. Furthermore, feature selection can remove irrelevant or redundant features, making the model more concise and efficient.

[0173] Method 2: The identity recognition server combines the candidate feature dimensions to obtain sets of feature dimensions, and performs the following operations for each set of feature dimensions: Combine the candidate feature dimensions to obtain sets of feature dimensions, and perform the following operations for each set of feature dimensions: Extract features from the identity recognition image and multiple candidate images according to a set of feature dimensions to obtain corresponding image features, and obtain an evaluation result corresponding to a set of feature dimensions based on the feature similarity between the image features of the identity recognition image and the image features of multiple candidate images; Based on the evaluation results corresponding to each set of feature dimensions, select a set of feature dimensions that meet the set evaluation conditions from each set of feature dimensions, and use the candidate feature dimensions in the selected set of feature dimensions as the identity recognition feature dimensions.

[0174] In this context, the feature similarity between the image features of the identity recognition image and the image features of multiple candidate images refers to the feature similarity between the image features of each candidate image and the image features of the identity recognition image. A feature similarity score characterizes the degree of similarity between the image features of the identity recognition image and the image features of a candidate image. Feature similarity can be implemented using distance measurement algorithms such as Euclidean distance, cosine similarity, and Jaccard similarity, but is not limited to these.

[0175] The evaluation results are used to assess the image discrimination capability of combinations of corresponding candidate feature dimensions. For example, the average or weighted average of the feature similarities between the image features of the identity recognition image and the image features of multiple candidate images can be used as the evaluation result for a set of feature dimensions. A higher evaluation result indicates greater similarity between the image features of the identity recognition image and the image features of multiple candidate images, and the lower the prediction accuracy of that set of feature dimensions. Conversely, a lower evaluation result indicates greater difference between the image features of the identity recognition image and the image features of multiple candidate images, and the higher the prediction accuracy of that set of feature dimensions. It should be noted that in practical applications, the evaluation result can also be obtained based on the feature difference between the image features of the identity recognition image and the image features of multiple candidate images. A feature difference is used to characterize the degree of difference between the image features of the identity recognition image and the image features of a candidate image. The feature difference can be implemented using distance metrics such as Euclidean distance, cosine similarity, and Jaccard similarity, which will not be elaborated upon here.

[0176] For example, combining feature similarity, we first define a consistency score S1 between the predicted matching result and the actual matching result: S1 = 1 when they are consistent, and S1 = 0 when they are inconsistent. Then, we calculate the average feature similarity between the image features of the identity recognition image and the image features of multiple candidate images. Where M is the number of candidate images, S jLet E be the feature similarity between the j-th candidate image and the identity recognition image (calculated in the same way as the cosine similarity above). The final evaluation result E = αS1 + (1-α)S2, where α is the weight coefficient, ranging from [0,1]. Evaluation conditions include, but are not limited to: the evaluation result E being greater than the set evaluation threshold E. T This indicates that the prediction was accurate.

[0177] Referring again to Figure 9A, assuming that the actual matching result is that the identity recognition image matches the candidate image 1, the candidate feature dimensions include 10 feature dimensions such as feature 1 (palm size), feature 2 (finger length), and feature 3 (finger spacing). The identity recognition server combines each candidate feature dimension to obtain 3 sets of feature dimensions, namely combination 1, combination 2, and combination 3. Combination 1 includes candidate feature dimensions such as feature 1, feature 2, and feature 4; combination 2 includes candidate feature dimensions such as feature 2, feature 3, and feature 4; and combination 3 includes candidate feature dimensions such as feature 1, feature 3, and feature 4.

[0178] Referring to Figure 10, for combination 1, the identity recognition server extracts features from the hand image to be recognized and each candidate hand image according to candidate feature dimensions such as feature 1, feature 2, and feature 4, obtaining the hand features corresponding to the hand image to be recognized and each of the multiple candidate hand images. Then, it calculates the feature similarity between the hand features of the multiple candidate hand images and the hand features of the hand image to be recognized, and uses the average of the calculated similarities as the evaluation result of combination 1. Similarly, the evaluation results of combination 2 and combination 3 can be obtained. Assuming that the evaluation condition is set to the lowest evaluation result value, and that combination 2 has the lowest evaluation result value among the three combinations, then based on the evaluation results corresponding to each set of feature dimensions, the set of feature dimensions with the lowest evaluation result value is selected from each set of feature dimensions, i.e., combination 2, and the candidate feature dimensions in combination 2 are used as the identity recognition feature dimensions.

[0179] In the above implementation, a distance metric is used to measure the similarity or difference between data points. This ensures that the selected features have good discriminative power in terms of distance metric, thereby effectively distinguishing highly similar data.

[0180] Method 3: Based on Method 1, combine Method 2 for feature extraction.

[0181] The identity recognition server combines the candidate feature dimensions to obtain sets of feature dimensions, and performs the following operations for each set of feature dimensions: It extracts features from the identity recognition image and multiple candidate images according to each set of feature dimensions to obtain corresponding image features. Based on the predicted matching results and the actual matching results, and combining the feature similarity between the image features of the identity recognition image and the image features of multiple candidate images, it obtains an evaluation result corresponding to a set of feature dimensions. Based on the evaluation results corresponding to each set of feature dimensions, it selects a set of feature dimensions that meet the set evaluation conditions from each set of feature dimensions, and uses the candidate feature dimensions in the selected set of feature dimensions as the identity recognition feature dimensions.

[0182] For example, the evaluation result can be calculated in, but is not limited to, the following manner: A result evaluation value is obtained by comparing the predicted matching result and the actual matching result. Then, a distance evaluation value is obtained based on the feature similarity between the image features of the identity recognition image and the image features of multiple candidate images. The distance evaluation value can be the mean or weighted mean of the feature similarity, etc. Afterwards, the result evaluation value and the distance evaluation value are averaged or weighted averaged to obtain an evaluation result corresponding to a set of feature dimensions. Since method three is similar to methods one and two, it will not be described in detail here.

[0183] The result evaluation value is obtained by comparing the predicted matching result and the actual matching result during the feature selection process. It is used to evaluate the accuracy of the combination of candidate feature dimensions. For example, when the predicted matching result and the actual matching result are consistent, the result evaluation value is recorded as 1, indicating accurate prediction; if they are inconsistent, it is recorded as 0, indicating incorrect prediction. The distance evaluation value is a value calculated based on the feature similarity between the image features of the identity recognition image and the image features of multiple candidate images. It is used to measure the degree of similarity between the identity recognition image and multiple candidate images. It can be the average value or a weighted average value of the feature similarity. The higher the evaluation value, the more similar the image features of the identity recognition image and the image features of multiple candidate images are, and the lower the prediction accuracy of that set of feature dimensions; the lower the value, the greater the difference, and the higher the prediction accuracy.

[0184] In the above implementation, the combination of candidate feature dimensions is evaluated using the prediction matching results. At the same time, the similarity or difference between data points is measured by combining distance metrics. In this way, the selected features can have good discriminative power in terms of distance metrics, thereby improving the accuracy of identity recognition.

[0185] When a new palm image of the object to be identified is received again, features are extracted from the new palm image using at least one identification feature dimension to obtain new image features. Using these new image features, an accurate identification result can be obtained. Since the process of identification using new image features is similar to the process of identification using palm features from the palm image of the object to be identified, it will not be described in detail here. Please refer to S701 for details.

[0186] In some implementations, the identity recognition process is typically implemented using an identity recognition model, whereby the identity recognition server inputs the image of the hand to be recognized into the identity recognition model to obtain a matching result.

[0187] In one possible implementation, the identity recognition server associates the palm image to be recognized with matching candidate objects (i.e., candidate objects that match the object to be detected). Then, using the matching candidate objects and their associated palm images to be recognized, combined with multiple candidate objects that successfully match the palm images to be recognized of the matching candidate objects and their associated candidate palm images, the identity recognition model is trained to obtain the trained identity recognition model, which can be used for subsequent identity recognition.

[0188] The specific method for storing the association can be to create an association table in the database. This table contains fields such as the identifier of the hand image to be recognized, the identity identifier of the matching candidate, and the association time. When the identity recognition server determines that the hand image to be recognized matches a candidate, it inserts the identifier of the hand image to be recognized and the identity identifier of the matching candidate into the association table and records the association time for subsequent querying and use. The association table is a structured table in the database used to record and store the correspondence between the hand image to be recognized and the matching candidate. By setting fields such as the identifier of the hand image to be recognized, the identity identifier of the matching candidate, and the association time, it saves the association information in an ordered and queryable manner, facilitating the management, querying, and use of this association data in the subsequent identity recognition business process, thereby providing data support for the optimization of the identity recognition algorithm and business processing. The association time is the time recorded in the association table when the association between the hand image to be recognized and the matching candidate is established.

[0189] It should be noted that in this embodiment, the identity recognition model can be trained after obtaining the identity information of a highly similar object to be identified, or after obtaining the identity information of a certain number of highly similar objects to be identified, or by using the identity information of highly similar objects to be identified within a set update cycle. No limitation is imposed on this approach. This document only uses the process of training the identity recognition model using the identity information of a single object to be identified as an example.

[0190] Specifically, the identity recognition server labels the palm image to be recognized, obtaining the corresponding label for the palm image. The label is used to represent the true identity of the palm image. Then, the palm image to be recognized is input into the identity recognition model to obtain the predicted matching result. Specifically, the identity recognition model first extracts features from the palm image to be recognized to obtain the corresponding sample palm features. Then, the sample palm features of the palm image to be recognized are matched with the sample palm features of each candidate palm image to obtain the predicted matching result. Finally, based on the predicted matching result and the true matching result, the model loss is obtained, and the parameters are tuned based on the model loss.

[0191] As an example, let the predicted matching result be P and the actual matching result be T. The model loss is calculated using the cross-entropy loss function L, i.e. Where m is the number of categories.

[0192] Referring to Figure 11, assuming that the candidate objects that successfully match the object to be identified include A, B, and C, the identity recognition server determines that the object to be identified is A based on the identity information transmitted by the object. The identity recognition server associates the hand image to be identified with A and labels the hand image to be identified, obtaining the label corresponding to the hand image. The label is used to represent that the object associated with the hand image to be identified is A. Then, the hand image to be identified is input into the identity recognition model to obtain the predicted matching result. The predicted matching result represents that the object associated with the hand image to be identified is B. Afterwards, based on the predicted matching result and the true matching result, the model loss is obtained, and the parameters are tuned based on the model loss.

[0193] As an example, the parameter tuning method is as follows: using stochastic gradient descent, let the model parameters be θ and the learning rate be α, then the parameter update formula is: Where L is the model loss.

[0194] In one possible implementation, regularization techniques can be used during model training. Regularization is an important technique in machine learning and deep learning, used to prevent overfitting and improve the model's generalization ability. By adding a regularization term to the loss function, regularization techniques can penalize complex models and encourage them to learn simpler, more general patterns. In the embodiments of this application, regularization techniques include, but are not limited to: L1 regularization, L2 regularization, Dropout, data augmentation, early stopping, etc.

[0195] Taking L2 regularization as an example, an L2 regularization term is added to the original loss function L. Where λ is the regularization coefficient, w i These are the parameters of the model. Then the new loss function L...new for During training, when updating model parameters using stochastic gradient descent, it is necessary to consider not only the gradient of the original loss function but also the gradient of the regularization term. Therefore, the parameter update formula becomes... Where α is the learning rate and θ are the model parameters. This method penalizes excessively large parameters in the model, preventing overfitting and improving the model's generalization ability.

[0196] The present application will now be described with reference to several specific embodiments.

[0197] Example 1:

[0198] Referring to Figure 12, it is a schematic diagram of a passage verification process based on palm verification recognition in a subway passage scenario provided in an embodiment of this application.

[0199] S121. The object to be identified places its palm in the image acquisition area of ​​the palm verification device, triggering the palm verification operation.

[0200] S122. The palm verification device responds to the palm verification operation triggered by the object to be identified, acquires the palm image of the object to be identified, and sends the palm image to the identity recognition server.

[0201] S123. The identity recognition server performs identity recognition based on the palm image to be recognized. Specifically, based on the identity recognition image and the candidate images of each candidate object, the object to be recognized is matched with each candidate object.

[0202] S124. If the matching result indicates that the object to be identified is successfully matched with multiple candidate objects, the identity recognition server will transfer the resources to be transferred of the object to be identified through the set public account.

[0203] S125. The identity verification server sends a resource transfer exemption instruction to the palm verification device.

[0204] S126. The palm verification device displays a resource transfer exemption interface. The resource transfer exemption interface includes the resource gains that the object to be identified can obtain, as well as two ways to obtain resource gains. The two ways include the acquisition method based on NFC and the acquisition method based on message sharing.

[0205] S127. The palm verification device sends an opening instruction to the subway turnstile to control the opening of the subway turnstile.

[0206] S128. When the mobile phone of the object to be identified is brought close to the palm verification device, an NFC connection channel is established between the mobile phone and the palm verification device, and the identity information of the object to be identified is sent to the palm verification device through the NFC connection channel.

[0207] After acquiring the identity information, the palm verification device displays a gain acquisition interface with prompts. The prompts indicate that the identity information transmitted during the process of acquiring resource gain through the selected method is used as the recognition result of the identity recognition image.

[0208] S129. The palm verification device transmits the identity information of the object to be identified to the identity recognition server. Furthermore, the identity recognition server can use the identity information of the object to be identified to optimize the identity recognition algorithm.

[0209] Example 2:

[0210] Referring to Figure 13, it is a schematic diagram of the item purchase process based on palm verification recognition in an automated vending scenario provided in an embodiment of this application.

[0211] S131. The object to be identified places its palm in the image acquisition area of ​​the palm verification device, triggering the palm verification operation.

[0212] S132. The palm verification device responds to the palm verification operation triggered by the object to be identified, acquires the palm image of the object to be identified, and sends the palm image to the identity recognition server.

[0213] S133. The identity recognition server performs identity recognition based on the palm image to be recognized. Specifically, based on the identity recognition image and the candidate images of each candidate object, the object to be recognized is matched with each candidate object.

[0214] S134. If the matching result indicates that the object to be identified is successfully matched with multiple candidate objects, the identity recognition server will transfer the resources to be transferred of the object to be identified through the set public account.

[0215] S135. The identity verification server sends a resource transfer exemption instruction to the palm verification device.

[0216] S136. The palm verification device displays a resource transfer exemption interface. The resource transfer exemption interface includes the resource gains that the object to be identified can obtain, as well as two ways to obtain resource gains. The two ways include the acquisition method based on NFC and the acquisition method based on message sharing.

[0217] S137. The palm verification device sends a pop-up instruction to the vending machine to control the vending machine to pop up the items to be sold.

[0218] S138. The person to be identified publishes content with a password in a social application. The identity recognition server is also the backend server of the social application. When the person to be identified publishes content in the social application via mobile phone, the content will be uploaded to the identity recognition server. Alternatively, the identity recognition server can also be a server that can interact with the backend server of the social application. When the person to be identified publishes content in the social application via mobile phone, if the published content is to be shared, the content to be shared can be transmitted to the identity recognition server.

[0219] Furthermore, the identity verification server optimizes the identity verification algorithm based on the identity information associated with the account that issued the password.

[0220] In addition, the identity recognition server can also send an information receiving instruction to the palm verification device so that the palm verification device can display a gain acquisition interface containing prompt information. The prompt information is used to indicate that the identity information transmitted by the object to be identified during the process of acquiring resource gain through the selected method is used as the recognition result of the identity recognition image.

[0221] Based on the same inventive concept, this application provides a business processing device based on identity recognition. As shown in Figure 14, which is a structural schematic diagram of the business processing device 1400 based on identity recognition, it may include:

[0222] The identification unit 1401 is configured to, in response to an identity recognition operation triggered for an object to be identified, acquire an identity recognition image collected for the object to be identified, wherein the identity recognition image is used for identity recognition.

[0223] The exemption unit 1402 is used to display a resource transfer exemption interface when the identity recognition image is successfully matched with the candidate images of multiple candidate objects. The resource transfer exemption interface includes: the resource gain that the object to be identified can obtain, and at least one way to obtain the resource gain. The at least one way is used to transmit the identity information of the object to be identified, and the identity information is used to determine the identity recognition result.

[0224] In one possible implementation, if the at least one approach includes: an acquisition method based on short-range communication, then the exemption unit 1402 is further configured to:

[0225] In response to a short-range communication operation triggered by the object to be identified in relation to the acquisition method based on short-range communication, the identity information transmitted by the object to be identified during the process of acquiring the resource gain through the acquisition method based on short-range communication is obtained.

[0226] In one possible implementation, the exemption unit 1402 is further used for:

[0227] When a short-range communication request is received from the terminal device of the object to be identified, a connection channel is established between the terminal device of the object to be identified and the palm verification device; the connection channel is used to transmit the identity information of the object to be identified.

[0228] When the identity information of the object to be identified is obtained through the connection channel, it is determined that the object to be identified triggers a short-range communication operation in response to the acquisition method based on short-range communication.

[0229] In one possible implementation, if the at least one approach includes a message-sharing-based acquisition method, then the resource transfer exemption interface further includes: information to be shared, which is at least generated based on the device identifier of the identification device.

[0230] In one possible implementation, the information to be shared is at least a password generated based on the device identifier of the palm verification device, and the length of the password is less than a set length threshold.

[0231] In one possible implementation, the identity recognition device is a palm verification device, the identity recognition image is a palm image, and the recognition unit 1401 is further configured to:

[0232] Feature extraction is performed on the palm image to obtain the palm features of the object to be identified, and the palm features to be identified are matched with the palm features corresponding to the candidate images of each candidate object;

[0233] If the matching result indicates that there are multiple candidate palm features that successfully match the palm feature to be identified, then it is determined that the identity recognition image has successfully matched the candidate images of the multiple candidate objects.

[0234] Based on the same inventive concept, this application provides a business processing device based on identity recognition. As shown in Figure 15, which is a structural schematic diagram of the business processing device 1500 based on identity recognition, it may include:

[0235] The identity recognition unit 1501 is used to match the object to be recognized with each candidate object based on the identity recognition image and the candidate images of each candidate object when it receives the identity recognition image of the object to be recognized sent by the palm verification device.

[0236] The resource transfer unit 1502 is used to transfer the resources to be transferred of the object to be identified through a set public account if the matching result indicates that the object to be identified is successfully matched with multiple candidate objects, and send a resource transfer exemption instruction to the palm verification device so that the palm verification device displays a resource transfer exemption interface. The resource transfer exemption interface includes the resource gains that the object to be identified can obtain, and at least one way to obtain the resource gains.

[0237] The information acquisition unit 1503 is used to use the identity information as the recognition result of the identity recognition image when it receives the identity information transmitted by the object to be identified through a selected path in the at least one path.

[0238] In one possible implementation, the information acquisition unit 1503 is used for:

[0239] If the selected method is: acquisition method based on short-range communication, then the identity information is obtained through the identity recognition device;

[0240] If the selected method is: message sharing-based acquisition method, then when the object to be identified triggers a publishing operation for the information to be shared, the identity information associated with the object publishing the information to be shared will be used as the identity information of the object to be identified.

[0241] In one possible implementation, the identity recognition unit 1501 is further used for:

[0242] Based on the identity recognition image and the multiple candidate objects, feature selection is performed on each candidate feature dimension to obtain at least one identity recognition feature dimension.

[0243] When a new palm image of the object to be identified is received again, the feature extraction of the new palm image is performed using the at least one identity recognition feature dimension to obtain new image features, and the recognition result is obtained using the new image features.

[0244] In one possible implementation, when selecting features from each candidate feature dimension based on the identity recognition image and the plurality of candidate objects to obtain at least one identity recognition feature dimension, the identity recognition unit 1501 is specifically used for:

[0245] The candidate feature dimensions are combined to obtain sets of feature dimensions, and the following operations are performed on each set of feature dimensions:

[0246] Based on a set of feature dimensions, feature extraction is performed on the identity recognition image and the multiple candidate images respectively to obtain corresponding image features. Based on the obtained image features, corresponding prediction matching results are obtained. Based on the prediction matching results and the actual matching results, the evaluation results corresponding to the set of feature dimensions are obtained.

[0247] Based on the evaluation results corresponding to each set of feature dimensions, a set of feature dimensions that meet the set evaluation conditions is selected from each set of feature dimensions, and the candidate feature dimensions in the selected set of feature dimensions are used as identity recognition feature dimensions.

[0248] In one possible implementation, when obtaining the evaluation results corresponding to the set of feature dimensions based on the predicted matching results and the actual matching results, the identity recognition unit 1501 is specifically used to perform at least one of the following operations:

[0249] The evaluation results corresponding to the set of feature dimensions are obtained directly based on the predicted matching results and the actual matching results.

[0250] Based on the predicted matching results and the actual matching results, and combining the image features of the identity recognition image and the feature similarity between the image features of the multiple candidate images, the evaluation results corresponding to the set of feature dimensions are obtained.

[0251] In one possible implementation, when matching the object to be identified with each candidate object based on the identity recognition image and the candidate images of each candidate object, the identity recognition unit 1501 is specifically used for:

[0252] Feature extraction is performed on the identity recognition image to obtain the palm features of the object to be identified;

[0253] The similarity between the candidate palm features corresponding to the candidate images of each candidate object and the palm features to be identified is calculated to obtain the feature similarity between the corresponding candidate object and the object to be identified.

[0254] If there are multiple candidate objects that have a feature similarity to the object to be identified and meet the set similarity range, then a matching result is generated to indicate that the object to be identified has successfully matched with the multiple candidate objects.

[0255] For ease of description, the above sections are divided into modules (or units) according to their functions and described separately. Of course, in implementing this application, the functions of each module (or unit) can be implemented in one or more software or hardware components.

[0256] Regarding the apparatus in the above embodiments, the specific manner in which each unit executes the request has been described in detail in the embodiments related to the method, and will not be elaborated here.

[0257] Those skilled in the art will understand that various aspects of this application can be implemented as a system, method, or program product. Therefore, various aspects of this application can be specifically implemented in the following forms: a completely hardware implementation, a completely software implementation (including firmware, microcode, etc.), or a combination of hardware and software implementations, collectively referred to herein as a "circuit," "module," or "system."

[0258] Based on the same inventive concept, this application also provides an electronic device. In one embodiment, the electronic device can be a server or a terminal device. Referring to FIG16, which is a schematic diagram of the structure of a possible electronic device provided in an embodiment of this application, in FIG16, the electronic device 1600 includes: a processor 1610 and a memory 1620.

[0259] The memory 1620 stores a computer program that can be executed by the processor 1610. By executing the instructions stored in the memory 1620, the processor 1610 can perform the steps of the above-mentioned identity recognition-based business processing method.

[0260] Memory 1620 may be volatile memory, such as random-access memory (RAM); memory 1620 may also be non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid-state drive (SSD); or memory 1620 may be any other medium capable of carrying or storing desired program code having the form of instructions or data structures and accessible by a computer, but is not limited thereto. Memory 1620 may also be a combination of the above-mentioned memories.

[0261] The processor 1610 may include one or more central processing units (CPUs) or digital processing units, etc. The processor 1610 implements the aforementioned identity-based business processing method when executing a computer program stored in the memory 1620.

[0262] In some embodiments, the processor 1610 and the memory 1620 may be implemented on the same chip, while in other embodiments they may be implemented on separate chips.

[0263] This application embodiment does not limit the specific connection medium between the processor 1610 and the memory 1620. This application embodiment takes the connection between the processor 1610 and the memory 1620 via a bus as an example. The bus is depicted with a thick line in Figure 16. The connection methods between other components are only illustrative and not intended to be limiting. Buses can be divided into address buses, data buses, control buses, etc. For ease of description, Figure 16 uses only one thick line, but it does not depict only one bus or one type of bus.

[0264] Based on the same inventive concept, embodiments of this application provide a computer-readable storage medium including a computer program. When the computer program is run on an electronic device, it causes the electronic device to perform the steps of the aforementioned identity-based business processing method. In some possible implementations, various aspects of the identity-based business processing method provided in this application can also be implemented as a program product including a computer program. When the program product is run on an electronic device, it causes the electronic device to perform the steps in the aforementioned identity-based business processing method. For example, the electronic device can perform the steps shown in FIG3 or FIG7.

[0265] The program product may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: electrical connections having one or more wires, portable disks, hard disks, RAM, ROM, erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0266] The program product of the embodiments of this application may be a CD-ROM and include a computer program, and may run on an electronic device. However, the program product of this application is not limited thereto. In this document, the readable storage medium may be any tangible medium that contains or stores a computer program that may be used by or in conjunction with a command execution system, apparatus, or device.

[0267] A readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying a readable computer program. This propagated data signal may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable signal medium may also be any readable medium other than a readable storage medium, capable of sending, propagating, or transmitting a computer program for use by or in conjunction with a command execution system, apparatus, or device.

[0268] In summary, this application provides a business processing method, apparatus, device, storage medium, and computer program product based on identity recognition. When an identity recognition operation is triggered for an object to be identified, the identity recognition device acquires an identity recognition image for identity recognition. In traditional identity recognition scenarios, if the identity recognition image successfully matches candidate images of multiple candidate objects, a complex verification process is often required, which significantly increases business processing time. In this solution, when such a successful match occurs, a resource transfer exemption interface containing resource gains and acquisition methods is displayed. Exempting the object from the resource transfer operation allows the object to be identified to quickly complete the business process without waiting for the cumbersome verification process. For example, in scenarios requiring rapid passage, such as bus and subway turnstiles, passengers can quickly pass through the turnstiles, greatly improving business processing efficiency. At the same time, resource gains serve as an incentive mechanism, encouraging the object to be identified to actively provide identity information. Since the provided identity information is genuine, subsequent identity recognition can be based on accurate information, avoiding recognition errors caused by inaccurate information, thereby improving the accuracy of identity recognition.

[0269] Furthermore, if the acquisition of resource gains includes short-range communication (SRM) methods, the identification device will respond to the short-range communication operation triggered by the object to be identified, acquiring the identity information transmitted during the resource gain acquisition process. Short-range communication technologies (such as NFC or Bluetooth) are characterized by high data transmission speed and high stability. During data transmission, they can complete the exchange of large amounts of data in a short time, significantly reducing the time cost of information transmission compared to traditional verification methods, such as manually entering the last four digits of a mobile phone number or a verification code. For example, at subway turnstiles, passengers only need to bring their NFC-enabled mobile phones close to the turnstiles to quickly complete the transmission of identity information, enabling rapid passage, improving information acquisition efficiency, and thus increasing the overall speed of business processing.

[0270] When an object to be identified triggers a short-range communication-based acquisition method, the identification device, upon receiving a short-range communication request from the object's terminal device, establishes a connection channel between the terminal device and the handprint verification device. During this connection establishment process, the identification device and the terminal device negotiate communication parameters, such as transmission rate and encryption method. Negotiating the transmission rate ensures data is transmitted at the fastest and most stable speed, avoiding data congestion or loss; selecting an appropriate encryption method guarantees the security of identity information during transmission, preventing theft or tampering. This explicit triggering and information transmission mechanism, through negotiation of communication parameters, ensures the stability and reliability of identity information transmission, contributing to the accurate and timely acquisition of the object's identity information, thereby improving the efficiency and accuracy of identification.

[0271] If the resource gain acquisition method includes message sharing, the resource transfer exemption interface will include shareable information generated based on the device identifier of the identification device. After the target object publishes content containing the shareable information, the application server will send the identity information associated with the account that published the shareable information to the identity recognition server, which will then use this as the target object's identity information. This method increases the ways to obtain identity information, making identity recognition no longer limited to a single method and improving its flexibility. Simultaneously, since the shareable information is generated based on the device identifier of the identification device, and the device identifier is unique, this ensures the traceability of the shareable information. Through the device identifier, the identity recognition server can accurately determine the identification device performing the resource transfer exemption, and thus determine the corresponding identity recognition image, which helps to accurately associate identity information and identity recognition images, improving the accuracy of identity recognition.

[0272] Furthermore, the information to be shared must be at least a password generated based on the device identifier of the palm verification device, and the password length must be less than a set length threshold. Because the password is shorter, it is easier for users to remember and operate. In practical applications, users can easily remember the password and share it on social media platforms or other applications, reducing the difficulty of sharing. More users willing to share the information increases the scope and speed of information dissemination, thereby increasing the probability of sharing. A large number of sharing behaviors can provide the identity verification server with more identity information samples, which helps to optimize the identity verification algorithm and improve the accuracy and efficiency of identity verification.

[0273] When the identity verification device is a palm verification device and the identity verification image is a palm image, to determine whether the identity verification image matches successfully with the candidate images of multiple candidate objects, feature extraction needs to be performed on the palm image to obtain the palm features of the object to be identified, and then matched with the corresponding palm features of the candidate images of each candidate object. The features of a palm image are unique and difficult to replicate. Each person's palm geometric features (such as palm size, finger length, finger spacing, etc.) and texture features (such as palm lines, wrinkles, etc.) are unique, making them difficult to forge or tamper with. Extracting these features for identity verification ensures the accuracy and security of the identification process. Furthermore, using feature similarity to determine the matching result allows for a more accurate measurement of the similarity between the object to be identified and the candidate objects. Feature similarity can be calculated using various algorithms, such as Euclidean distance and cosine similarity. These algorithms accurately reflect the similarity between the features of two images, thereby improving the accuracy of the identity verification results.

[0274] When the identity recognition server receives the identity recognition image of the object to be identified from the identity recognition device, it matches the object to be identified with each candidate object based on that image and the candidate images of each candidate object. Traditional identity recognition methods may simply compare images, which is prone to misjudgment. In this solution, the server extracts features from the image of the hand to be identified, obtaining the hand features of the object to be identified, and then calculates the similarity between this and the corresponding candidate hand features of each candidate object. If multiple candidate objects have feature similarities with the object to be identified that meet a set similarity range, a matching result is generated indicating that the object to be identified has successfully matched with multiple candidate objects. Identity recognition through image features ensures accuracy and security because image features are difficult to copy or tamper with. Furthermore, feature similarity accurately measures the degree of similarity between two image features; using feature similarity to determine the matching result improves the accuracy of the identity recognition result.

[0275] If the matching result indicates that the object to be identified successfully matches multiple candidate objects, the identity recognition server will transfer the resources to be transferred for the object to be identified through a designated public account and send a resource transfer exemption instruction to the identity recognition device, causing the device to display the resource transfer exemption interface. In traditional business processing, when multiple matching objects appear, a dilemma of identity verification often arises, leading to business processing stagnation. This solution, by transferring resources through a public account, ensures the rapid passage of the object to be identified, avoiding business processing delays. Simultaneously, the resource gains and acquisition methods included in the resource transfer exemption interface incentivize the object to be identified to proactively provide identity information, improving the accuracy of identity recognition and the efficiency of business processing. For example, in a public transportation scenario, passengers can board quickly without waiting for identity verification, improving the efficiency of public transportation operations.

[0276] When the identity information of the object to be identified is received through a selected method from at least one approach, the identity recognition server uses this information as the recognition result of the identity image. If the selected method is based on short-range communication, the identity recognition server obtains the identity information through the identity recognition device. The identity recognition device acts as a bridge and relay in this process; it can perform preliminary verification and processing of the identity information to ensure the accuracy and reliability of the information transmitted to the server. Simultaneously, the speed of short-range communication ensures that the identity information reaches the server in a timely manner, helping the server to make timely judgments and improving the efficiency of identity recognition.

[0277] If the selected method is message-sharing-based acquisition, when the target object triggers a publishing operation for the information to be shared, the identity recognition server uses the identity information associated with the object publishing the information as the target object's identity information. Message sharing leverages the widespread reach of social media platforms or other applications, expanding the scope of identity information acquisition. During the process of publishing the information, the target object attracts more participants, increasing the opportunity to obtain identity information. Simultaneously, by associating the information with the publishing object's identity, the target object's identity can be accurately determined, improving the accuracy and reliability of identity recognition.

[0278] The identity recognition server also selects features from the identity recognition image and multiple candidate objects, obtaining at least one identity recognition feature dimension. When a new hand image of the object to be recognized is subsequently received, these identity recognition feature dimensions are used to extract features from the new hand image, obtaining new image features, and then using these new image features to obtain the recognition result. The original feature set may contain a large amount of redundant information and irrelevant features, which increases the computational burden on the model and reduces recognition efficiency. Feature selection allows for the selection of the most representative subset of features from numerous candidate feature dimensions, best distinguishing different categories, reducing the feature dimensionality and avoiding interference from redundant information. This makes the model more efficient in processing data, improving its efficiency and accuracy, and consequently enhancing the accuracy and efficiency of subsequent identity recognition.

[0279] During feature selection, the identity recognition server combines candidate feature dimensions to obtain sets of feature dimensions and performs operations on each set. Based on a set of feature dimensions, feature extraction is performed on the identity recognition image and multiple candidate images to obtain corresponding image features. Based on these image features, corresponding predicted matching results are obtained, and based on the predicted matching results and the actual matching results, an evaluation result corresponding to that set of feature dimensions is obtained. Based on the evaluation results corresponding to each set of feature dimensions, a set of feature dimensions that meet the set evaluation conditions is selected from each set, and the candidate feature dimensions in this selected set are used as the identity recognition feature dimensions. This comprehensive evaluation and selection method can fully consider the effectiveness of each feature dimension combination. By comparing the predicted matching results and the actual matching results, the performance of feature dimensions in practical applications can be observed, thereby accurately selecting the most discriminative and representative feature dimensions, improving the accuracy and effectiveness of feature selection, and contributing to improved accuracy in identity recognition.

[0280] When obtaining evaluation results for a set of feature dimensions based on predicted and actual matching results, the identity recognition server can employ multiple methods. One approach is to directly obtain the evaluation result based on both predicted and actual matching results; this method is simple, direct, and allows for quick assessment of the accuracy of the feature dimension combination. Another approach is to combine the image features of the identity recognition image with the feature similarity between the image features of multiple candidate images, taking into account both the accuracy of the matching results and the similarity of image features. This method comprehensively considers both the accuracy of the matching results and the similarity of image features, enabling a more comprehensive and accurate evaluation of the effectiveness of the feature dimensions. These diverse evaluation methods make the evaluation results more objective and accurate, allowing for more precise selection of suitable feature dimensions, improving the quality of feature selection, and further enhancing the accuracy of identity recognition.

[0281] When matching the object to be identified against each candidate object based on the identity recognition image and the candidate images of each candidate object, the identity recognition server extracts features from the identity recognition image to obtain the palm features of the object to be identified. Then, it calculates the similarity between the palm features of the candidate images of each candidate object and the palm features of the object to be identified. If there are multiple candidate objects whose feature similarity with the object to be identified meets the set similarity range, a matching result is generated indicating that the object to be identified has successfully matched with multiple candidate objects. Identity recognition through image features ensures accuracy and security because image features are difficult to copy or tamper with. Furthermore, feature similarity accurately measures the degree of similarity between two image features; using feature similarity to determine the matching result can improve the accuracy of the identity recognition result.

[0282] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0283] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A business processing method based on identity recognition, executed by an identity recognition device, comprising: In response to an identity recognition operation triggered for an object to be identified, an identity recognition image of the object to be identified is acquired, and the identity recognition image is used for identity recognition. and When the identity recognition image successfully matches the candidate images of multiple candidate objects, a resource transfer exemption interface is displayed. The resource transfer exemption interface includes: the resource gain that the object to be identified can obtain, and at least one way to obtain the resource gain. The at least one way is used to transmit the identity information of the object to be identified, and the identity information is used to determine the identity recognition result.

2. The method of claim 1, wherein the at least one approach comprises: Based on the acquisition method using short-range communication, the method further includes: In response to a short-range communication operation triggered by the object to be identified in relation to the acquisition method based on short-range communication, the identity information transmitted by the object to be identified during the process of acquiring the resource gain through the acquisition method based on short-range communication is obtained.

3. The method as described in claim 2, wherein the short-range communication operation is triggered in the following manner: When a short-range communication request is received from the terminal device of the object to be identified, a connection channel is established between the terminal device of the object to be identified and the palm verification device; the connection channel is used to transmit the identity information of the object to be identified. When the identity information of the object to be identified is obtained through the connection channel, it is determined that the object to be identified triggers a short-range communication operation in response to the acquisition method based on short-range communication.

4. The method according to any one of claims 1 to 3, wherein the at least one approach comprises: Based on the message sharing acquisition method, the resource transfer exemption interface also includes: information to be shared, which is generated at least based on the device identifier of the identity recognition device.

5. The method as described in claim 4, wherein the information to be shared is at least a password generated based on the device identifier of the palm verification device, and the length of the password is less than a set length threshold.

6. The method according to any one of claims 1-5, wherein the identity recognition device is a palm verification device, and the identity recognition image is a palm image; The successful matching of the identity recognition image with the candidate images of multiple candidate objects includes: Feature extraction is performed on the palm image to obtain the palm features of the object to be identified, and the palm features to be identified are matched with the palm features corresponding to the candidate images of each candidate object; If the matching result indicates that there are multiple candidate palm features that successfully match the palm feature to be identified, then it is determined that the identity recognition image has successfully matched the candidate images of the multiple candidate objects.

7. A business processing method based on palm verification and recognition, executed by an identity recognition server, comprising: When an identity recognition image of an object to be identified is received from an identity recognition device, the object to be identified is matched with each of the candidate objects based on the identity recognition image and the candidate images of each candidate object. If the matching result indicates that the object to be identified is successfully matched with multiple candidate objects, then the resources to be transferred of the object to be identified are transferred through the set public account, and a resource transfer exemption instruction is sent to the identity recognition device so that the identity recognition device displays a resource transfer exemption interface. The resource transfer exemption interface includes the resource gains that the object to be identified can obtain, and at least one way to obtain the resource gains. and When the identity information of the object to be identified is received through a selected method among the at least one methods, the identity information is used as the recognition result of the identity recognition image.

8. The method of claim 7, further comprising: If the selected method is: acquisition method based on short-range communication, then the identity information is obtained through the identity recognition device.

9. The method of claim 7 or 8, further comprising: If the selected method is: message sharing-based acquisition method, then when the object to be identified triggers a publishing operation for the information to be shared, the identity information associated with the object publishing the information to be shared will be used as the identity information of the object to be identified.

10. The method of any one of claims 7 to 9, further comprising: Based on the identity recognition image and the multiple candidate objects, feature selection is performed on each candidate feature dimension to obtain at least one identity recognition feature dimension. When a new palm image of the object to be identified is received again, the feature extraction of the new palm image is performed using the at least one identity recognition feature dimension to obtain new image features, and the recognition result is obtained using the new image features.

11. The method of claim 10, wherein the step of selecting features for each candidate feature dimension based on the identity recognition image and the plurality of candidate objects to obtain at least one identity recognition feature dimension includes: The candidate feature dimensions are combined to obtain sets of feature dimensions, and the following operations are performed on each set of feature dimensions: Based on a set of feature dimensions, feature extraction is performed on the identity recognition image and the multiple candidate images respectively to obtain corresponding image features. Based on the obtained image features, corresponding prediction matching results are obtained. Based on the prediction matching results and the actual matching results, the evaluation results corresponding to the set of feature dimensions are obtained. Based on the evaluation results corresponding to each set of feature dimensions, a set of feature dimensions that meet the set evaluation conditions is selected from each set of feature dimensions, and the candidate feature dimensions in the selected set of feature dimensions are used as identity recognition feature dimensions.

12. The method of claim 11, wherein obtaining the evaluation result corresponding to the set of feature dimensions based on the predicted matching result and the actual matching result includes at least one of the following operations: The evaluation results corresponding to the set of feature dimensions are obtained directly based on the predicted matching results and the actual matching results. Based on the predicted matching results and the actual matching results, and combining the image features of the identity recognition image and the feature similarity between the image features of the multiple candidate images, the evaluation results corresponding to the set of feature dimensions are obtained.

13. The method according to any one of claims 7-12, wherein matching the object to be identified with each candidate object based on the identity recognition image and the candidate images of each candidate object comprises: Feature extraction is performed on the identity recognition image to obtain the palm features of the object to be identified; The similarity between the candidate palm features corresponding to the candidate images of each candidate object and the palm features to be identified is calculated to obtain the feature similarity between the corresponding candidate object and the object to be identified. If there are multiple candidate objects that have a feature similarity to the object to be identified and meet the set similarity range, then a matching result is generated to indicate that the object to be identified has successfully matched with the multiple candidate objects.

14. A business processing device based on palm verification and recognition, comprising: The identification unit is configured to, in response to an identity recognition operation triggered for an object to be identified, acquire an identity recognition image collected for the object to be identified, wherein the identity recognition image is used for identity recognition. and An exemption unit is used to display a resource transfer exemption interface when the identity recognition image successfully matches the candidate images of multiple candidate objects. The resource transfer exemption interface includes: the resource gain that the object to be identified can obtain, and at least one way to obtain the resource gain. The at least one way is used to transmit the identity information of the object to be identified, and the identity information is used to determine the identity recognition result.

15. A business processing device based on palm verification and recognition, comprising: The identity recognition unit is used to match the object to be recognized with each candidate object based on the identity recognition image and the candidate images of each candidate object when it receives the identity recognition image of the object to be recognized sent by the palm verification device. The resource transfer unit is used to transfer the resources to be transferred of the object to be identified through a set public account if the matching result indicates that the object to be identified is successfully matched with multiple candidate objects, and send a resource transfer exemption instruction to the palm verification device so that the palm verification device displays a resource transfer exemption interface. The resource transfer exemption interface includes the resource gains that the object to be identified can obtain, and at least one way to obtain the resource gains. and The information acquisition unit is used to use the identity information as the recognition result of the identity recognition image when it receives the identity information transmitted by the object to be identified through a selected path in the at least one approach.

16. An electronic device comprising a processor and a memory, wherein, The memory stores a computer program that, when executed by the processor, causes the processor to perform the steps of any of the methods described in claims 1 to 6, or to perform the steps of any of the methods described in claims 7 to 13.

17. A computer-readable storage medium comprising a computer program, which, when executed on an electronic device, causes the electronic device to perform the steps of any of the methods of claims 1 to 6, or to perform the steps of any of the methods of claims 7 to 13.

18. A computer program product comprising a computer program stored in a computer-readable storage medium, wherein a processor of an electronic device reads from the computer-readable storage medium and executes the computer program, causing the electronic device to perform the steps of any of the methods of claims 1 to 6, or to perform the steps of any of the methods of claims 7 to 13.

Images