Transaction risk control method, device and computer equipment based on biometric identification

Abstract

The present application relates to a transaction risk control method, device and computer equipment based on biometrics. The method includes: identifying the target part data collected from the target object, obtaining the biometric result, and determining the target account that matches the target object based on the biometric result; obtaining the operation data associated with the target account, and performing a multi-dimensional risk assessment on the target account based on the operation data to obtain the risk level of the target account; obtaining the transaction data associated with the target account, and performing a multi-dimensional risk assessment on the transaction based on the transaction data to obtain the transaction risk result of the transaction; determining the target risk control method of the transaction based on the account risk level and the transaction risk result, and performing risk control on the transaction based on the target risk control method. The use of this method can achieve flexible risk control.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a transaction risk control method, apparatus, computer equipment, storage medium and computer program product based on biometrics. Background Art

[0002] With the development of image recognition technology, verifying the identity of an object by identifying its biometric features has become a common payment verification method. During the payment process, the collected biometric features are matched with the biometric features pre-stored in the database to verify the identity of the object.

[0003] However, although the current image recognition algorithms have gradually matured, they still cannot avoid risks in the payment verification process. For example, the subject’s account may be hacked, resulting in the leakage or tampering of biometric data, resulting in the biometrics provided during verification not being the subject’s own biometrics.

[0004] Therefore, there is an urgent need for a method to fundamentally improve the security of biometric verification payment methods. Summary of the invention

[0005] Based on this, it is necessary to provide a biometric-based transaction risk control method, apparatus, computer equipment, computer-readable storage medium and computer program product that can improve the security of biometric-based transactions in response to the above-mentioned technical problems.

[0006] On the one hand, the present application provides a transaction risk control method based on biometrics. The method comprises:

[0007] Identify the target part data collected from the target object to obtain a biometric recognition result, and determine a target account that matches the target object based on the biometric recognition result;

[0008] Acquire operation data associated with the target account, and perform a multi-dimensional risk assessment on the target account according to the operation data to obtain a risk level of the target account;

[0009] Acquire transaction data associated with the target account, and perform multi-dimensional risk identification on the transaction based on the transaction data to obtain a transaction risk result of the transaction;

[0010] Based on the account risk level and the transaction risk results, the target risk control method for this transaction is determined, and the risk of this transaction is managed and controlled according to the target risk control method.

[0011] On the other hand, the present application also provides a transaction risk control device based on biometrics. The device comprises:

[0012] An identification module, used to identify the target part data collected from the target object, obtain a biometric identification result, and determine a target account matching the target object based on the biometric identification result;

[0013] A risk assessment module, used to obtain operation data associated with the target account, and perform a multi-dimensional risk assessment on the target account according to the operation data to obtain a risk level of the target account;

[0014] A risk identification module is used to obtain transaction data associated with the target account, and to perform multi-dimensional risk identification on the transaction based on the transaction data to obtain a transaction risk result of the transaction;

[0015] The determination module is used to determine the target risk control method of this transaction based on the account risk level and the transaction risk result, and to perform risk management and control on this transaction according to the target risk control method.

[0016] On the other hand, the present application further provides a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned transaction risk control method based on biometrics when executing the computer program.

[0017] On the other hand, the present application also provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the above-mentioned transaction risk control method based on biometric identification when executed by a processor.

[0018] On the other hand, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the steps of the transaction risk control method based on biometrics are implemented.

[0019] The above-mentioned transaction risk control method, device, computer equipment, storage medium and computer program product based on biometric identification, by identifying the target part data collected from the target object, determine the target account that matches the target object according to the biometric identification result obtained by identification, and perform preliminary verification in the form of biometric identification. After the target account is obtained through preliminary verification, on the one hand, the operation data associated with the target account is obtained, so as to perform a multi-dimensional risk assessment on the target account and obtain the risk level of the target account; on the other hand, the transaction data associated with the target account is obtained, so as to perform a multi-dimensional risk discrimination on this transaction and obtain the transaction risk result of this transaction. Finally, the risk situation of this transaction is further discriminated in combination with the account risk level and the transaction risk result, and finally the target risk control method of this transaction is comprehensively determined, and the risk control of this transaction is performed according to the target risk control method. Different risk control strategies are implemented for different risk levels, which can effectively control the risk of the entire biometric identification process and improve the security of transactions based on biometric identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is an application environment diagram of a transaction risk control method based on biometrics in one embodiment;

[0021] Figure 2 is a flowchart of a transaction risk control method based on biometric identification in one embodiment;

[0022] Figure 3 A schematic diagram of the principle of risk grading based on risk score distribution in one embodiment;

[0023] Figure 4 It is a schematic diagram of the principle of risk grading based on risk score distribution in another embodiment;

[0024] Figure 5 A schematic diagram of the principle of mining users with high palm print similarity in one embodiment;

[0025] Figure 6 A schematic diagram of a palmprint payment process in an embodiment;

[0026] Figure 7 A schematic diagram of the principle of palmprint image recognition in one embodiment;

[0027] Figure 8 A schematic diagram of the process of risk management and control in a palmprint payment scenario in one embodiment;

[0028] Fig. 9 A schematic diagram of the timing of risk management and control in a palmprint payment scenario in one embodiment;

[0029] Fig.10is a schematic diagram of the principle of a transaction risk control method based on biometric identification in an embodiment;

[0030] Fig.11 is a structural block diagram of a transaction risk control device based on biometrics in one embodiment;

[0031] Fig.12 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0032] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0033] The transaction risk control method based on biometric identification provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the acquisition device 102 is connected to the server 104 for communication. The acquisition device 102 and the server 104 can be directly or indirectly connected by wired or wireless communication, and this application does not limit this. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other servers.

[0034] Among them, the collection device 102 is used to collect the target part data of the target object and transmit the target part data to the server 104. The server 104 identifies the target part data collected from the target object, obtains a biometric recognition result, and determines the target account that matches the target object based on the biometric recognition result. After determining the target account, on the one hand, the server 104 obtains the operation data associated with the target account, and performs a multi-dimensional risk assessment on the target account based on the operation data to obtain the risk level of the target account. On the other hand, the server 104 obtains the transaction data associated with the target account, and performs a multi-dimensional risk assessment on this transaction based on the transaction data to obtain the transaction risk result of this transaction. Finally, the server 104 determines the target risk control method of this transaction based on the account risk level and the transaction risk result, and performs risk management and control on this transaction according to the target risk control method.

[0035] In some embodiments, the acquisition device 102 is used to acquire an image of a target part of a target object. To this end, the acquisition device 102 includes at least an image acquisition element, which is used to capture objects within a visible range. The visible range of the image acquisition element is determined based on the field of view. In one embodiment, the image acquisition element is, for example, a camera, a video camera, a camera module integrated with an optical system or a CCD (Charge Coupled Device) chip, or a camera module integrated with an optical system and a CMOS (Complementary Metal Oxide Semiconductor) chip. The image acquisition element can be integrated into the acquisition device 102 or can be set independently of the acquisition device 102. For example, the image acquisition element can be externally connected to the acquisition device 102 and communicated in a wired or wireless manner.

[0036] In one embodiment, the acquisition device 102 may also include a display element. The display element is used to provide an interface for the target object to view the status of the target part currently being acquired in real time, or it may also be used to display prompt messages to the target object to interact with the target object, etc. In one embodiment, the display element is, for example, a liquid crystal display screen or a projector. Among them, the display element may be integrated into the acquisition device 102, or it may be set independently of the acquisition device 102. For example, the display element may be set externally to the acquisition device 102, and communicated in a wired or wireless manner.

[0037] In some embodiments, the acquisition device 102 may also be, but is not limited to, one or more of various desktop computers, laptop computers, smart phones, tablet computers, IoT devices, or portable wearable devices, etc. The IoT devices may be one or more of smart speakers, smart TVs, smart air conditioners, or smart car-mounted devices, etc. The portable wearable device may be one or more of smart watches, smart bracelets, or head-mounted devices, etc.

[0038] In some embodiments, the collection device 102 may be provided with a variety of APP (Application) applications, and provide transaction services, such as payment services, to the objects through the applications.

[0039] Among them, server 104 can be an independent physical server, or 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 communications, middleware services, domain name services, security services, CDN (Content Delivery Network), as well as big data and artificial intelligence platforms.

[0040] Big data refers to a collection of data that cannot be captured, managed, and processed by conventional software tools within a certain time frame. It is a massive, high-growth, and diversified information asset that requires new processing models to have stronger decision-making power, insight discovery, and process optimization capabilities. Technologies applicable to big data include large-scale parallel processing databases, data mining, distributed file systems, distributed databases, cloud computing platforms, the Internet, and scalable storage systems.

[0041] In one embodiment, Figure 2 As shown in FIG. 1 , a transaction risk control method based on biometric identification is provided. The method can be executed by a server alone, or by a collection device and a server in collaboration; if the hardware conditions are sufficient, the collection device alone can also be executed. Figure 1 The computer device in the example is used to illustrate, and the computer device includes a server or a collection device. The method includes the following steps:

[0042] Step S202, identifying the target part data collected from the target object to obtain a biometric recognition result, and determining a target account that matches the target object based on the biometric recognition result.

[0043] The target object refers to the target living body currently being collected, such as a person. The target part refers to the part of the target object, such as the face, palm, eye, fingertip, etc.

[0044] In some embodiments, the acquisition device responds to the transaction initiation request, acquires the target part of the target object, obtains the target part data, and sends the target part data to the computer device for processing. The target part data can be an image, such as a facial image, a palm print image, etc.; the target part data can also be point cloud data, such as three-dimensional point cloud data, etc.

[0045] It should be noted that the acquisition device can directly transmit the acquired original target part data to the computer device, or can pre-process the target part data before transmitting it to the computer device. The pre-processing includes but is not limited to cropping, rotation, or binarization.

[0046] Specifically, the computer device receives the target part data collected from the target object, and identifies the target part data to obtain a biometric recognition result. The computer device identifies the target part data by performing image recognition on an image of the target part, or by modeling point cloud data or converting it into a depth image for image recognition. In a manner not limited to the above, the computer device identifies the target part data and extracts part features corresponding to the target part data. Part features are, for example, one or more of facial features, palm print features, iris features, or fingerprint features.

[0047] Based on the biometric recognition result, the computer device determines the target account that matches the target object. Specifically, the computer device uses the account to which each preset biological feature indicated in the biometric recognition result belongs as a candidate account. The computer device further screens the candidate accounts, thereby determining the final account as the target account. The target account is the account that matches the target object. Thus, the preliminary verification based on the biometric collection data is completed.

[0048] In some embodiments, for the identified part features, the computer device will search in the database whether there is a preset biometric feature that matches it to obtain a biometric recognition result. The preset biometric feature is a stored biometric feature, including but not limited to one or more of facial features, palm print features, iris features, or fingerprint features. The preset biometric features entered by each object before this collection can be used to form a biometric feature library.

[0049] Exemplarily, the computer device matches the identified part features with each preset biometric feature pre-stored in the biometric feature library, and uses the matched preset biometric feature as the biometric recognition result of the target part data.

[0050] In some embodiments, assuming that there is no preset biometric feature that matches it in the biometric feature library, it means that the target object may be performing the corresponding biometric recognition for the first time, and the computer device can create a new account for the target object and use the new account as the target account that matches the target object. In addition, the computer device associates the biometric feature recognized for the first time with the target account and stores it in the biometric feature library for the next matching.

[0051] Step S204, obtaining operation data associated with the target account, and performing a multi-dimensional risk assessment on the target account based on the operation data to obtain a risk level of the target account.

[0052] The operation data refers to the data generated by the target account in various historical operations. The historical operations include but are not limited to one or more of login operations, identification operations, matching operations, or social operations.

[0053] Login operation data includes but is not limited to one or more of when and where (for example, one or more of the province, city, district, business district, or specific IP address), or what device (for example, one or more of the device type, device model, or device address) the subject logs into his account. Big data analysis may also include the time and frequency corresponding to the account login address, login device, or usage environment, for example, through big data analysis or threshold analysis, it is determined that the frequently logged-in device associated with the target account is device A.

[0054] Identification operation data includes, but is not limited to, historical biometric identification results, or one or more of the merchants that have been identified. Matching operation data includes, but is not limited to, historical account matching results, or one or more of the scores obtained for each match. Social operations include, but are not limited to, social payment relationships, or one or more of the activity statistics of objects and merchants.

[0055] In some embodiments, various types of operational data are stored in a big data storage platform and analyzed through a big data analysis platform.

[0056] Specifically, after determining the target account that matches the target object, the computer device obtains the operation data associated with the target account, which may include historical operation data or current operation data. When the computer device is a collection device, the collection device sends a request through the server to receive the operation data associated with the target account. Based on the operation data, the computer device evaluates the risk of the target account from multiple dimensions, thereby obtaining the risk level of the target account.

[0057] In some embodiments, the computer device evaluates the risk of the target account from multiple dimensions to obtain the evaluation scores of the target account in each dimension; then, based on the evaluation scores of the target account in each dimension, the risk level of the target account is comprehensively determined. Exemplarily, the risk level can be set to a high risk level, a medium risk level, and a low risk level, etc. Alternatively, the risk level can also be set to L1 level, L2 level...Ln level, etc. Each risk level corresponds to its own evaluation score, or evaluation score range.

[0058] Step S206, acquiring transaction data associated with the target account, and performing multi-dimensional risk identification on the transaction based on the transaction data to obtain a transaction risk result of the transaction.

[0059] Among them, transaction data includes but is not limited to what kind of transaction (transaction type, such as online transfer or offline code scanning, etc.) was conducted with the target account at what time, where, and with which merchant (for example, including one or more of merchant type, merchant name, or merchant code, etc.), the transaction amount, frequency, frequency of transactions with each merchant, or historical intercepted transactions, etc. In short, transaction data includes but is not limited to one or more of abnormal transaction amount, abnormal transaction frequency, abnormal transaction relationship, or risky transaction behavior, etc.

[0060] Specifically, after determining the target account that matches the target object, the computer device obtains the transaction data associated with the target account, which may include historical transaction data or current transaction data. When the computer device is a collection device, the collection device sends a request to the server to receive the transaction data associated with the target account, or obtains the transaction data associated with the target account by itself. For example, when the object pays offline by face recognition, the collection device obtains the transaction amount data based on the amount entered by the merchant on the collection device.

[0061] The computer device determines the risk of the target account from multiple dimensions based on the transaction data, thereby obtaining the transaction risk result of the transaction. In some embodiments, the transaction risk result includes but is not limited to one or more of abnormal transaction amount, abnormal transaction frequency, abnormal transaction relationship, risky transaction behavior, etc.

[0062] Step S208, based on the account risk level and the transaction risk result, determine the target risk control method for this transaction, and perform risk management and control on this transaction according to the target risk control method.

[0063] Among them, various risk control methods are preset, and each risk control method stipulates specific measures for risk control of transactions, such as no control, additional verification, limit amount, increase limit, interception of transactions, message reminders, asset freezing, or account ban, etc. One or more of the above.

[0064] Specifically, the computer device selects one or more of the preset risk control methods as the target risk control method for this transaction according to the account risk level and the risk degree of this transaction represented by any one or a combination of the transaction risk results. Thus, the computer device executes the specific measures specified by the target risk control method for this transaction according to the determined target risk control method.

[0065] For example, if the risk level of this transaction represented by the computer device according to the account risk level and any one or combination of the transaction risk results is low, the target risk control method of this transaction can be determined as no control, and the transaction can be directly executed; or the transaction limit of the target account can be increased. In this way, the impact of various risk controls on transactions of low-risk accounts can be avoided, the payment experience of the object can be improved, and the object can also be encouraged to maintain good transaction habits to keep the account in a low-risk state.

[0066] For another example, if the computer device determines that the risk level of this transaction is medium based on the account risk level and any one or combination of the transaction risk results, the target risk control method for this transaction can be determined as additional verification (for example, through a verification code or link, etc.) or limiting the amount, etc., thereby improving the security of this transaction.

[0067] For another example, if the risk level of this transaction is high according to the account risk level and any one or combination of the transaction risk results, the computer device can determine that the target risk control method for this transaction is to intercept this transaction, or to block the account when the risk level is abnormally high. In this way, high-risk transaction behaviors can be intercepted in a timely manner to avoid abnormal asset flow in the target account, effectively protect the security of the target account, and safeguard the assets of the target.

[0068] In the above-mentioned transaction risk control method based on biometrics, the target part data collected from the target object is identified, and the target account matching the target object is determined according to the biometric results obtained by the identification, and preliminary verification is performed in the form of biometrics. After the target account is obtained through preliminary verification, on the one hand, the operation data associated with the target account is obtained, so as to conduct a multi-dimensional risk assessment on the target account and obtain the risk level of the target account; on the other hand, the transaction data associated with the target account is obtained, so as to conduct a multi-dimensional risk discrimination on this transaction and obtain the transaction risk result of this transaction. Finally, the risk situation of this transaction is further discriminated in combination with the account risk level and the transaction risk result, and finally the target risk control method of this transaction is comprehensively determined, and the risk control of this transaction is carried out according to the target risk control method. Different risk control strategies are implemented for different risk levels, which can effectively control the risk of the entire biometric process and improve the security of transactions based on biometrics.

[0069] In some embodiments, target risk control methods can be divided into mitigation control methods and mandatory control methods, etc.

[0070] The mitigation control method represents the degree of risk that falls within the scope of preset control. Under the mitigation control method, the risk control result of the computer equipment for this transaction is to execute this transaction, but in some cases this transaction may be restricted, such as limiting the maximum transaction amount of this transaction.

[0071] The mitigation control method includes but is not limited to one or more of an additional verification method or a limit update method. The additional verification method refers to requiring the object to perform additional verification to ensure the security of payment. The limit update method refers to adjusting the limit that can be used for this transaction, which can be an increase in the limit or a limit on the limit.

[0072] The mandatory control method indicates that the risk level exceeds the preset control scope. Under the mandatory control method, the risk control result of the computer equipment for this transaction is to reject this transaction, such as intercepting transaction orders or banning accounts.

[0073] Among them, the mandatory control methods include but are not limited to one or more of the following methods: transaction interception or account ban. The transaction interception method refers to intercepting the transaction to prevent abnormal outflow of funds. The account ban method refers to banning the target account or freezing the account assets associated with the target account.

[0074] In some embodiments, the target risk control method of the transaction is determined based on the account risk level and the transaction risk result, including: when the risk level represented by the account risk level and the risk transaction result belongs to the preset control scope, the target risk control method of the transaction is determined to be a moderate control method, and the moderate control method includes at least one of an additional verification method and a quota update method; when the risk level represented by the account risk level and the risk transaction result exceeds the preset control scope, the target risk control method of the transaction is determined to be a compulsory control method, and the compulsory control method includes at least one of a transaction interception method and an account ban method.

[0075] Specifically, the computer device determines the target risk control method for this transaction according to the risk level of this transaction represented by the account risk level or any one of the transaction risk results. For example, when the account risk level or any one of the transaction risk results represents a high risk level for this transaction, such as the account risk level is a high risk level, or the transaction risk results contain any of abnormal transaction amount, abnormal transaction frequency, abnormal transaction relationship, risky transaction behavior, etc., the computer device determines that the target risk control method for this transaction is a moderate control method or a mandatory control method.

[0076] In other embodiments, the computer device determines the target risk control method of this transaction according to the risk level of this transaction represented by the combination result of the account risk level or the transaction risk results. For example, the account risk level is preset with a high risk level, a medium risk level, and a low risk level, and the transaction risk results include normal transaction amount, abnormal transaction amount, normal transaction frequency, abnormal transaction frequency, normal transaction relationship, abnormal transaction relationship, non-risk transaction behavior, and risky transaction behavior, etc., then the account risk level and the transaction risk result can be combined to obtain at least 24 combination results. Therefore, according to the combination result, the computer device determines that the target risk control method of this transaction is a moderate control method or a mandatory control method.

[0077] In the above embodiment, by combining the account risk level and the transaction risk result, a systematic risk control strategy is provided, and different risk control methods are adopted under different risk levels. Transaction behaviors can be controlled in a targeted manner, and while ensuring the convenience and efficiency of object payment, the security of biometric payment can be effectively guaranteed.

[0078] In some embodiments, as the risk level of account transactions of each object is updated, the available credit limit of each account set by default can also be dynamically adjusted accordingly. For example, as shown in Tables 1 and 2 below, Table 1 is, for example, the initial average credit limit, and Table 2 is, for example, the adjusted average credit limit. The computer device can pre-set credit limit thresholds of different gears for each risk level L1 to L6. As the operation data and transaction data associated with the account of each object are updated, the credit limit thresholds of different gears are also dynamically adjusted accordingly. Among them, the credit limit thresholds of each gear can be determined according to the corresponding ranking of each gear. For example, the highest 20% of the credit limit corresponding to all object accounts is divided into a high-level credit limit, the second highest 20% to 60% is divided into a medium-level credit limit, and the rest is a low-level credit limit, and the average credit limit is set for each level with the average of each gear, that is, the credit limit of this transaction, etc.

[0079] grade <500 500-700 700-800 800-900 900-1000 >1000 L1 100 200 300 400 500 600 L2 100 300 200 200 200 600 L3 200 400 300 300 400 700 L4 200 400 400 500 600 800 L5 300 600 600 700 800 1000 L6 500 700 800 900 1000 2000

[0080] Table 1

[0081]

[0082]

[0083] Table 2

[0084] Compared with facial collection, palm collection is usually more private and portable. Taking the target part data as a palm print image as an example, in some embodiments, the target part data collected from the target object is identified to obtain a biometric recognition result, including: obtaining a target palm print image collected from the target object, the target palm print image including at least one of a color palm print image, a depth palm print image, or an infrared palm print image; searching for a preset palm print image matching the target palm print image based on a preset palm print image library to obtain a matching result, and determining a biometric recognition result based on the matching result.

[0085] The target palmprint image refers to a palmprint image collected from a target object. According to the hardware conditions of the acquisition device in the actual scene, the target palmprint image can be a multimodal image, including but not limited to one or more of a color palmprint image taken by a camera, a depth palmprint image collected by a depth camera, or an infrared palmprint image collected by an infrared camera.

[0086] Specifically, the computer device acquires the target palm print image of the target object acquired by the acquisition device, recognizes the target palm print image, and extracts the palm print features. The biometric feature library includes, for example, a palm print image library, and the palm print images in the palm print image library are referred to as preset palm print images. In some embodiments, out of consideration for the privacy protection of the object, the palm print image library may only store the palm print feature images without retaining the original palm print images. The computer device searches the preset palm print image library to see whether there is a preset palm print image that matches the target palm print image to obtain a matching result. The computer device uses the matching result as the biometric print recognition result.

[0087] The computer device matches the identified palm print features with the palm print features of each preset palm print image pre-stored in the biometric database, and uses the matched preset palm print image as the biometric recognition result of the target part data.

[0088] The matching between palm print features may be to calculate the similarity between two palm print features, etc. Exemplarily, the computer device matches the identified palm print features with the palm print features of each preset palm print image pre-stored in the biometric database, calculates the matching scores corresponding to each preset palm print image, and thereby selects the top N (N is the number of selections, which can be set according to actual conditions) with the highest similarity, or M preset palm print images greater than the similarity threshold, as the biometric recognition result of the target part data.

[0089] In some embodiments, assuming that there is no preset palm print image matching it in the palm print image library, it means that the target object may be swiping its palm for the first time, or is in the process of activating the palm swiping payment function. Then the computer device can create a new account for the target object and use the new account as the target account to match the target object. In addition, the computer device associates the palm print image recognized for the first time with the target account and stores it in the palm print image library for the next matching.

[0090] In the above embodiment, the identity of the target object is initially verified through multimodal palm print image recognition, which can perform basic risk identification and prevent the identity of the object from being forged to a certain extent, thereby ensuring the account security and asset security of the object.

[0091] After the initial verification, although the transaction security can be guaranteed to a certain extent, the biometric recognition scheme may encounter the problem of being unable to recognize or misrecognizing when encountering similar palm prints, forged palm prints, stolen accounts, etc. In order to solve this problem and truly improve transaction security from the root, the embodiments of the present application provide some systematic security authentication and risk control methods based on biometric recognition to maximize transaction security.

[0092] To this end, in some embodiments, determining the target account that matches the target object based on the biometric recognition result includes: when the biometric recognition result indicates the presence of at least one preset biometric feature that matches the target part data, obtaining at least one candidate account that matches the preset biometric feature; obtaining verification information associated with the target object, and based on the verification information, determining the target account that matches the target object from the at least one candidate account.

[0093] Specifically, after obtaining the biometric identification result, the computer device searches for the target account according to the biometric identification result. That is, the computer device determines the accounts that match the preset biometric features, i.e., the candidate accounts, according to the matching preset biometric features indicated in the biometric identification result, and then further determines the final target account from the candidate accounts.

[0094] Among them, the biometric recognition result may indicate that there is one preset biometric feature that matches the target part data, there are two or more preset biometric features that match the target part data, or there is no preset biometric feature that matches the target part data, etc.

[0095] In the case where the biometric recognition result indicates the existence of a preset biometric feature that matches the target part data, the computer device can directly determine, based on the matching preset biometric feature, the corresponding candidate account as the target account that matches the target object.

[0096] In the case where the biometric recognition result indicates that there are two or more preset biometric features that match the target part data, the computer device obtains each candidate account that matches the preset biometric feature for further confirmation. The computer device can obtain verification information associated with the target object and perform further confirmation based on the verification information, thereby determining the target account that matches the target object from among the candidate accounts.

[0097] The verification information may be obtained based on the device-related information uploaded by the acquisition device. The device-related information includes but is not limited to one or more of device attributes, device environment, device model, or device address. For example, the computer device obtains the device environment uploaded by the acquisition device, and then further matches each candidate account based on the device environment, thereby determining the target account from each candidate account.

[0098] In some embodiments, the verification information may also be obtained by interacting with the target object. For example, the computer device sends a verification request to the collection device for secondary verification; the computer device receives the secondary verification related information returned by the collection device, and determines the target account from each candidate account accordingly.

[0099] In the above embodiment, when multiple similar identification results are obtained through biometric identification, further verification is performed based on the verification information to determine the final target account number. This can eliminate problems such as similar biometric features and inaccurate identification, and prevent others from impersonating the person for biometric verification, thereby effectively ensuring the security of the transaction and the account and asset security of the object.

[0100] In some embodiments, determining a target account that matches the target object from among the at least one candidate account based on the verification information includes: obtaining account feature information associated with each candidate account; and based on a match between the verification information and each account feature information, using the matching candidate account as a target account that matches the target object.

[0101] Specifically, when each candidate account is obtained, the computer device obtains account feature information associated with each candidate account. The account feature information characterizes the characteristics of each candidate account formed under various historical activities, such as the characteristics of the device used in transaction activities, or the characteristics of the region in which the social activities are located, or the biometric features historically entered during biometric activities. Thus, the computer device can analyze the historical information of each candidate account, thereby determining the candidate account that is most likely to be the target account among the candidate accounts.

[0102] The computer device matches the verification information with the account feature information of each candidate account, determines whether there is a matching candidate account, and uses the matching candidate account as the target account that matches the target object.

[0103] In some cases, the verification information includes current device related information uploaded by the acquisition device that acquires the target part data, and the account characteristic information includes at least historical device related information. Accordingly, based on the matching of the verification information and each account characteristic information, the matching candidate account is used as the target account that matches the target object, including: determining the target device corresponding to each candidate account according to the historical device related information corresponding to each candidate account; determining whether the acquisition device matches the target device based on the current device related information; and using the candidate account corresponding to the target device that matches the acquisition device as the target account that matches the target object.

[0104] In this case, the computer device analyzes the historical device related information corresponding to each candidate account to determine the target device corresponding to each candidate account, wherein the target device can be a device that meets a certain frequency obtained based on the historical device related information.

[0105] For example, the computer device determines the device model commonly used by each candidate account based on the historical device attribute information (including one or more of device ID, device address, or device model, etc.) corresponding to each candidate account, and then uses the most commonly used device as the target device. The device address is, for example, a MAC (Media Access Control Address) address.

[0106] In addition, the target device corresponding to each candidate account can also be determined according to the regional location of the device. The regional location is, for example, an IP (Internet Protocol Address) address, a WI-FI address, a GPS (Global Positioning System) location, or an LBS (Location Based Services) location.

[0107] Therefore, the computer device compares the current device related information uploaded by the acquisition device with the target devices corresponding to each candidate account to determine whether the acquisition device matches the target device. The computer device uses the candidate account corresponding to the target device that matches the acquisition device as the target account that matches the target object.

[0108] For example, the target device of candidate account S1 is a T1 model smartphone, which often conducts transactions in area A, and the target device of candidate account S2 is a T2 model smartwatch, which often conducts transactions in area B. The computer device determines whether the acquisition device matches the target device of candidate account S1 or the target device of candidate account S2 based on the current model or area of ​​the acquisition device. Assuming that the acquisition device matches the target device of candidate account S2, the computer device uses candidate account S2 as the target account that matches the target object.

[0109] Therefore, based on the current device-related information uploaded by the current collection device and compared and matched with the historical device-related information of each candidate account, the target account associated with the target object can be automatically determined, and the risk of the target account can be accurately judged.

[0110] In other cases, the verification information includes cross-validation information obtained in response to a cross-validation operation. For example, the computer device may send a cross-validation request to the acquisition device to cause the acquisition device to trigger a cross-validation operation and interact with the target object again to obtain cross-validation information. For another example, the computer device may trigger multiple verifications of the target part data of multiple modalities originally collected in response to a cross-validation operation.

[0111] Wherein, the cross-verification information at least includes the current biometric features that are distinguished from the target part data. For example, if the target part data originally collected is a palm print image, an image that is distinguished from the palm print image should be obtained during cross-verification, which can be a palm print image of another hand, or a facial image, an iris image, or a fingerprint image, etc., to obtain biometric features that are distinguished from the palm print features of the palm print image, thereby improving the accuracy and security of cross-verification. Accordingly, the account feature information corresponding to each candidate account at least includes historical biometric features, such as one or more of the facial features of the historically recorded facial image, the palm print features of the palm image, the fingerprint features of the finger image, or the iris features of the eye image.

[0112] Correspondingly, based on the matching between the verification information and the characteristic information of each account, the matching candidate account is used as the target account that matches the target object, including: matching the current biometric feature with the historical biometric feature associated with each candidate account; and the candidate account associated with the historical biometric feature that matches the current biometric feature is used as the target account that matches the target object.

[0113] Specifically, the computer device matches the current biometric feature with the historical biometric features associated with each candidate account respectively; and uses the candidate account associated with the historical biometric feature that matches the current biometric feature as the target account that matches the target object.

[0114] The current biometric feature is matched with the historical biometric features associated with each candidate account respectively; the candidate account associated with the historical biometric feature that matches the current biometric feature is used as the target account that matches the target object.

[0115] Therefore, secondary verification is performed based on the cross-verification information obtained through interaction with the object, and the account is cross-confirmed based on multiple types of biometric information, which can improve the accuracy of matching the target account. In addition, the risk of others impersonating or attacking the account can be avoided through secondary verification, effectively improving the security of the target account.

[0116] After determining the final target account, preliminary verification based on biometrics was achieved. On this basis, risk assessment and identification were carried out to determine the specific risk management method, so as to carry out targeted risk management for this transaction.

[0117] To this end, on the one hand, the computer device performs a multi-dimensional assessment of the risk level of the target account based on the operational data related to the target account.

[0118] In some embodiments, the obtaining of operation data associated with the target account and performing a multi-dimensional risk assessment on the target account based on the operation data to obtain the risk level of the target account includes: obtaining historical operation data associated with the target account; performing a multi-dimensional risk assessment on the target account based on the historical operation data, at least from an account behavior dimension, a remote jump dimension, a personnel anomaly dimension, and an account attack dimension to obtain an assessment score corresponding to each dimension; and obtaining the risk level of the target account based on the assessment score corresponding to each dimension.

[0119] Specifically, the computer device obtains historical operation data associated with the target account, wherein the historical operation data is operation data of historical records. The computer device performs risk assessment from various dimensions based on the historical operation data of the target account to obtain a more accurate risk assessment result.

[0120] The dimensions of risk assessment include, but are not limited to, one or more of the account behavior dimension, the remote jump dimension, the personnel abnormality dimension, or the account attack dimension. The account behavior dimension includes, for example, whether the target account has a risk of being stolen. The remote jump dimension includes, for example, whether the target account has a risk of remote login. The personnel abnormality dimension includes, for example, whether the target account is operated by the person himself. The account attack dimension includes, for example, whether the target part data is forged, or whether the target account is frequently used multiple times.

[0121] In addition to the dimensions mentioned above, more or fewer dimensions can be set according to actual needs to consider complex and changeable transaction situations in actual scenarios, thereby improving the accuracy of risk assessment of the target account.

[0122] Taking the above dimensions as an example, the computer device performs multi-dimensional risk assessment on the target account from the dimensions of account behavior, remote jump, personnel anomaly, and account attack, thereby obtaining the assessment scores corresponding to each dimension. Then, the computer device can finally assess the risk level of the target account based on the assessment scores corresponding to each dimension.

[0123] In the above embodiment, by judging the current risk situation of the target account based on the historical operation data of the target account, the activity situation and historical records of the target account over a period of time can be integrated, the impact of a single abnormality on the overall risk level of the target account can be avoided, and the transaction can be judged based on the overall historical situation, so as to better assess the risk of the transaction and provide more accurate risk control measures.

[0124] Wherein, in the case of setting various dimensions, the weight coefficients of the various dimensions are also set respectively to calculate the final risk level. To this end, in some embodiments, the risk level of the target account is obtained based on the evaluation scores corresponding to the various dimensions, including: based on the weight coefficients pre-set for each dimension, the evaluation scores corresponding to each dimension are weighted summed to obtain the risk score of the target account; based on the comparison result of the risk score and the preset risk classification, the risk level of the target account is obtained.

[0125] The preset risk classification is obtained by constructing a risk score distribution based on the historical evaluation scores of multiple accounts. For a period of time in history, the computer device can establish a risk score distribution based on the historical evaluation scores of each account, and classify the risk scores based on the classification criteria, thereby obtaining each preset risk classification.

[0126] For example, Figure 3 and Figure 4 As shown, the computer device can use the historical evaluation scores of multiple accounts as a benchmark, perform normal distribution standardization, and classify them according to the risk standard evaluation method. For example, the KS (Kolmogorov-Smirnov, a discrimination evaluation index) value is used as the classification standard. For example, for Figure 3 As shown in the data graph, the best cut point is around 0.45. Generally, the middle value or the low-risk customer group takes 0.45 as the grade standard, and then the normal distribution simulation is performed according to the object magnitude to make the final classification. Figure 4 As shown, the AUC (Area Under Curve, the area under the ROC curve and enclosed by the coordinate axis) value can also be used as the basis for division.

[0127] According to the risk levels L1 to LN, for each account i, the grading formula is, for example:

[0128] L i =[F i ,F i+1 )i∈[1,2,3…N]

[0129] Specifically, after obtaining the evaluation scores corresponding to each dimension, the computer device performs weighted summation based on the weight coefficients pre-set for each dimension to obtain the risk score of the target account. Finally, the computer device compares the risk score with the preset risk classification, thereby obtaining the risk level of the target account according to the comparison result.

[0130] Therefore, by analyzing and constructing the distribution of huge amounts of data through big data analysis, historical data can be fully utilized. At the same time, since historical data can be updated over time, the constructed risk score distribution will also be dynamically adjusted, which can improve the accuracy of risk assessment.

[0131] In some embodiments, the evaluation scores of each dimension can be obtained by classifying the classification models pre-trained for each dimension. Exemplarily, the classification models corresponding to each dimension include, but are not limited to, one or more of an account behavior model, a remote jump model, a personnel anomaly model, or an account attack model. Each classification model can be a logistic regression, an ensemble learning XGB (eXtreme Gradient Boosting, extreme gradient boosting) classification model, etc., and this application does not limit this.

[0132] For the account behavior model, the historical operation data mainly used include but are not limited to one or more of common environmental information, common equipment, commonly identified merchants, time and frequency of historical transactions, or amount distribution of historical transactions. Exemplarily, an account behavior model is pre-established, and the data of accounts with historical theft anomalies are used as negative samples, and negative samples are sampled from the data of normal sample accounts for training. Exemplarily, the account behavior model ultimately outputs an abnormal probability value of 0-1 for subsequent risk assessment.

[0133] The environmental information includes but is not limited to one or more of the following: installing a Frida plug-in (a reverse analysis plug-in), a Hook plug-in (a redirection plug-in), being in a Root (root authority) environment, or having the device transmission data repackaged.

[0134] For the off-site jump model, the historical operation data mainly used include but are not limited to one or more of common environment information, common equipment, common identification object merchants, common regional location lists, or real-time multi-location jump locations, etc. Exemplarily, the off-site jump model finally outputs an abnormal probability value of 0-1, for example, new equipment and the current new province, new city, and new merchant have a high risk value of 1, and the risk of common equipment but new area is a medium value of 0.5, and so on.

[0135] For the personnel anomaly model, historical operation data is mainly combined with the real-name identity information (such as identity ID, etc.) and account ID of the current target account for joint matching. The historical operation data mainly used by the personnel anomaly model include but are not limited to one or more of common environmental information, common equipment, commonly identified merchants, time and frequency of historical transactions, or amount distribution of historical transactions. Exemplarily, a personnel anomaly model is established in advance, and the data of abnormal accounts that have been used by someone other than the person himself or by multiple people are used as negative samples, and negative samples are sampled from the data of normal sample accounts for training. Exemplarily, the off-site jump model ultimately outputs an abnormal probability value of 0-1.

[0136] For the account attack model, the historical operation data mainly used include but are not limited to one or more of device risk labels, transaction request frequency, liveness detection results, common environment information, common equipment, commonly identified merchants, historical transaction time, historical transaction frequency, or historical transaction amount. Exemplarily, a personnel anomaly model is pre-established, with the data of accounts that have been attacked in the past as negative samples, and negative samples are sampled from the data of normal sample accounts for training. Exemplarily, assuming that the abnormal probability value output by the personnel anomaly model is >0.9, and combined with the device environment, it is determined that the risk is root risk. In the case of multiple frequent operations, it can be determined that the risk of account attack is high.

[0137] After obtaining the abnormal probability values ​​output by each classification model, the evaluation scores corresponding to each dimension are obtained. The computer device then performs weighted summation of each abnormal probability value according to the weight coefficient to obtain the risk score of the target account. For example, the computer device can use the following formula for calculation:

[0138]

[0139] Among them, a i is the weight coefficient corresponding to each classification model. The weight coefficient can be dynamically adjusted according to the actual scenario. i ) is the abnormal probability value output by the i-th classification model. N is the total number of classification models.

[0140] In order to make the assessment process as risk-sensitive as possible, in some cases, the risk level of the target account is obtained based on the assessment scores corresponding to each dimension, including: when the assessment score corresponding to any dimension exceeds the threshold under the corresponding dimension, the risk level of the target account is obtained based on the assessment score exceeding the threshold.

[0141] Specifically, if in the process of calculating the evaluation scores of each dimension, the evaluation scores corresponding to any one or more dimensions calculated exceed the threshold under the corresponding dimension, it means that the probability of abnormal risk under this dimension is very high, so it can be directly determined that there is a corresponding risk under this dimension. The thresholds of each dimension can be set adaptively, for example, the threshold of the account behavior dimension is 0.8, the threshold of the remote jump dimension is 0.85, etc.

[0142] Accordingly, the risk level of the target account can be directly set to a certain level, or the risk level of the target account can be increased according to the evaluation score exceeding the threshold, etc. That is, the computer device obtains the risk level of the target account according to the evaluation score exceeding the threshold.

[0143] Therefore, through the above method, we can remain highly sensitive to various risk situations as much as possible, avoid any possible risks as much as possible, and manage possible risks, which effectively improves the security of transactions and protects the account and asset security of the object.

[0144] On the other hand, the computer equipment makes a multi-dimensional assessment of the risk of the transaction based on the transaction data related to the target account.

[0145] In some embodiments, the acquiring of transaction data associated with the target account, and performing multi-dimensional risk assessment on the transaction based on the transaction data to obtain a transaction risk result for the transaction, includes: acquiring historical transaction data and current transaction data associated with the target account; based on the historical transaction data and current transaction data, performing multi-dimensional risk assessment on the transaction at least from the transaction amount dimension, the transaction frequency dimension, and the transaction relationship dimension to obtain risk assessment results corresponding to each transaction dimension; acquiring a risk label associated with the target account; obtaining a risk assessment result under the label dimension based on the risk label; and obtaining a transaction risk result for the transaction based on the risk assessment results corresponding to each dimension.

[0146] Specifically, the computer device obtains transaction data associated with the target account. The transaction data includes historical transaction data and current transaction data. The computer device performs risk identification from various transaction dimensions based on the historical transaction data and current transaction data of the target account to obtain a more accurate risk identification result.

[0147] Among them, the transaction dimensions for risk identification include but are not limited to one or more of the transaction amount dimension, transaction frequency dimension, or transaction relationship dimension, etc. Correspondingly, the risk identification result includes one or more of normal transaction amount, abnormal transaction amount, normal transaction frequency, abnormal transaction frequency, normal transaction relationship, or abnormal transaction relationship, etc.

[0148] The transaction amount dimension determines whether the amount of this transaction exceeds the normal range from the perspective of the transaction amount of each transaction (including this transaction), for example, whether the amount of this transaction exceeds the maximum value obtained from historical statistics. The transaction data used in the transaction amount dimension include but are not limited to one or more of the historical daily transaction amount, the historical monthly transaction amount, etc.

[0149] The transaction frequency dimension determines whether the current transaction is an abnormal frequency transaction operation from the perspective of the frequency of historical transactions. For example, whether the time difference between the current transaction and the historical transaction is less than the threshold, if so, it is determined to be abnormal frequency. The transaction data used in the transaction frequency dimension includes but is not limited to one or more of the historical transaction time, historical daily transaction frequency, or historical monthly transaction frequency.

[0150] The transaction relationship dimension determines whether there is a risk in the transaction relationship of this transaction from the perspective of the merchant and environment at the time of the transaction, for example, whether it is a brand new payment relationship, etc. The transaction data used in the transaction relationship dimension includes but is not limited to one or more of historical identification of merchants, historical regions / locations, or historical multi-location adjustment locations.

[0151] In addition to the transaction dimensions mentioned above, more or fewer transaction dimensions can be set according to actual needs to consider the complex and changeable transaction situations in actual scenarios, thereby improving the accuracy of risk assessment of this transaction.

[0152] Taking the above-mentioned transaction dimension as an example, the computer device conducts multi-dimensional risk assessment on the transaction from the dimensions of transaction amount, transaction frequency, and transaction relationship, thereby obtaining risk assessment results corresponding to each transaction dimension.

[0153] In some embodiments, the computer device may obtain the transaction risk result of the transaction based on the risk identification results corresponding to each transaction dimension. The transaction risk result may be a combination of each risk identification result, or a risk degree obtained according to each risk identification result. Exemplarily, the transaction risk result may also be represented by a pre-classified level.

[0154] In other embodiments, in addition to the above transaction dimensions, the computer device also refers to the risk tag carried by the transaction to comprehensively determine the transaction risk result of the transaction. Specifically, the computer device obtains the risk tag associated with the target account. Risk tags include, for example, illegal tags or batch aggregation operation tags.

[0155] Risk labels can be provided externally, such as obtaining risk labels marked on transactions by banks; they can also be obtained based on historical data analysis, such as judging that an illegal transaction is based on data such as address jumps and frequency anomalies in historical transaction data, and then marking the target account with a risk label.

[0156] Thus, the computer device obtains the risk identification result under the label dimension according to whether the corresponding risk label exists. Corresponding to the risk label, the risk identification result under the label dimension includes illegal transactions, batch aggregation transactions, etc.

[0157] Therefore, based on the risk identification results of each transaction dimension and the risk identification results under the label dimension, the computer device can obtain the risk identification results corresponding to each dimension respectively, and finally obtain the transaction risk result of this transaction.

[0158] In the above embodiment, by judging the risk of the transaction based on the transaction data of the target account, the transaction history records of the target account over a period of time can be integrated to avoid the impact of a single abnormality on the overall risk level of the target account, and the transaction can be judged based on the overall historical situation, so as to better assess the risk of the transaction and provide more accurate risk control measures.

[0159] In some special scenarios, the results of biometrics are extremely similar. For example, the target part data collected is the part data of one of the twins. If they are twins, they may be very familiar with each other's usage habits, or one of them may directly use the other's device to pay, with or without the other's knowledge.

[0160] To this end, in some embodiments, when the biometric recognition result indicates that there is at least one preset biometric feature that matches the target part data, and the similarity between the preset biometric features that match is less than a threshold, after determining the target account that matches the target object based on the biometric recognition result, the method further includes: setting a biometric anomaly label for the target account. The biometric anomaly label is used to indicate that the biometric recognition results are extremely similar, which may be a scenario of twins or multiple births.

[0161] Accordingly, the target risk control method for the transaction is determined based on the account risk level and the transaction risk result, and further includes: determining that the target risk control method for the transaction at least includes a limit update method based on the biometric abnormality tag. Specifically, the computer device determines that the target risk control method for the transaction at least includes a limit update method based on the biometric abnormality tag corresponding to the target account, such as limiting the limit of the transaction.

[0162] In the above embodiment, for the special scenario of twins or multiple births, by setting labels to mark the relevant accounts, targeted risk management and control can be carried out for this scenario, thereby improving the security of the subject's account and assets.

[0163] In some embodiments, after performing transaction control on each object, the computer device also performs abnormal analysis and audit on the transaction. To this end, the above method also includes: in response to a complaint request initiated for risk control of this transaction, re-evaluating at least one of the account risk level or risk transaction results of the target account; based on the re-evaluated evaluation result, updating at least one of the account risk level or risk transaction results of the target account, and re-determining the target risk control method of this transaction based on the updated data.

[0164] Specifically, the computer device may re-evaluate the target account's account risk level, risk transaction results, or account risk level and risk transaction results based on the appeal request initiated for the risk control of this transaction. For example, for a certain transaction, the computer device determines that the target risk control method for this transaction is a limit amount based on the risk level determined by the assessment. If the object files an appeal against the risk control of this transaction, the computer device will re-evaluate based on the appeal request.

[0165] In some embodiments, the computer device updates at least one of the operation data or transaction data of the target account based on the data update request to prevent evaluation errors caused by errors in the data itself. Based on at least one of the updated operation data or transaction data, the account risk level and risk transaction results of the target account are re-evaluated and judged, which can reduce the impact on the current transaction and avoid affecting the transaction experience.

[0166] In other embodiments, the computer device updates the models of each classification model based on the data update request, thereby improving the accuracy of the output risk assessment results, and further improving the accuracy of the judgment of the risk level of the target account. The model parameters of each classification model can be updated based on the current and historical operation data, and new negative samples and positive samples are extracted therefrom to train the classification model, thereby updating the classification model; or based on the updated operation data, new negative samples and positive samples are extracted therefrom, and the classification model is trained, thereby updating the classification model, etc.

[0167] Therefore, based on the evaluation result obtained by the re-evaluation, the computer device updates at least one of the account risk level or the risk transaction result of the target account, and re-determines the target risk control method of this transaction based on the updated data.

[0168] In the above embodiment, by providing a complaint method to re-evaluate and update the risk control method of the transaction, data processing errors can be updated in a timely manner, and the evaluation strategy can be updated in a timely manner, which can improve the accuracy of risk management.

[0169] In some embodiments, the computer device also obtains an updated risk tag provided by an external party to re-determine whether there is illegal transaction or other behavior for the current transaction or historical transaction. Exemplarily, for a target account that is later determined to have illegal transaction or other behavior, the computer device can update the operation data and transaction data of the target account. For example, for a target account that is later determined to have a batch aggregation operation, the computer device updates the transaction frequency in its transaction data.

[0170] In some embodiments, the above method further includes: updating the credit rating corresponding to the target account based on the account risk level of the target account and the transaction risk result. Each credit rating is preset with a corresponding initial credit limit. After the credit rating of the target account is updated, the credit limit of the target account is also updated accordingly. The specific method can refer to the embodiment of the credit limit update in the above embodiment. For example, for a target account that is characterized by a low risk in terms of the account risk level and the transaction risk result, it belongs to a high-quality credit account, so it can be determined that the target risk control method for this transaction is to increase the credit limit, etc.

[0171] It is easy to understand that the above-mentioned types of operation data, transaction data, and device information are only examples and can be appropriately adjusted according to actual conditions in specific application scenarios; those skilled in the art should be aware that reasonable deformation and appropriate adjustments made to the above-mentioned various data and information are within the scope of protection of this application.

[0172] The present application also provides an application scenario, which applies the above-mentioned transaction risk control method based on biometrics. In actual scenarios, data storage and analysis capabilities can be provided by big data cloud servers. Exemplarily, big data storage platforms such as Redis (Remote Dictionary Server), HDFS (Hadoop Distributed File System), etc., big data analysis platforms MapReduce (a programming model, a core component of Hadoop), Spark SQL (Spark component module for structured data processing), Pytorch (open source Python machine learning library), etc.

[0173] Specifically, the cloud server accumulates and summarizes various data based on the operation data and transaction data of each object, thereby obtaining at least the following types of data portraits:

[0174] The basic payment feature portrait includes the transaction amount statistical portrait, which mainly includes monthly / cumulative transaction times / amounts, and social / commercial transaction times / amounts.

[0175] Payment relationship portrait, including social payment relationship, device list, common area list, WI-FI positioning, and LBS positioning, etc.

[0176] Merchant statistical profile, including the subject's frequently used merchant list, the total transaction amount / days with each merchant, and merchant activity statistics, etc.

[0177] Biometric profiling, including identification records, matching scores, and transaction behavior statistics.

[0178] Device security profile, including risk data such as device Frida / Hook / Root / repackaging, as well as the target’s commonly used devices, IoT devices, etc.

[0179] Historical anomaly portraits, including statistics on historical interception, intervention, and continued trading behaviors.

[0180] Similar relationship portraits include target customer groups with highly similar biological characteristics, target customer groups with real-name acquaintance relationships, target customer groups with highly similar payment environment and region, and target customer groups with highly similar or high overlap in social relationships.

[0181] Label portraits, including illegal risk labels, object / merchant labels (including object / merchant labels of different credit levels), complaint risk labels (number of complaints, amount, related complaints, etc.), etc.

[0182] By establishing a full-link risk control portrait of the transaction process, we can build core basic capabilities for the transaction's risk management strategy.

[0183] Exemplarily, as shown in Table 3 below, taking the scenario of palmprint payment as an example, the palmprint payment risk control portrait includes but is not limited to the basic payment risk control portrait, the basic portrait of palmprint payment, the similar relationship behavior portrait, the abnormal risk portrait mining, etc.

[0184]

[0185]

[0186] Table 3

[0187] For example, Figure 5 As shown, by using big data to mine users with high palm print similarity, it is possible to identify customer groups with high stranger similarity and high acquaintance similarity.

[0188] The biometric similarity module mainly determines the similarity of user palmprint features and compares the similarity of palmprint images. In addition, in order to improve accuracy and match new algorithms, the palmprint database will be regularly updated for palmprints that have historical acquaintances, and the relationship database will be updated.

[0189] The relationship similarity module mainly determines user similarity through associated data, such as transaction similarity in environment, transaction relationship intimacy, location similarity in region, social attributes in relationship, and specific time and environment.

[0190] The real-name identity verification module mainly verifies the relationship of acquaintances found through biometric acquaintance and behavioral relationships. The data used include similar names, similar ages, gender distinctions, common regions with the same real names, and binding relationship associations. For example, if the palm prints are similar and the data corresponding to the two accounts are the same age and surname, then there is a high probability that they are twins; they can then be distinguished by gender as twin siblings, brothers, sisters, etc.

[0191] In a specific example, Figure 6 As shown, the application of the transaction risk control method based on biometrics in this application scenario is as follows: Palm swiping multimodal feature acquisition: the acquisition device acquires multimodal palm print images and reports them to the cloud server. Wherein, in the case of cross-verification, facial images can be acquired by the acquisition device for reporting. Palmprint identity recognition security authentication risk control: account matching is performed based on palm print images to determine the identity of the target object, thereby determining the target account that matches the target object. Abnormal risk identification of palm swiping payment transaction behavior: risk assessment and risk identification of the target account are performed based on the operation data and transaction data of the target account. Differentiated identity authentication risk control: dynamic quota control is performed for different risk levels and categories. Differentiated secondary identity authentication is performed when necessary. Comprehensive abnormal risk audit: abnormal audits such as transaction loops and illegal behaviors are performed afterwards. Reverse iterative classification model and strategy matching: dynamic adjustment of classification models and strategies.

[0192] Among them, Figure 7 As shown, in the process of image acquisition and reporting, the multimodal palm print images collected by the acquisition device may include color images, infrared images, and depth images. After being reported to the cloud server, the cloud server extracts biometric features based on a neural network and matches them with the biometric library to determine candidate accounts with similar identities, i.e., recognition results. The cloud server can also perform mutual recognition verification on multimodal images based on a neural network to perform identity cross-verification. If cross-verification is required, the acquisition device can also collect facial images or videos to further determine the target account based on additional biometric features.

[0193] Of course, it is not limited to this. The transaction risk control method based on biometrics provided in this application can also be applied in other application scenarios, such as access control, etc.

[0194] In a specific example, taking palm prints as an example, Figure 8As shown, the transaction risk control method based on biometrics provided in the embodiment of the present application includes:

[0195] The subject scans his palm through the acquisition device, and the acquisition device collects the palm print image.

[0196] The computer device obtains a target palm print image of at least one modality collected from the target object, and searches for a preset palm print image matching the target palm print image based on a preset palm print image library to obtain a matching result, wherein the matching result includes the existence of multiple similar palm prints, the existence of one similar palm print, and the absence of similar palm prints.

[0197] Thus, the computer device determines the palm print recognition result based on the matching result, and determines the target account that matches the target object based on the palm print recognition result. In the case where there are multiple similar palm prints, the computer device determines the object account 1, ... object account N associated with each similar palm print based on similar palm print 1, ... similar palm print N, and then determines the target account.

[0198] In the case that there is a similar palm print, the computer device directly matches the candidate account associated with the palm print image as the target account.

[0199] In the case that there is no similar palm print, the computer device determines that it is the first time to scan the palm, registers the target account for the target object and stores the palm print in association.

[0200] Among them, in the case of multiple similar palm prints, one method is: the computer device determines the target device corresponding to each candidate account according to the historical device related information corresponding to each candidate account corresponding to each similar palm print; based on the current device related information, determines whether the acquisition device matches the target device; and uses the candidate account corresponding to the target device that matches the acquisition device as the target account that matches the target object.

[0201] Another way is: the computer device matches the current biometric feature with the historical biometric features associated with the candidate accounts corresponding to each similar palm print; and uses the candidate account associated with the historical biometric feature that matches the current biometric feature as the target account that matches the target object.

[0202] Therefore, after determining the target account, on the one hand, the computer device obtains the historical operation data associated with the target account, and based on the historical operation data, performs a multi-dimensional risk assessment on the target account from at least the account behavior dimension, the remote jump dimension, the personnel abnormality dimension, and the account attack dimension, and obtains the assessment score corresponding to each dimension. Then, based on the weight coefficients pre-set for each dimension, the computer device performs a weighted summation of the assessment scores corresponding to each dimension to obtain the risk score of the target account; finally, based on the comparison result of the risk score and the preset risk classification, the risk level of the target account is obtained. In particular, when the assessment score corresponding to any dimension exceeds the threshold under the corresponding dimension, the computer device obtains the risk level of the target account based on the assessment score exceeding the threshold.

[0203] On the other hand, the computer device obtains the historical transaction data and the current transaction data associated with the target account; performs multi-dimensional risk identification on the current transaction from at least the transaction amount dimension, the transaction frequency dimension, and the transaction relationship dimension based on the historical transaction data and the current transaction data, and obtains the risk identification results corresponding to each transaction dimension; obtains the risk label associated with the target account; and obtains the risk identification results under the label dimension based on the risk label. Finally, the computer device obtains the transaction risk result of the current transaction based on the risk identification results corresponding to each dimension.

[0204] Therefore, based on the risk level of the target account (such as high risk, medium risk, and low risk) and the transaction risk result of this transaction (including device-related anomalies, transaction-related anomalies, etc.), the computer device finally determines the target risk control method for this transaction, such as one or more combinations of secondary verification, interception of transactions, quota limits, or quota increases, etc. Among them, for highly similar anomalies such as twins (the target account is marked with a biometric anomaly label), at least risk control methods such as quota limits are adopted.

[0205] In a specific example, the overall palm payment technology process is as follows: Fig. 9 As shown, the identification background and risk control services can be set up on the server.

[0206] The acquisition device collects the multimodal image of the user's palm print and initiates a request for activation, identification, or payment. The palm print feature recognition is then performed through the palm print recognition background to identify the object.

[0207] The identification backend aggregates data to initiate a risk control service request. The risk control service sends a risk query request to the feature service, obtains various account, device, relationship, abnormal label and other features of the current identification request through the feature service, and then calls the decision engine for real-time risk identification, and then returns the identification result to the identification backend.

[0208] If the current request needs to be verified, the recognition background will return the verification request to the collection device, and the user will enter the information as required. After submitting the information, the information will be verified, and then the verified result will be returned to the object.

[0209] The final risk determination and identification results will be reported to the data platform. After the data platform conducts data integration and analysis, the calculated features and labels will be updated to the feature service.

[0210] The above method can be applied to the risk control of transactions in large-scale object scenarios, avoiding the problems of similar biometrics, biometric forgery, or account theft that exist in the single use of biometrics, and can provide a systematic payment security authentication and risk control method. While ensuring the convenience of transactions by objects through biometrics, it effectively guarantees the security of transactions and the security of objects' accounts and assets. With the accumulation of payment behaviors, relevant features and strategies are dynamically adjusted to ensure transaction security and convenience.

[0211] Compared with the existing payment security authentication scheme, the above method can more effectively mine biometrics and risky accounts and risky transactions during payment. By grading the risk level and performing differentiated risk control on accounts and transactions with different risk levels, for example, stricter risk control can be performed on high-risk objects, while the payment experience of low-risk and normal objects can be better ensured to avoid excessive intervention, thereby more effectively and specifically responding to different risks. At the same time, through risk level assessment and transaction risk identification, abnormal transaction situations can be intervened in a timely manner, thereby avoiding abnormal asset loss.

[0212] In a specific example, Fig.10 As shown, the embodiment of the present application is mainly divided into palmprint identity identification, risk behavior identification, transaction behavior identification, and risk modeling audit.

[0213] In the palmprint identity identification part, human palm images are collected from multiple angles, and multi-modal palmprint image recognition is performed to cross-verify the user's identity, achieving highly accurate palmprint identity authentication. In this way, the risk of identity forgery is prevented to a certain extent, and highly similar palmprints are matched with the historical palmprint feature library to accumulate palmprint similarity libraries for subsequent high similarity identification.

[0214] In the risk behavior identification part, the user's account and payment behavior are modeled and analyzed, and the account identity is risk graded. This part is mainly based on the various operation data accumulated in the payment scenario, combined with the basic features of the palm print and the similarity relationship library, and uses big data and artificial intelligence analysis technology to comprehensively identify the various risk levels of the user's account in the payment scenario.

[0215] In the transaction behavior identification part, the risk of the transaction is identified in real time based on the current transaction amount, frequency, region, etc., and intervention is made for abnormalities such as high amounts and overfrequency.

[0216] In the risk modeling audit, we mainly conduct offline comprehensive multi-dimensional identification audits of identity anomalies, transaction anomalies, credit anomalies, and fund circuit anomalies. We conduct audits based on the full process data of swiping palm payment to identify suspicious users and transactions, and provide feedback to identify risky behaviors and strategies.

[0217] Therefore, through differentiated risk management, high-risk accounts can be effectively controlled to ensure the payment experience of most normal low-risk customers. Secondary verification can be performed for stranger-like customers to eliminate risks such as mis-swipes; additional verification such as scanning codes for acquaintances with high similarity at the beginning of payment can ensure that the payment is made by the person himself and ensure the safety of funds. With the accumulation of payment behaviors, relevant features and strategies are dynamically adjusted to ensure transaction security and convenience.

[0218] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0219] Based on the same inventive concept, the embodiment of the present application also provides a biometric-based transaction risk control device for implementing the above-mentioned biometric-based transaction risk control method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above-mentioned method, so the specific limitations in one or more embodiments of the biometric-based transaction risk control device provided below can refer to the limitations of the biometric-based transaction risk control method above, and will not be repeated here.

[0220] In one embodiment, Fig.11 As shown, a transaction risk control device 1100 based on biometric identification is provided, comprising: an identification module 1101, a risk assessment module 1102, a risk identification module 1103 and a determination module 1104, wherein:

[0221] The identification module 1101 is used to identify the target part data collected from the target object, obtain a biometric identification result, and determine a target account that matches the target object based on the biometric identification result.

[0222] The risk assessment module 1102 is used to obtain operation data associated with the target account, and perform a multi-dimensional risk assessment on the target account according to the operation data to obtain a risk level of the target account.

[0223] The risk identification module 1103 is used to obtain transaction data associated with the target account, and perform multi-dimensional risk identification on the transaction based on the transaction data to obtain a transaction risk result of the transaction.

[0224] The determination module 1104 is used to determine the target risk control method of the transaction based on the account risk level and the transaction risk result, and perform risk management and control on the transaction according to the target risk control method.

[0225] In some embodiments, the target part data includes a palmprint image, and the recognition module is further used to obtain a target palmprint image collected from the target object, wherein the target palmprint image includes at least one of a color palmprint image, a depth palmprint image, or an infrared palmprint image; based on a preset palmprint image library, a preset palmprint image matching the target palmprint image is searched to obtain a matching result, and a biometric print recognition result is determined based on the matching result.

[0226] In some embodiments, the identification module is also used to obtain at least one candidate account that matches the preset biometric feature when the biometric identification result indicates the existence of at least one preset biometric feature that matches the target part data; obtain verification information associated with the target object, and based on the verification information, determine a target account that matches the target object from the at least one candidate account.

[0227] In some embodiments, the identification module is further used to obtain account feature information associated with each candidate account; based on the matching between the verification information and each account feature information, the matching candidate account is used as a target account that matches the target object.

[0228] In some embodiments, the verification information includes current device-related information uploaded by a collection device that collects target part data, and the account feature information includes at least historical device-related information; the identification module is also used to determine the target device corresponding to each candidate account based on the historical device-related information corresponding to each candidate account; based on the current device-related information, determine whether the collection device matches the target device; and use the candidate account corresponding to the target device that matches the collection device as the target account that matches the target object.

[0229] In some embodiments, the verification information includes cross-verification information obtained in response to a cross-verification operation; the cross-verification information includes at least a current biometric feature that is distinguishable from the target part data; the account feature information includes at least a historical biometric feature; the identification module is also used to match the current biometric feature with the historical biometric features associated with each candidate account; and the candidate account associated with the historical biometric feature that matches the current biometric feature is used as the target account that matches the target object.

[0230] In some embodiments, the risk assessment module is also used to obtain historical operation data associated with the target account; based on the historical operation data, the target account is subjected to a multi-dimensional risk assessment at least from the dimensions of account behavior, remote jump, personnel anomaly, and account attack, to obtain an assessment score corresponding to each dimension; based on the assessment score corresponding to each dimension, the risk level of the target account is obtained.

[0231] In some embodiments, the risk assessment module is also used to perform weighted summation of the assessment scores corresponding to each dimension based on the weight coefficients pre-set for each dimension to obtain the risk score of the target account; based on the comparison result of the risk score and the preset risk classification, obtain the risk level of the target account, wherein the preset risk classification is obtained by constructing a risk score distribution through historical assessment scores of multiple accounts.

[0232] In some embodiments, the risk assessment module is further used to obtain the risk level of the target account based on the assessment score exceeding the threshold when the assessment score corresponding to any dimension exceeds the threshold under the corresponding dimension.

[0233] In some embodiments, the risk identification module is also used to obtain historical transaction data and current transaction data associated with the target account; based on the historical transaction data and the current transaction data, perform multi-dimensional risk identification on the transaction at least from the transaction amount dimension, transaction frequency dimension, and transaction relationship dimension, and obtain risk identification results corresponding to each transaction dimension; obtain the risk label associated with the target account; according to the risk label, obtain the risk identification result under the label dimension; based on the risk identification results corresponding to each dimension, obtain the transaction risk result of this transaction.

[0234] In some embodiments, the above-mentioned device also includes a high similarity module, which is used to set a biometric abnormality label for the target account when the biometric recognition result indicates that there is at least one preset biometric feature that matches the target part data, and the similarity between each matching preset biometric feature is less than a threshold. The high similarity module is connected to the recognition module and to the determination module. Accordingly, the determination module is also used to determine that the target risk control method of this transaction at least includes a credit limit update method based on the biometric abnormality label.

[0235] In some embodiments, the determination module is also used to determine that the target risk control method of this transaction is a mild control method when the risk level of the account and the risk degree represented by the risk transaction results belong to the preset control scope, and the mild control method includes at least one of an additional verification method or a quota update method; when the risk level of the account risk level and the risk degree represented by the risk transaction results exceed the preset control scope, determine that the target risk control method of this transaction is a compulsory control method, and the compulsory control method includes at least one of a transaction interception method or an account ban method.

[0236] In some embodiments, the above-mentioned device also includes an update module, which is used to re-evaluate the account risk level or at least one of the risk transaction results of the target account in response to a complaint request initiated for the risk management of this transaction; based on the re-evaluation result, update the account risk level or at least one of the risk transaction results of the target account, and redetermine the target risk control method of this transaction based on the updated data.

[0237] Each module in the above-mentioned transaction risk control device based on biometrics can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the corresponding operations of each of the above modules.

[0238] In one embodiment, a computer device is provided, which may be a server or a collection device. The following description takes the computer device as a server as an example, and its internal structure diagram may be as follows: Fig.12 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store biometric features, operation data, and transaction data, etc. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a transaction risk control method based on biometrics is implemented.

[0239] Those skilled in the art will understand that Fig.12 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0240] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0241] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0242] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0243] It should be noted that the objects / object information (including but not limited to device information, transaction information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the objects / objects or fully authorized by all parties, and the collection, use and processing of relevant data need to comply with relevant laws, regulations and standards of relevant countries and regions.

[0244] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0245] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0246] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A transaction risk control method based on biometrics, characterized in that: The method comprises: Acquire target part data collected from a target object; the target part data includes a target palm print image; searching for a preset palmprint image matching the target palmprint image based on a preset palmprint image library; In the case where there are multiple similar palm prints matching the target palm print image, respectively obtaining candidate accounts matching each of the similar palm prints, and obtaining verification information associated with the target object; Based on the verification information, determining a target account that matches the target object from among the candidate accounts; Acquire operation data associated with the target account, and perform a multi-dimensional risk assessment on the target account according to the operation data to obtain a risk level of the target account; Acquire transaction data associated with the target account, and perform multi-dimensional risk identification on the transaction based on the transaction data to obtain a transaction risk result of the transaction; Based on the account risk level and the transaction risk result, the target risk control method of this transaction is determined, and the risk of this transaction is controlled according to the target risk control method; wherein, when there are multiple similar palm prints matching the target palm print image and the similarity between each of the similar palm prints is less than a threshold, the target risk control method at least includes a credit limit update method.

2. The method according to claim 1, characterized in that The target palmprint image includes at least one of a color palmprint image, a depth palmprint image, or an infrared palmprint image.

3. The method according to claim 1, characterized in that The step of determining a target account matching the target object from among the candidate accounts based on the verification information includes: Obtain account feature information associated with each candidate account; Based on the matching between the verification information and the characteristic information of each account, the matching candidate account is used as the target account that matches the target object.

4. The method according to claim 3, characterized in that The verification information includes current device related information uploaded by the acquisition device that acquires the target part data, and the account characteristic information includes at least historical device related information; The matching of the verification information with the characteristic information of each account number, taking the matching candidate account number as the target account number matching the target object, includes: According to the historical device related information corresponding to each candidate account, determine the target device corresponding to each candidate account; Based on the current device related information, determine whether the acquisition device matches the target device; The candidate account corresponding to the target device that matches the acquisition device is used as the target account that matches the target object.

5. The method according to claim 3, characterized in that: The verification information includes cross-verification information obtained in response to a cross-verification operation; the cross-verification information includes at least a current biometric feature that is distinguishable from the target part data; the account feature information includes at least a historical biometric feature; The matching of the verification information with the characteristic information of each account number, taking the matching candidate account number as the target account number matching the target object, includes: Matching the current biometric feature with the historical biometric features associated with each candidate account; A candidate account number associated with the historical biometric feature that matches the current biometric feature is used as a target account number that matches the target object.

6. The method according to claim 1, characterized in that The acquiring operation data associated with the target account, and performing a multi-dimensional risk assessment on the target account according to the operation data to obtain a risk level of the target account, includes: Obtaining historical operation data associated with the target account; Based on the historical operation data, a multi-dimensional risk assessment is performed on the target account from at least the account behavior dimension, the remote jump dimension, the personnel abnormality dimension, and the account attack dimension to obtain an assessment score corresponding to each dimension; Based on the evaluation scores corresponding to each dimension, the risk level of the target account is obtained.

7. The method according to claim 6, characterized in that The risk level of the target account is obtained based on the evaluation scores corresponding to each dimension, including: Based on the weight coefficients pre-set for each dimension, weighted summation is performed on the evaluation scores corresponding to each dimension to obtain the risk score of the target account; Based on the comparison result of the risk score and the preset risk classification, the risk level of the target account is obtained, wherein the preset risk classification is obtained by constructing a risk score distribution through historical evaluation scores of multiple accounts.

8. The method according to claim 6, characterized in that The risk level of the target account is obtained based on the evaluation scores corresponding to each dimension, including: When the evaluation score corresponding to any dimension exceeds the threshold value under the corresponding dimension, the risk level of the target account is obtained based on the evaluation score exceeding the threshold value.

9. The method according to claim 1, characterized in that: The acquiring of the transaction data associated with the target account, and performing multi-dimensional risk identification on the transaction according to the transaction data to obtain the transaction risk result of the transaction, includes: Obtain historical transaction data and current transaction data associated with the target account; Based on the historical transaction data and the current transaction data, a multi-dimensional risk assessment is performed on the current transaction at least from the transaction amount dimension, the transaction frequency dimension, and the transaction relationship dimension to obtain a risk assessment result corresponding to each transaction dimension; Obtaining a risk tag associated with the target account; According to the risk label, a risk identification result under the label dimension is obtained; Based on the risk identification results corresponding to each dimension, the transaction risk result of this transaction is obtained.

10. The method according to claim 1, characterized in that In the case where the biometric recognition result indicates that there is at least one similar palmprint matching the target palmprint image, and the similarity between the matching similar palmprints is less than a threshold, the method further includes: Setting a biometric anomaly tag for the target account; Determining the target risk control method of the transaction based on the account risk level and the transaction risk result also includes: Based on the biometric anomaly tag, determine that the target risk control method for this transaction at least includes a credit limit update method.

11. The method according to claim 1, characterized in that: Determining the target risk control method for this transaction based on the account risk level and the transaction risk result includes: When the risk level of the account and the risk degree represented by the risk transaction result fall within the preset control scope, the target risk control method of this transaction is determined to be a mitigation control method, and the mitigation control method includes at least one of an additional verification method or a quota update method; When the risk level of the account and the risk transaction result represent a risk beyond the preset control scope, the target risk control method for this transaction is determined to be a compulsory control method, and the compulsory control method includes at least one of a transaction interception method and an account ban method.

12. The method according to any one of claims 1 to 11, characterized in that The method further comprises: In response to an appeal request initiated by the risk management and control for the transaction, re-evaluate at least one of the account risk level or the risk transaction result of the target account; Based on the re-evaluation result, at least one of the account risk level or the risk transaction result of the target account is updated, and the target risk control method of the transaction is re-determined based on the updated data.

13. A transaction risk control device based on biometrics, characterized in that: The device comprises: The identification module is used to obtain target part data collected from the target object; the target part data includes a target palm print image; based on a preset palm print image library, a preset palm print image matching the target palm print image is searched; in the case where there are multiple similar palm prints matching the target palm print image, candidate accounts matching each of the similar palm prints are respectively obtained, and verification information associated with the target object is obtained; based on the verification information, a target account matching the target object is determined from each of the candidate accounts; A risk assessment module, used to obtain operation data associated with the target account, and perform a multi-dimensional risk assessment on the target account according to the operation data to obtain a risk level of the target account; A risk identification module is used to obtain transaction data associated with the target account, and to perform multi-dimensional risk identification on the transaction based on the transaction data to obtain a transaction risk result of the transaction; A determination module is used to determine a target risk control method for this transaction based on the account risk level and the transaction risk result, and to perform risk management on this transaction according to the target risk control method; wherein, when there are multiple similar palm prints matching the target palm print image and the similarity between each of the similar palm prints is less than a threshold, the target risk control method at least includes a credit limit update method.

14. The device according to claim 13, characterized in that The target palmprint image includes at least one of a color palmprint image, a depth palmprint image, or an infrared palmprint image.

15. The device according to claim 13, characterized in that The identification module is also used to obtain account feature information associated with each candidate account; based on the matching between the verification information and each account feature information, the matching candidate account is used as a target account that matches the target object.

16. The device according to claim 15, characterized in that The verification information includes current device related information uploaded by the acquisition device that acquires the target part data, and the account characteristic information includes at least historical device related information; The identification module is also used to determine the target device corresponding to each candidate account according to the historical device related information corresponding to each candidate account; Based on the current device related information, determine whether the acquisition device matches the target device; and use the candidate account corresponding to the target device that matches the acquisition device as the target account that matches the target object.

17. The device according to claim 15, characterized in that The verification information includes cross-verification information obtained in response to a cross-verification operation; the cross-verification information includes at least a current biometric feature that is distinguishable from the target part data; the account feature information includes at least a historical biometric feature; The identification module is also used to match the current biometric feature with the historical biometric features associated with each candidate account respectively; and use the candidate account associated with the historical biometric feature that matches the current biometric feature as the target account that matches the target object.

18. The device according to claim 13, characterized in that The risk assessment module is also used to obtain historical operation data associated with the target account; based on the historical operation data, the target account is subjected to a multi-dimensional risk assessment at least from the dimensions of account behavior, remote jump, personnel anomaly, and account attack, to obtain an assessment score corresponding to each dimension; based on the assessment score corresponding to each dimension, the risk level of the target account is obtained.

19. The device according to claim 18, characterized in that The risk assessment module is further used to perform weighted summation of the assessment scores corresponding to each dimension based on the weight coefficients pre-set for each dimension to obtain the risk score of the target account; Based on the comparison result of the risk score and the preset risk classification, the risk level of the target account is obtained, wherein the preset risk classification is obtained by constructing a risk score distribution through historical evaluation scores of multiple accounts.

20. The device according to claim 18, characterized in that The risk assessment module is also used to obtain the risk level of the target account based on the assessment score exceeding the threshold when the assessment score corresponding to any dimension exceeds the threshold under the corresponding dimension.

21. The device according to claim 13, characterized in that The risk identification module is further used to obtain historical transaction data and current transaction data associated with the target account; based on the historical transaction data and the current transaction data, perform multi-dimensional risk identification on the current transaction at least from the dimensions of transaction amount, transaction frequency, and transaction relationship, and obtain risk identification results corresponding to each transaction dimension; Obtaining a risk tag associated with the target account; According to the risk label, a risk identification result under the label dimension is obtained; Based on the risk identification results corresponding to each dimension, the transaction risk result of this transaction is obtained.

22. The device according to claim 13, characterized in that When the biometric result indicates that there is at least one similar palm print matching the target palm print image, and the similarity between each matching similar palm print is less than a threshold, the transaction risk control device based on biometric identification also includes a high similarity module for setting a biometric anomaly label for the target account; based on the biometric anomaly label, determining that the target risk control method for this transaction at least includes a credit limit update method.

23. The device according to claim 13, characterized in that The determination module is also used to determine that the target risk control method of this transaction is a moderate control method when the risk level of the account and the risk degree represented by the risk transaction result belong to the preset control scope, and the moderate control method includes at least one of an additional verification method or a quota update method; when the risk level of the account risk level and the risk degree represented by the risk transaction result exceed the preset control scope, determine that the target risk control method of this transaction is a compulsory control method, and the compulsory control method includes at least one of a transaction interception method or an account ban method.

24. The device according to claim 13, characterized in that The transaction risk control device based on biometrics further includes an updating module for re-evaluating at least one of the account risk level or the risk transaction result of the target account in response to a complaint request initiated by the risk control for the transaction; Based on the re-evaluation result, at least one of the account risk level or the risk transaction result of the target account is updated, and the target risk control method of the transaction is re-determined based on the updated data.

25. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 12 are implemented.

26. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 12 are implemented.

27. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 12 are implemented.

Images