Payment method and apparatus, electronic device, computer-readable storage medium, and computer program product

Abstract

The present application provides a payment method and apparatus, an electronic device, a computer-readable storage medium, and a computer program product. The method comprises: in response to a payment request sent by a terminal, acquiring geographic location information of the current location of the terminal, the payment request comprising biometric identification information of a target object; on the basis of the geographic location information, determining an area feature library of an area where the terminal is currently located; in response to a matching failure of the biometric identification information in the area feature library, matching the biometric identification information by means of a full feature library to obtain identity information of the target object, the full feature library comprising area feature libraries of a plurality of areas; acquiring scenario recognition information of the terminal, and determining a scenario recognition result of the terminal on the basis of the scenario recognition information; and executing payment processing on the basis of the scenario recognition result and the identity information of the target object.

Description

Payment method, apparatus, electronic device, computer-readable storage medium and computer program product

[0001] Cross-reference to Related Applications

[0002] This application is based on and claims priority to Chinese Patent Application No. 202410634272.0, filed on May 20, 2024, the entire contents of which are incorporated herein by reference. TECHNICAL FIELD

[0003] The present application relates to the technical field of computers, and particularly relates to a payment method, apparatus, electronic device, computer-readable storage medium and computer program product. BACKGROUND

[0004] With the development of artificial intelligence technology, mobile payment methods are more and more, such as face recognition payment, two-dimensional code payment and fingerprint recognition payment.

[0005] In the related art, palm payment is a mobile payment method based on biometric technology, which performs identity verification and payment authorization by scanning the texture and pattern of the palm. Compared with face recognition payment, two-dimensional code payment and fingerprint recognition payment, palm payment has the advantages of reducing privacy information leakage and convenient operation.

[0006] However, in the related art, users are prone to misbrushing when performing palm payment operations. During palm payment, incorrect identity information that does not match the user may be identified, and thus incorrect payment is performed based on the incorrect identity information. In particular, when the user performs cross-region payment, the misbrushing rate is higher. Therefore, the payment method in the related art has low accuracy and security. SUMMARY

[0007] The payment method, apparatus, electronic device, computer-readable storage medium and computer program product provided by the embodiments of the present application can improve the accuracy and security of payment.

[0008] The technical solutions of the embodiments of the present application are implemented as follows:

[0009] An embodiment of the present application provides a payment method, the method is executed by an electronic device, and the method comprises the following steps: in response to a payment request sent by a terminal, acquiring geographical position information of a current position of the terminal; the payment request comprises biological feature identification information of a target object; based on the geographical position information, determining a regional feature library of a region where the terminal is currently located; in response to a failure of the biological feature identification information in the regional feature library, matching the biological feature identification information through a full-amount feature library to obtain identity information of the target object; the full-amount feature library comprises regional feature libraries of multiple regions; acquiring scene identification information of the terminal, and determining a scene identification result of the terminal based on the scene identification information; and performing payment processing based on the scene identification result and the identity information of the target object.

[0010] An embodiment of the present application provides a payment device, the device comprises: a position acquisition module configured to acquire geographical position information of a current position of a terminal in response to a payment request sent by the terminal; the payment request comprises biological feature identification information of a target object; a feature library acquisition module configured to determine a regional feature library of a region where the terminal is currently located based on the geographical position information; a matching module configured to match the biological feature identification information through a full-amount feature library to obtain identity information of the target object in response to a failure of the biological feature identification information in the regional feature library; the full-amount feature library comprises regional feature libraries of multiple regions; a scene identification module configured to acquire scene identification information of the terminal, and determine a scene identification result of the terminal based on the scene identification information; and a payment module configured to perform payment processing based on the scene identification result and the identity information of the target object.

[0011] An embodiment of the present application provides an electronic device, the electronic device comprises: a memory configured to store computer executable instructions or computer programs; and a processor configured to execute the computer executable instructions or computer programs stored in the memory to implement the payment method provided by the embodiment of the present application.

[0012] An embodiment of the present application provides a computer readable storage medium, which stores computer programs or computer executable instructions, and the computer executable instructions or computer programs are executed by a processor to implement the payment method provided by the embodiment of the present application.

[0013] An embodiment of the present application provides a computer program product, which comprises computer programs or computer executable instructions, and the computer programs or computer executable instructions are executed by a processor to implement the payment method provided by the embodiment of the present application.

[0014] The embodiments of the present application have the following beneficial effects: in the scenario that the target object uses a terminal to perform a payment operation, the biological feature recognition information sent by the terminal can be matched from the area feature library of the area where the terminal is currently located, so as to determine the identity information of the target object. If the matching fails in the area feature library, the biological feature recognition information is matched from the full-amount feature library to determine the identity information of the target object. Compared with directly using the full-amount feature library to match the identity information of the target object in the related art, the embodiments of the present application pre-establish the area feature library for different areas, first match the biological feature recognition information of the target object by using the area feature library of the area where the terminal is currently located, reduce the size of the feature library, and can more accurately identify the identity information of the target object, thereby reducing the probability of false brushing and improving the accuracy and security of payment. Moreover, when the matching fails in the area feature library, that is, the target object is currently performing cross-area payment, the scene recognition result is further determined based on the scene recognition information of the terminal, the payment processing is performed in combination with the identity information matched from the full-amount feature library and the scene recognition result, so as to ensure the security when performing cross-area payment, and thus the accuracy and security of the entire payment process are improved. BRIEF DESCRIPTION OF DRAWINGS

[0015] FIG. 1 is a structural schematic diagram of a payment system architecture provided by an embodiment of the present application;

[0016] FIG. 2 is a structural schematic diagram of an electronic device provided by an embodiment of the present application;

[0017] FIG. 3 is a first flow schematic diagram of a payment method provided by an embodiment of the present application;

[0018] FIG. 4 is a second flow schematic diagram of a payment method provided by an embodiment of the present application;

[0019] FIG. 5 is a third flow schematic diagram of a payment method provided by an embodiment of the present application;

[0020] FIG. 6 is a fourth flow schematic diagram of a payment method provided by an embodiment of the present application;

[0021] FIG. 7 is another optional flow schematic diagram of a payment method provided by an embodiment of the present application;

[0022] FIG. 8a is a schematic diagram of a palm brushing payment device provided by an embodiment of the present application;

[0023] FIG. 8b is a schematic diagram of another palm brushing payment device provided by an embodiment of the present application;

[0024] FIG. 9 is a flow schematic diagram of palm brushing payment provided by an embodiment of the present application;

[0025] FIG. 10 is an interface schematic diagram of mobile phone number verification provided by an embodiment of the present application;

[0026] FIG. 11 is a schematic diagram of a convolutional neural network structure according to an embodiment of the present application.

[0027] It should be noted that the above-mentioned "first", "second", "third", "fourth" are only used to distinguish different schemes, and do not represent the degree of superiority or priority in the implementation process. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical scheme and advantages of the present application more clear, the present application will be further described in detail below with reference to the drawings, and the described embodiments should not be regarded as limiting the present application, and all other embodiments obtained by those skilled in the art without creative labor shall fall within the scope of protection of the present application.

[0029] In the following description, "some embodiments" are referred to, which describe a subset of all possible embodiments, but it can be understood that "some embodiments" can be the same subset or different subset of all possible embodiments, and can be combined with each other without conflict.

[0030] If the similar description of "first / second" appears in the application file, the following description is added, in the following description, the terms "first\second\third" are only used to distinguish similar objects, and do not represent a specific order of the objects. It can be understood that "first\second\third" can be interchanged in a specific order or sequence as allowed, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein.

[0031] In the embodiments of the present application, the term "module" or "unit" refers to a computer program or a part of a computer program with a predetermined function, and works together with other related parts to achieve a predetermined target, and can be implemented entirely or partially by using software, hardware (such as processing circuit or memory) or combination thereof. Similarly, one processor (or multiple processors or memory) can be used to implement one or more modules or units. In addition, each module or unit can be a part of an overall module or unit that includes the functions of the module or unit.

[0032] Unless otherwise defined, all technical and scientific terms used in the embodiments of the present application have the same meanings as those commonly understood by those skilled in the art. The terms used in the embodiments of the present application are only for the purpose of describing the embodiments of the present application and are not intended to limit the present application.

[0033] In the embodiments of the present application, the relevant data collection process should strictly comply with the requirements of relevant national laws and regulations, obtain the informed consent or separate consent of the personal information subject, and within the scope of authorization of laws and regulations and the personal information subject, carry out subsequent data use and processing.

[0034] Before the embodiments of the present application are further described in detail, first, the terms and terminologies involved in the embodiments of the present application are explained, and the terms and terminologies involved in the embodiments of the present application are applicable to the following explanations.

[0035] 1) Palm payment: a new biometric payment method that does not require the use of traditional payment methods (such as cash or bank cards, etc.), based on biometric technology, through scanning the palm texture and pattern of the object to perform identity verification, payment authorization and payment transactions, etc., that is, the object only needs to lightly scan the palm on the terminal device supporting palm payment to complete the payment. Palm payment mainly relies on palmprint recognition technology and palm vein recognition technology. Palmprint recognition generates a unique digital identifier by scanning the unique texture and shape of the user's palm, while palm vein recognition achieves identity authentication by collecting palm vein images and analyzing the unique vein distribution characteristics of the palm vein images. In the payment process, the palmprint and palm vein data of the object are encrypted and then transmitted to the payment system through a secure channel for verification.

[0036] 2) Biometric information: refers to the use of physiological characteristics or behavioral characteristics inherent to living organisms to identify individual identities. These physiological characteristics or behavioral characteristics are unique to each individual and difficult to replicate, so they can be used as reliable identity verification means. Biometric information includes biometric data used to identify and verify individual identities, such as fingerprints, faces, irises, voiceprints, gaits, deoxyribonucleic acid (DNA), palm veins, and palmprints. Biometric information can be collected and recorded by sensors, cameras or other biometric collection devices, and is commonly used in identity verification, security access control and payment verification to ensure accuracy and security.

[0037] 3) Full feature library: refers to a database or storage system containing all available feature data. In the process of biometric identification, the full feature library contains the feature data of all registered users, and the full feature library is used for matching and identifying biometric information.

[0038] 4) Regional feature library: refers to a database or feature data storage area divided according to geographical location or other specific conditions, different geographical locations correspond to respective regional feature libraries, and the regional feature library contains the feature data of all registered users in the geographical location corresponding to the regional feature library. For example, in biometric identification, the regional feature library of region A includes the feature data of all objects in region A. The full feature library includes the regional feature libraries of all regions, and the regional feature library can be regarded as a sub-feature library in the full feature library. The feature data in all regional feature libraries is aggregated to obtain the feature data in the full feature library.

[0039] 5) Geographical location information: refers to data describing or identifying the location of a point or area on Earth, which can include latitude and longitude coordinates, address, and place name information. Geographical location information can be obtained through Global Positioning System (GPS), wireless network, and base station positioning technology.

[0040] 6) Point of Interest (POI): refers to a geographical location with specific attributes or functions, such as restaurants, stores, scenic spots, gas stations, and highway service areas.

[0041] 7) Payment risk level: refers to the division of payment process into different risk levels according to payment behavior and transaction risk. Different payment risk levels will adopt corresponding risk control strategies, such as different payment verification methods in the embodiments of the present application.

[0042] 8) Risk control strategy: a series of measures for assessing and controlling potential risks, such as the strategy for controlling risks in payment process in the embodiments of the present application.

[0043] 9) Bluetooth Beacon: a wireless signal transmitting device based on Bluetooth Low Energy (BLE) technology, which can broadcast specific information to nearby devices through Bluetooth signals. These information is usually used for positioning, navigation, and push notification, etc., so as to realize indoor positioning and scene recognition functions.

[0044] In related technologies, when a user performs a palm swiping payment operation, it is easy to cause mis-swiping, that is, the identity information that does not match the user may be identified, and thus an error payment is made based on the identity information that does not match the user. Especially when the user performs cross-city payment, the mis-swiping rate is higher. In order to solve the problem of high mis-swiping rate in cross-city, the palm swiping payment usually adopts a risk control strategy based on fixed geographical location, that is, when the user performs a palm swiping payment operation, the system will perform risk control interception according to the geographical location information of the user. If the geographical location information of the user changes, the mobile phone number verification is triggered. This method works well when dealing with users in the same area (region), but in the cross-city scenario, the mobile phone number verification needs to be triggered, which increases the operation difficulty of the user and the user experience is poor. In addition, the palm swiping payment method in related technologies relies only on the geographical location information of the device for risk control interception, such as GPS coordinates. This method can identify the geographical location of the user to some extent, but cannot accurately determine the specific scene where the user is located, such as specific scenes such as gas stations and highway service stations. Therefore, in these specific scenes, the risk control strategy is less effective, which easily leads to mis-swiping and poor user experience.

[0045] In view of at least one of the problems existing in the related art, embodiments of the present application provide a payment method, apparatus, electronic device, computer-readable storage medium, and computer program product. The method is a payment method for out-of-place dynamic Alipay risk control and interception based on geographic location and scene detection, which can improve the accuracy and security of payment. In the payment method provided by the embodiments of the present application, first, in response to a payment request sent by a terminal, geographic location information of a location where the terminal is currently located is obtained. The payment request includes biological feature recognition information of a target object. Then, based on the geographic location information, a regional feature library of a region where the terminal is currently located is determined. When the biological feature recognition information fails to match in the regional feature library, the biological feature recognition information is matched through a full-amount feature library to obtain identity information of the target object. The full-amount feature library includes regional feature libraries of multiple regions. Then, scene recognition information of the terminal is obtained, and a scene recognition result of the terminal is determined based on the scene recognition information. Finally, payment processing is performed based on the scene recognition result and the identity information of the target object.

[0046] The following describes an exemplary application of a payment device provided by the embodiments of the present application, which is an electronic device for implementing the payment method. The payment device (i.e., the electronic device) provided by the embodiments of the present application can be implemented as any terminal with a biological feature recognition information recognition function or a payment function, such as a notebook computer, a tablet computer, a desktop computer, a set-top box, a smart phone, a smart speaker, a smart watch, a smart television, a vehicle-mounted terminal, etc., or as a server. The server can be a standalone physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and basic cloud computing services such as big data and artificial intelligence platforms. The terminal and the server can be directly or indirectly connected through wired or wireless communication, which is not limited in the embodiments of the present application. The following describes an exemplary application when the payment device is implemented as a server.

[0047] In the case that the payment device is implemented as a server, the payment device is included in the payment system. Referring to FIG. 1, which is a schematic diagram of an architecture of a payment system according to an embodiment of the present application, to perform a payment operation, a payment application can be provided, which can be an application dedicated to payment or a functional module in another application (such as a payment module in an instant messaging application). The payment system 100 according to an embodiment of the present application includes at least a terminal 400, a network 300 and a server 200, wherein the server 200 is a server of the payment application. The server 200 can constitute the payment device according to an embodiment of the present application, that is, the payment method according to an embodiment of the present application is implemented through the server 200. The terminal 400 is connected to the server 200 through the network 300, and the network 300 can be a wide area network or a local area network, or a combination of the two.

[0048] In an embodiment of the present application, the terminal 400 can include an image acquisition device, which is used to acquire biometric information of a user. Referring to FIG. 1, a user (i.e., a target object) can perform an interactive operation on a client of the payment application through the terminal 400, which can be, for example, a click payment operation. After receiving the interactive operation of the target object, the terminal 400 acquires biometric information of the target object, encapsulates the biometric information into a payment request, and sends the payment request to the server 200 through the network 300. After receiving the payment request, the server 200 acquires geographic location information of a current location of the terminal in response to the payment request sent by the terminal, wherein the payment request includes the biometric information of the target object. The server 200 determines a regional feature library of a region where the terminal is currently located based on the geographic location information, and matches the biometric information in the regional feature library. When the biometric information fails to be matched in the regional feature library, the server 200 matches the biometric information through a full-amount feature library to obtain identity information of the target object, wherein the full-amount feature library includes regional feature libraries of multiple regions. Meanwhile, the server 200 also acquires scene recognition information of the terminal, determines a scene recognition result of the current location of the terminal based on the scene recognition information, and finally performs payment processing based on the scene recognition result and the identity information of the target object to obtain a payment result. The server 200 can also send the payment result to the terminal 400 to show the payment result to the user.

[0049] In some embodiments, the steps in the payment method can also be performed by the terminal 400, that is, the terminal 400 acquires the geographic location information of the current location; then, the terminal 400 determines the regional feature library of the region where the terminal is currently located based on the geographic location information; in response to the biological feature recognition information of the target object failing to match in the regional feature library, the biological feature recognition information is matched through the full-amount feature library to obtain the identity information of the target object; the full-amount feature library includes regional feature libraries of multiple regions; then, the terminal 400 acquires the scene recognition information of the terminal and determines the scene recognition result of the terminal based on the scene recognition information; finally, the terminal 400 performs payment processing based on the scene recognition result and the identity information of the target object.

[0050] The payment method provided by the embodiments of the present application can also be implemented based on a cloud platform and through cloud technology, for example, the server 200 described above can be a cloud server. The cloud server acquires the geographic location information of the current location of the terminal in response to the payment request sent by the terminal; and the cloud server determines the regional feature library of the region where the terminal is currently located based on the geographic location information; and the cloud server matches the biological feature recognition information through the full-amount feature library when the biological feature recognition information fails to match in the regional feature library to obtain the identity information of the target object; the full-amount feature library includes regional feature libraries of multiple regions; and the cloud server acquires the scene recognition information of the terminal and determines the scene recognition result of the terminal based on the scene recognition information; or the cloud server performs payment operation based on the scene recognition result and the identity information of the target object.

[0051] In some embodiments, there can also be a cloud storage, and the regional feature library and the full-amount feature library can be stored in the cloud storage. In this way, when the payment request is received, the regional feature library and the full-amount feature library can be directly acquired from the cloud storage, so as to realize the identification and matching of the biological feature recognition information of the target object based on the regional feature library and the full-amount feature library, thereby improving the payment response rate.

[0052] It needs to be explained here that cloud technology refers to the unified hardware, software and network resources in a wide area network or local area network to realize the calculation, storage, processing and sharing of data. Cloud technology is based on the network technology, information technology, integration technology, management platform technology and application technology of cloud computing business model application, which can form a resource pool, and can be used as needed, and is flexible and convenient. Cloud computing technology will become an important support. The background service of the technical network system needs a large amount of computing resources and storage resources, such as video websites, picture websites and more portal websites. With the high development and application of the Internet industry, every item may have its own identification mark in the future, and it needs to be transmitted to the background system for logical processing. Different levels of data will be processed separately, and various industry data need strong system support. Various industry data can be realized through cloud computing.

[0053] Referring to FIG. 2, FIG. 2 is a structural schematic diagram of an electronic device provided by an embodiment of the present application. The electronic device shown in FIG. 2 can be a payment device, and the payment device includes at least one processor 410, a memory 450, at least one network interface 420 and a user interface 430. The various components in the terminal 400 are coupled together through a bus system 440. It can be understood that the bus system 440 is used to realize the connection and communication between the components. In addition to including a data bus, the bus system 440 also includes a power bus, a control bus and a status signal bus. However, for the purpose of clear illustration, all kinds of buses are marked as the bus system 440 in FIG. 2.

[0054] The processor 410 can be an integrated circuit chip with signal processing capability, such as a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc., wherein the general-purpose processor can be a microprocessor or any conventional processor.

[0055] The user interface 430 includes one or more output devices 431 that enable presentation of media content, including one or more speakers and one or more visual display screens. The user interface 430 also includes one or more input devices 432 that facilitate user input, such as a keyboard, a mouse, a microphone, a touch screen display, a camera, other input buttons and controls.

[0056] The memory 450 can be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid-state memory, hard drives, optical drives, and the like. The memory 450 optionally includes one or more storage devices remotely located from the processor(s) 410. The memory 450 comprises a volatile memory or non-volatile memory, and can also include both volatile and non-volatile memory. Non-volatile memory can be read only memory (ROM), volatile memory can be random access memory (RAM). The memory 450 described in embodiments of the present application is intended to include any suitable type of memory. In some embodiments, the memory 450 is capable of storing data to support various operations, examples of which include programs, modules, and data structures or a subset or superset thereof, which are described below.

[0057] The operating system 451 includes system programs for processing various basic system services and performing hardware-related tasks, such as a framework layer, a core library layer, a driver layer, and the like, for implementing various basic services and processing hardware-based tasks; a network communication module 452 for reaching other electronic devices via one or more (wired or wireless) network interfaces 420, example network interfaces 420 including Bluetooth, wireless fidelity (WiFi), universal serial bus (USB), and the like; a presentation module 453 for enabling the presentation of information via one or more output devices 431 associated with the user interface 430 (e.g., a display screen, a speaker, and the like) (e.g., a user interface for operating peripheral devices and displaying content and information); and an input processing module 454 for detecting and translating one or more user inputs or interactions from one or more input devices 432.

[0058] In some embodiments, the payment device provided by embodiments of the present application can be implemented in a software manner, and FIG. 2 shows a payment device 455 stored in the memory 450, which can be software in the form of programs and plug-ins, including the following software modules: a location acquisition module 4551, a feature library acquisition module 4552, a matching module 4553, a scene recognition module 4554, and a payment module 4555. These modules are logical, and thus can be combined or further split according to the implemented functions. The functions of each module will be described below.

[0059] In some embodiments, the payment device provided by the embodiments of the present application can be implemented in a hardware manner. For example, the device provided by the embodiments of the present application can be a hardware decoding processor programmed to perform the payment method provided by the embodiments of the present application. For example, the hardware decoding processor can be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), or other electronic elements.

[0060] It should be noted that the examples of the payment method described below are based on the payment scenario of swiping palms. Based on the understanding of the following description, those skilled in the art can apply the payment method provided by the embodiments of the present application to other payment scenarios based on biological recognition technology, such as face recognition payment, fingerprint recognition payment, iris payment, and voiceprint payment.

[0061] FIG. 3 is a first flowchart of the payment method provided by the embodiments of the present application. The following will be described in combination with the steps shown in FIG. 3. As shown in FIG. 3, the payment method is described by taking a server as an example of an execution subject of the payment method. The method includes the following steps S101-S105:

[0062] In step S101, in response to a payment request sent by a terminal, geographic location information of a current location of the terminal is obtained.

[0063] Here, the payment request can include biological feature recognition information of a target object. The payment request is a request sent by a target object to a merchant or a service provider through a terminal, and is used to request payment for goods or services. The target object can be a user who performs a payment operation through the terminal. The biological feature recognition information of the target object is biological feature data used to identify and verify the identity information of the target object, such as fingerprints, faces, irises, voiceprints, or palm prints.

[0064] In some embodiments, in addition to the biological feature recognition information, the payment request can also include geographic location information of a current location of the terminal. The terminal carries the geographic location information in the payment request and sends it to the server at the same time when sending the payment request to the server.

[0065] The geographic location information of the current location of the terminal can include at least one of the following: longitude and latitude coordinates of the current location of the terminal (the longitude and latitude coordinates are used to accurately describe the location of the terminal on the surface of the earth), height information (also referred to as altitude information, the height information is used to provide the three-dimensional position of the terminal), cell identification (Cell ID) (in a cell coverage-based positioning method, the location of the terminal can be estimated by the cell identification of the cell to which the terminal belongs, the cell identification is related to the geographic location of the base station, and the location of the terminal can be roughly determined by the location information of the base station), angle of arrival (AoA) (by measuring the angle of arrival of the terminal signal at the base station, in combination with the location information of the base station, the relative position of the terminal can be calculated), pseudo-range (in a satellite positioning system, the terminal calculates the pseudo-range from each satellite by receiving the signals of the satellites, and then calculates the accurate position of the terminal), indoor positioning information (in an indoor environment, the location of the terminal can be determined by a nearby room-dividing antenna or an indoor distribution system, and by calculating the wireless propagation distance between the terminal and the room-dividing antenna, in combination with the location information of the antenna, the specific location of the terminal in the indoor environment can be determined), and fingerprint database data (the fingerprint database stores signal characteristic data such as signal strength and signal delay at different locations. By matching the currently measured signal characteristics of the terminal with the data in the fingerprint database, the location of the terminal can be determined), and the like.

[0066] In some other embodiments, if the geographic location information is not carried in the payment request, or the server fails to successfully parse the geographic location information from the payment request, the server can further send a geographic location acquisition request to the terminal in response to the payment request, so that the terminal acquires and sends the geographic location information to the server. In the embodiments of the present application, the method for acquiring the geographic location information is not limited, and for example, the geographic location information can be acquired by means of GPS, wireless network, base station positioning, and the like.

[0067] For example, in a palm payment scenario, the terminal can be a palm payment device installed with a payment application, and the biometric information can be an image of the palm of the target object. The target object can click to purchase a product or the like through the client of the payment application on the terminal, so as to enter a payment interface. The terminal acquires the image of the palm of the target object through an image acquisition device (for example, a camera), encapsulates the image of the palm into a payment request, and sends the payment request to the server. The palm payment device can further be internally provided with an application with GPS function, for example, a map application. The palm payment device acquires the geographic location information in real time through the GPS function of the map application, encapsulates the geographic location information into the payment request, and sends the payment request to the server.

[0068] In step S102, the area feature library of the area where the terminal is currently located is determined based on the geographic location information.

[0069] Here, the geographic location information can include longitude and latitude coordinates of the terminal, and the region where the terminal is currently located can be determined through the longitude and latitude coordinates in the geographic location information. The region is a pre-divided geographical range. When dividing the region, at least the following division principles need to be considered: the principle of combining zonality and non-zonality, the principle of main factors and comprehensive analysis, the principle of relative consistency, the principle of genesis, the principle of regional conjugation, and the principle of applicability. Among them, zonality refers to the characteristics of natural geographical elements regularly replacing each other in a certain direction on the earth's surface, such as latitude zonality, longitude zonality, and vertical zonality. Non-zonality refers to geographical phenomena caused by internal energy of the earth or other local factors. The principle of main factors and comprehensive analysis is that when dividing the region, not only the dominant factors (such as climate and terrain, etc.) need to be considered, but also a variety of geographical elements need to be comprehensively analyzed to ensure the scientificity and rationality of the division result. The principle of relative consistency is that the similarity of geographical elements within the region should be as large as possible, and the difference between different regions should be as large as possible. The principle of genesis is that the divided region should have consistent generation, that is, the geographical phenomena within the region have similarity in the formation and development process. The principle of regional conjugation is that each specific division unit must be a continuous and unique regional unit. The principle of applicability is that the division of the region should have a clear service direction, such as serving identity recognition, etc.

[0070] In the embodiments of the present application, the region can be divided by any one of the following division methods: natural geographical region division, administrative region division, economic region division, and humanistic and cultural region division, etc. Among them, the basis of natural geographical region division is natural geographical characteristics. The basis of administrative region division is administrative division boundary. The basis of economic region division is economic development level, industrial structure, and resource distribution, etc. The basis of humanistic and cultural region division is the regional differentiation and similarity of humanistic elements. It should be noted that the division rules of the region in the embodiments of the present application are not limited, and can be set according to actual needs, for example, division is performed with cities as the dimension.

[0071] The region feature library includes the biometric feature data of all objects in the region. For example, the region feature library of region A includes the palmprint feature data of all objects in city A. When the object in the region registers a payment account, the object can be prompted to collect a biometric feature image, and the biometric feature data obtained by feature extraction on the biometric feature image is stored in the region feature library of the region.

[0072] In step S103, in response to the failure of the biometric feature recognition information matching in the region feature library, the biometric feature recognition information is matched through the full feature library to obtain the identity information of the target object.

[0073] Here, the biometric feature recognition information can include a biometric feature recognition image. The biometric feature recognition image is an image of a certain part of a target object, such as a face image or a palm image, etc. The region feature library includes image features of multiple objects. The image feature is biometric feature data obtained by performing feature extraction on an uploaded biometric feature recognition image when an object registers a payment account. The identity information of the target object is data or information used to identify the identity of the target object, including but not limited to name, birthday, address, phone number, email address, and ID number, etc.

[0074] In the embodiments of the present application, after obtaining the biometric feature recognition information of the target object, the biometric feature recognition information can be matched in the region feature library of the region where the terminal is located. When matching the biometric feature recognition information, the biometric feature recognition image can be first feature-extracted to obtain the image feature to be identified. Then, the similarity between the image feature to be identified and the image feature of each object in the region feature library is calculated. When there is at least one similarity greater than or equal to a preset similarity threshold, the matching is successful, and the identity information of the object corresponding to the image feature with the greatest similarity and greater than or equal to the preset similarity threshold is determined as the identity information of the target object. Then, the payment account corresponding to the identity information of the target object can be obtained, and the payment request corresponding to the deduction operation is executed in the payment account to generate a payment result. It should be noted that the method of feature-extracting the biometric feature recognition image in the embodiments of the present application is not limited, for example, the feature extraction can be performed through a convolution network. The method of calculating the similarity between the image feature to be identified and the image feature is also not limited in the embodiments of the present application, for example, it can be a cosine similarity calculation method, the cosine distance between the image feature to be identified and the image feature is calculated, and the cosine distance is taken as the cosine similarity between the image feature to be identified and the image feature. The preset similarity threshold can be set based on actual needs, and the greater the similarity threshold, the higher the accuracy of the identity information obtained.

[0075] When the similarity between the image feature of each object and the image feature to be identified is less than the preset similarity threshold, the matching fails, indicating that the image feature of the target object does not exist in the region feature library. That is, the image feature of the target object does not belong to the region feature library of the current region, and the target object is performing a cross-region payment operation. At this time, the biometric feature recognition information can be matched through the full-amount feature library to obtain the identity information of the target object. The full-amount feature library includes the region feature libraries of all divided regions, that is, the full-amount feature library includes the biometric feature data of all objects that have registered payment accounts.

[0076] In some embodiments, if the biometric information also fails to match through the full-quantity feature library, the terminal can be sent prompt information that the object has not opened the payment function, and the prompt information can be displayed on the current interface of the terminal to prompt the target object to open the payment function.

[0077] In some embodiments, the biometric information includes a biometric image. Referring to FIG. 4, the biometric information is matched through the full-quantity feature library in step S103 to obtain the identity information of the target object, which can be implemented through the following steps S1031 to S1033:

[0078] In step S1031, the biometric image is subjected to feature extraction to obtain image features to be recognized.

[0079] Here, the biometric image is an image of a part of the target object, such as a face image or a palm image, which is collected by the terminal through an image collection device. The image features to be recognized include specific feature information extracted from the biometric image, including but not limited to pixel values, edge information, and texture features, etc. The image features to be recognized are represented in the form of a vector. The method of feature extraction is not limited in the embodiments of the present application. For example, the method of feature extraction can be implemented based on deep learning, and specifically, a convolutional neural network can be used for feature extraction.

[0080] In step S1032, the similarity between the image features to be recognized and the image features of each object in the full-quantity feature library is determined.

[0081] Here, the image features of each object in the full-quantity feature library are feature vectors obtained by feature extraction on the biometric image uploaded during registration of a payment account. For the image features of each object, the similarity between the image features and the image features to be recognized can be determined. Here, several calculation methods for calculating the similarity between the image features and the image features to be recognized are listed: distance-based similarity calculation and vector angle-based similarity calculation, etc.

[0082] In the distance-based similarity calculation method, the Euclidean distance between the image features and the image features to be recognized can be calculated through the following formula (1):

[0083] wherein A represents the feature vector of the image features, A i represents the i-th element in the feature vector A; B represents the feature vector of the image features to be recognized, B i represents the i-th element in the feature vector B; and n represents the total number of elements in the feature vector A (or the total number of elements in the feature vector B).

[0084] Alternatively, the Manhattan Distance between the image feature and the image feature to be identified can also be calculated by the following formula (2):

[0085] Alternatively, the Mean Squared Error (MSE) between the image feature and the image feature to be identified can also be calculated by the following formula (3):

[0086] Wherein, I1 and I2 are pixel values of two images respectively, and m*n represents the number of pixels of the image.

[0087] In the similarity calculation based on the vector angle, the Cosine Similarity between the image feature and the image feature to be identified can be calculated, and the Cosine Similarity measures the similarity by calculating the cosine value of the angle between two feature vectors, as shown in the following formula (4):

[0088] The embodiments of the present application do not limit the selection of the calculation method for calculating the similarity between the image feature to be identified and the image feature, for example, the Cosine Similarity calculation method can be selected to calculate the similarity between the image feature to be identified and the image feature.

[0089] In step S1033, the identity information of the object corresponding to the image feature with the maximum similarity is determined as the identity information of the target object.

[0090] Here, the similarity between the image feature to be identified and the image feature of each object in the full feature library can be sorted in ascending order or descending order to obtain a similarity sequence, and then the maximum similarity is determined from the similarity sequence. The identity information of the object corresponding to the image feature with the maximum similarity can be directly determined as the identity information of the target object. Alternatively, in another embodiment, it can also be judged whether the maximum similarity is greater than or equal to a preset similarity threshold. When the maximum similarity is greater than or equal to the similarity threshold, the identity information of the object corresponding to the image feature with the maximum similarity is determined as the identity information of the target object. When the maximum similarity is less than the similarity threshold, it is determined that the matching fails, and the terminal can be sent a prompt information that the object does not open the payment function, and the prompt information is displayed on the current interface of the terminal to prompt the target object to open the payment function.

[0091] The preset similarity threshold in the embodiments of the present application can be set based on actual needs, wherein the greater the similarity threshold is set, the higher the accuracy of the identity information obtained by identification is.

[0092] In step S104, scene recognition information of the terminal is acquired, and a scene recognition result of the terminal is determined based on the scene recognition information.

[0093] Here, the scene recognition information can include one or more of point of interest information, network information, a wireless connection signal, and a background image. The scene recognition result is used to indicate which type of scene the terminal is currently in, and the scene recognition result can include a first type of scene or a second type of scene. The first type of scene is a scene in which a proportion of payment operations performed on a cross-regional object exceeds a preset proportion threshold, and the second type of scene is any other scene other than the first type of scene. For example, the first type of scene can include a gas station or a highway service area, and the like, in which a proportion of payment operations performed on a cross-regional object exceeds a preset proportion threshold.

[0094] In some embodiments, the scene recognition information can include point of interest information, which is scene information within a preset range of a current location of the terminal. In step S104, the scene recognition information of the terminal is acquired, and the scene recognition result of the terminal is determined based on the scene recognition information, which can be implemented in the following manner: first, a plurality of point of interest information is acquired from the scene recognition information; then, for each point of interest information, a preset first keyword is searched in the point of interest information; and when the first keyword exists in at least one point of interest information, it is determined that the scene recognition result of the terminal is that the terminal is currently located in the first type of scene.

[0095] Here, the preset range can be set by actual needs, for example, the preset range can be a range of 100 meters from a current location of the terminal, that is, a circular area range with the current location of the terminal as the center and 100 meters as the radius. The point of interest information includes a name, a type, and an address of the point of interest, and the like. The point of interest refers to a geographical location with specific attributes or functions. In the embodiments of the present application, the point of interest is used to refer to a scene, such as a restaurant, a store, a scenic spot, a gas station, a highway service area, and the like. The first keyword is a word or a term used to represent a scene name of the first type of scene, such as the keywords "highway service area" and "gas station". The first keyword can be set by actual needs. When the first keyword exists in at least one point of interest information, it can be determined that the scene recognition result of the terminal is that the terminal is currently located in the first type of scene; and when the first keyword does not exist in a plurality of point of interest information, it can be determined that the scene recognition result of the terminal is that the terminal is currently located in the second type of scene.

[0096] For example, the preset range is 100 m, a circular area is obtained by taking the current location of the terminal as the center and the preset range as the radius, and three point-of-interest information in the circular area is obtained. For any point-of-interest information, it can be checked whether the first keyword such as "highway service area" or "gas station" exists in the name and type of the point of interest in the point-of-interest information. If the names of the three points of interest are "XXX restaurant", "XXX supermarket", and "XXX gas station", the first keyword "gas station" exists in "XXX gas station", and therefore, it can be determined that the scene recognition result of the terminal is that the terminal is currently located in the first type of scene. If the names of the three points of interest are "XXX restaurant", "XXX supermarket", and "XXX school", and the first keyword does not exist in the names of the three points of interest, it can be determined that the scene recognition result of the terminal is that the terminal is currently located in the second type of scene.

[0097] The embodiment of the present application determines the type of the scene in which the terminal is currently located by using the point-of-interest information, which is convenient and fast, can improve the speed of scene recognition, and further improve the response speed of the payment request.

[0098] In some embodiments, the scene recognition information can further include network information. The scene recognition information of the terminal is obtained in step S104, and the scene recognition result of the terminal is determined based on the scene recognition information, which can be achieved by the following method: first, obtaining the network information of the network accessed by the terminal from the scene recognition information; then, determining the network name and signal strength of the current access network from the network information; finally, determining the scene recognition result of the terminal based on the network name and signal strength.

[0099] Here, the terminal is a kind of networking device, and can obtain the current access network of the terminal. The current access network of the terminal can be a wireless network (such as WiFi) or a wired network. The network information is the information of the current access network of the terminal, and can include the network name and signal strength. Taking WiFi as an example, the network information can be the name of WiFi and the signal strength of WiFi received by the terminal. The signal strength can be used to represent the distance between the location of the terminal and the location of WiFi. Therefore, the scene recognition result of the terminal can be determined by the network name and signal strength.

[0100] In the embodiment of the present application, the scene recognition result of the terminal is determined based on the network name and signal strength, which can be achieved by the following method: first, searching for a preset second keyword in the network name; then, when the second keyword exists in the network name and the signal strength reaches or is greater than a preset strength threshold, it is determined that the scene recognition result of the terminal is that the terminal is currently located in the first type of scene.

[0101] Here, the second keyword is a word or phrase used to represent the name of the scene of the first type, such as "highway service area", "gas station", etc. The second keyword can be the same as the first keyword. For example, the WiFi name connected by the terminal is "XXX gas station", and the signal strength of the WiFi is usually represented by a signal strength indicator (Signal Strength Indicator). A common representation is to represent the negative value (dBm) of the signal strength, and the larger the value, the stronger the signal. The signal strength of the "XXX gas station" WiFi connected by the terminal is -40 dBm. The preset strength threshold is -60 dBm, and the search result is that the second keyword exists in the network name and the signal strength is greater than the preset strength threshold, so it can be determined that the scene recognition result of the terminal is that the terminal is currently located in the first type of scene.

[0102] The embodiments of the present application determine the specific type of the scene in which the terminal is located through the network name. This determination method is convenient and fast, which can improve the speed of scene recognition, and in turn improve the response speed of the payment request. At the same time, the determination of the signal strength can improve the accuracy of scene recognition, and in turn improve the accuracy and security of payment.

[0103] In some embodiments, the scene recognition information includes a wireless connection signal. The scene recognition information of the terminal is obtained in step S104, and the scene recognition result of the terminal is determined based on the scene recognition information. This can be achieved by the following method: first, obtaining the wireless connection signal received by the terminal from the scene recognition information; then, determining the to-be-identified signal name of the wireless connection signal; finally, when the to-be-identified signal name is the same as the target signal name corresponding to the beacon signal, determining that the scene recognition result of the terminal is that the terminal is currently located in the first type of scene; the beacon signal is a signal transmitted by a wireless connection beacon preset in the first type of scene.

[0104] The wireless connection signal includes a Bluetooth signal or a broadcast signal, and the like. Taking the Bluetooth signal as an example, the wireless connection beacon is a Bluetooth beacon. The Bluetooth beacon is a small wireless device based on Bluetooth low energy technology. The Bluetooth beacon can periodically broadcast specific signals, which can be received and analyzed by other devices (such as smart phones, tablets, and the like), thereby realizing various location-based services and functions. The Bluetooth beacon can periodically broadcast data packets containing unique identifiers through Bluetooth low energy technology. When the terminal enters the signal coverage range of the Bluetooth beacon, the application program on the terminal can analyze the data packets, thereby realizing various location-based services. The beacon signal is a signal sent by the Bluetooth beacon. The Bluetooth beacon sending the beacon signal can be deployed in advance in the first type of scene. If the terminal is in the first type of scene, the terminal will automatically connect to the Bluetooth beacon and receive the beacon signal sent by the Bluetooth beacon. Therefore, when the to-be-identified signal name of the wireless connection signal received by the terminal is the same as the target signal name corresponding to the beacon signal, the scene recognition result can be obtained as the terminal currently being in the first type of scene. For example, the Bluetooth beacon is deployed in a first type of scene “gas station”, and the target signal name corresponding to the beacon signal sent by the Bluetooth beacon is set as “XXX gas station”. When the to-be-identified signal name of the wireless connection signal received by the terminal is “XXX gas station”, it indicates that the terminal is currently in the first type of scene “gas station”.

[0105] The embodiments of the present application determine the type of the scene in which the terminal is located through the wireless connection signal. This determination method is convenient and fast, which can improve the speed of scene recognition and the response speed of the payment request.

[0106] In some embodiments, the scene recognition information includes a background image. The scene recognition information of the terminal is obtained in step S104, and the scene recognition result of the terminal is determined based on the scene recognition information. This can be achieved by the following method: first, obtaining the background image collected in the current scene of the terminal from the scene recognition information; then, determining the probability that the image content in the background image belongs to different preset scenes by using a pre-trained scene detection model; finally, determining the scene recognition result of the terminal based on the probability that the image content in the background image belongs to different preset scenes; and the scene recognition result is that the terminal is located in the preset scene corresponding to the maximum probability.

[0107] Here, the background image is a scene photo collected by the terminal through an image collection device (for example, a camera). For example, the terminal can take a scene photo when it is started for the first time, and upload the background image to the server. Then, the background image can be input into the pre-trained scene detection model to obtain an image classification result. The image classification result includes probabilities that the image content in the background image belongs to different preset scenes. The preset scenes include a first type of scene and a second type of scene. The preset scene with the maximum probability is determined as the current scene where the terminal is located, and a scene recognition result is obtained. For example, after the background image is input into the pre-trained scene detection model, it is obtained that the probability of belonging to the first type of scene is 0.8 and the probability of belonging to the second type of scene is 0.2. The probability of the first type of scene is the maximum probability, so the scene recognition result that the terminal is located in the first type of scene can be obtained. It should be noted that the scene detection model is not limited in the embodiments of the present application, and can be a convolutional neural network model (CNN).

[0108] The scene detection model can be trained by the following method. Sample background images and sample labels are obtained. The sample background image is a photo collected by a sample terminal in a current scene, and the sample label is used to label the current scene where the sample terminal is located. The sample background images captured by multiple sample terminals after starting for the first time can be uploaded to the server, and the server stores each sample background image uploaded by each terminal. At the same time, the sample background images are manually labeled to obtain sample labels indicating whether the sample background images are first type scenes or second type scenes. Then, the convolutional neural network model to be trained is used to predict the scenes of the sample background images to obtain sample scene recognition results of the sample terminals. Based on the sample scene recognition results and the sample labels, a loss result is calculated, and an optimization algorithm such as Stochastic Gradient Descent (SGD) or Adaptive Moment Estimation (Adam) is used to train the convolutional neural network model to be trained until the loss of the convolutional neural network model to be trained converges, and a pre-trained scene detection model is obtained.

[0109] The embodiments of the present application identify the background image taken by the terminal through the scene detection model to obtain a scene recognition result. Since the scene detection model is a model trained based on a large amount of sample data containing sample labels, the scene recognition result predicted by the scene detection model can improve the accuracy of scene recognition, and further improve the accuracy and security of payment.

[0110] In step S105, payment processing is performed based on the scene recognition result and the identity information of the target object.

[0111] Here, when the scene recognition result is that the terminal is located in the first type of scene, the payment operation corresponding to the payment request can be directly performed in the payment account corresponding to the identity information of the target object, and after the deduction is completed, a payment result is generated, thereby realizing the process of payment processing. When the scene recognition result is that the terminal is located in the second type of scene, the target object needs to perform a preset other verification operation, and after the verification succeeds, the payment operation corresponding to the payment request is performed in the payment account corresponding to the identity information of the target object. If the verification fails, the target object is displayed with a payment failure result.

[0112] In some embodiments, referring to FIG. 5, FIG. 5 shows that the payment operation based on the scene recognition result and the identity information of the target object in step S105 can be implemented through the following steps S1051 to S1052:

[0113] Step S1051, adjusting the payment risk control level of the terminal based on the scene recognition result.

[0114] Here, the payment risk control level of the terminal is used to control the payment verification mode of the terminal. The default payment risk control level of the terminal is the first level. The payment verification mode corresponding to the first level is that when the target object performs cross-region payment, the reserved verification information needs to be verified, and the payment operation is continued only after the verification succeeds. If the verification fails, the payment operation is terminated. The cross-region payment of the target object means that the biometric information of the target object fails to match in the regional feature library of the current region of the terminal. When the target object does not perform cross-region payment, the payment operation corresponding to the payment request can be directly performed in the payment account corresponding to the identity information of the target object. The embodiments of the present application can adjust the payment risk control level of the terminal based on the scene recognition result.

[0115] In the embodiments of the present application, the scene recognition result includes a first scene recognition result and a second scene recognition result, which refer to scene recognition results obtained after scene recognition based on different scene recognition information. The scene recognition information corresponding to the first scene recognition result is different from the scene recognition information corresponding to the second scene recognition result. In step S1051, the payment risk control level of the terminal is adjusted based on the scene recognition result, which can be implemented in the following manner: when the first scene recognition result and the second scene recognition result are both that the terminal is currently located in the first type of scene, the payment risk control level of the terminal is adjusted from the first level to the second level; when any one of the first scene recognition result and the second scene recognition result is that the terminal is currently located in the first type of scene, the payment risk control level is adjusted from the first level to the third level; when the first scene recognition result and the second scene recognition result are both that the terminal is currently located in the second type of scene, the payment risk control level is maintained at the first level; the second type of scene is any scene other than the first type of scene, and the first level, the second level and the third level correspond to different payment verification manners, wherein the process complexity of the first payment verification manner corresponding to the first level is higher than the process complexity of the second payment verification manner corresponding to the second level; the process complexity of the third payment verification manner corresponding to the third level is lower than the process complexity of the first payment verification manner, and the process complexity of the third payment verification manner is higher than the process complexity of the second payment verification manner. That is, the process complexity of the payment verification manner corresponding to the first level is higher than the process complexity of the payment verification manner corresponding to the second level; the process complexity of the payment verification manner corresponding to the third level is lower than the process complexity of the payment verification manner corresponding to the first level, but the process complexity of the payment verification manner corresponding to the third level is higher than the process complexity of the payment verification manner corresponding to the second level, that is, the process complexity of the payment verification manner corresponding to the third level is between the process complexity of the payment verification manner corresponding to the first level and the process complexity of the payment verification manner corresponding to the second level.

[0116] For example, the first scene recognition result can be a scene recognition result corresponding to point of interest information, and the second scene recognition result can be a scene recognition result corresponding to a background image. The default payment risk control level of the terminal is a first level. When the first scene recognition result and the second scene recognition result are both that the terminal is currently located in a first type of scene, the payment risk control level is adjusted from the first level to a second level. The payment verification manner corresponding to the second level has a simpler process than the payment verification manner corresponding to the first level. When the first scene recognition result is that the terminal is currently located in the first type of scene and the second scene recognition result is that the terminal is currently located in a second type of scene, the payment risk control level is adjusted from the first level to a third level. The payment verification manner corresponding to the third level has a simpler process than the payment verification manner corresponding to the first level, but the payment verification manner corresponding to the third level has a more complex process than the payment verification manner corresponding to the second level. When the first scene recognition result and the second scene recognition result are both that the terminal is currently located in the second type of scene, the payment risk control level is maintained as the first level.

[0117] In the embodiments of the present application, after obtaining the scene recognition information of the terminal, because the scene recognition information of the terminal includes different information, for example, point of interest information, network information of a network currently accessed by the terminal, a wireless connection signal received by the terminal, and a background image collected by the terminal in a current scene, scene recognition can be performed based on these information respectively, and corresponding scene recognition results are obtained. In the specific implementation process, the scene recognition results obtained based on each kind of information can be the same or different. When the scene recognition results obtained based on each kind of information are different, the different scene recognition results can be further verified, and finally the final scene recognition result is determined based on the verified result.

[0118] When verifying the scene recognition result, new information can be re-obtained for scene recognition. For example, when verifying the scene recognition result determined based on the point of interest information, the point of interest information of the current location of the terminal can be re-obtained, and then the scene recognition result of the terminal is determined based on the re-obtained point of interest information. For another example, when verifying the scene recognition result determined based on the background image, the background image of the current scene of the terminal can be re-collected, and then the scene recognition result of the terminal is determined based on the re-collected background image.

[0119] It should be noted that the embodiments of the present application are only exemplified by taking that the scene recognition result includes the first scene recognition result and the second scene recognition result, and in actual application, the scene recognition result can also include a plurality of scene recognition results, and the payment risk control level can also be divided into other levels, which are not limited by the embodiments of the present application.

[0120] In step S1052, based on the adjusted payment risk control level and the identity information of the target object, payment processing corresponding to the payment request is performed to obtain a payment result.

[0121] According to the scene recognition result of the terminal, the payment risk control level can be dynamically adjusted, so as to reduce the payment operation complexity of the target object in a specific scene and improve the payment experience of the user.

[0122] In some embodiments, in step S1052, based on the adjusted payment risk control level and the identity information of the target object, payment processing corresponding to the payment request is performed to obtain a payment result, which can be implemented in the following manner: first, the adjusted payment risk control level is determined, and when the adjusted payment risk control level is a first level, the reserved verification information of the target object is obtained; then, a verification request is sent to the terminal, wherein the terminal will display a verification interface in response to the verification request after receiving the verification request; thereafter, the terminal receives the to-be-verified information input by the target object through the verification interface, and sends the to-be-verified information to the server. After receiving the to-be-verified information sent by the terminal, the server judges the to-be-verified information and the reserved verification information. If the to-be-verified information is the same as the reserved verification information, the payment operation is performed on the payment account corresponding to the identity information of the target object, and a payment result is generated, thereby completing the payment processing process corresponding to the payment request.

[0123] Here, the payment verification mode corresponding to the first level can include an additional verification operation on all cross-region payment objects. When the payment risk control level of the terminal is the first level, and at this time the target object is performing a cross-region payment, the target object's reserved verification information is obtained while matching the target object's identity information through the full-amount feature library. The reserved verification information is data or information uploaded by the target object when registering a payment account to identify the target object's identity, including but not limited to name, birthday, address, phone number, and email address, etc. For example, the reserved verification information can be the last four digits of the target object's mobile phone number. After obtaining the target object's reserved verification information, a verification request can be sent to the terminal, which is used to request the target object to input the to-be-verified information to re-verify the target object's identity. After receiving the verification request, the terminal displays a verification interface in response to the verification request. The verification interface displays prompt information and an input box, and the prompt information can be a prompt to input the to-be-verified information. For example, the verification interface displays four input boxes for the last four digits of the mobile phone number, and a prompt for the target object to input the last four digits of the mobile phone number. After the target object inputs the to-be-verified information through the verification interface, the terminal sends the to-be-verified information to the server. After receiving the to-be-verified information, the server determines whether the to-be-verified information is the same as the reserved verification information. When the to-be-verified information is the same as the reserved verification information, the target object's identity information verification is successful, and the payment operation is performed on the payment account corresponding to the target object's identity information, and a payment result is generated. For example, the target object inputs the to-be-verified information as "9920" through the verification interface, and the target object's reserved verification information is "9920". Since the to-be-verified information is the same as the reserved verification information, the target object's identity information verification is successful.

[0124] When the payment risk control level is the first level, the embodiments of the present application need to perform reserved verification information verification on the target object of the cross-region payment, which ensures that the target object's identity information matching is correct, reduces the misbrush rate, and improves the accuracy and security of the payment.

[0125] In some embodiments, when the payment risk control level is the first level, in addition to the above-mentioned reserved verification information verification method, any one of the following verification methods can be used to verify the target object's identity information: multi-factor authentication (MFA), digital certificate, behavior pattern recognition, payment tokenization, and risk assessment and detection.

[0126] When performing multi-factor authentication, a password + hardware token method can be used, and the target object needs to enter a password and use a one-time password (OTP) generated by a hardware token for verification. Alternatively, a password + mobile application verification method can also be used, and the target object enters a password and then uses a verification code generated by an authentication application on the mobile phone for verification. When performing digital certificate authentication, identity verification based on digital certificates can be implemented. The target object performs identity verification through a digital certificate, and the digital certificate is issued by a trusted certificate authority to ensure the authenticity and security of the identity. Alternatively, a public key infrastructure (PKI) technology can also be used to perform identity verification using PKI to ensure the security of transactions through digital certificates and public and private key encryption technology. When performing behavior pattern recognition authentication, the behavior characteristics of the target object, such as operation habits and device usage patterns, can be analyzed to identify abnormal behavior and perform risk assessment. When performing payment tokenization authentication, payment information encryption can be used to replace sensitive payment information with unidentifiable tokens to ensure that even if the data is stolen, hackers cannot use this information to conduct illegal transactions. When performing risk assessment and detection, real-time risk assessment can be performed. The payment platform analyzes transaction behavior in real time to identify abnormal patterns, and once suspicious transactions are detected, the system will take immediate action to conduct manual verification or reject the transaction.

[0127] It should be noted that when performing authentication, one authentication method can be selected for authentication, or multiple authentication methods can be selected for authentication. These authentication methods can be selected and combined according to different payment scenarios and security requirements to provide a safer and more convenient payment experience.

[0128] In some embodiments, in step S1052, based on the adjusted payment risk control level and the identity information of the target object, the payment operation corresponding to the payment request is performed to generate a payment result, which can also be implemented in the following way: when the adjusted payment risk control level is the second level, the payment operation is performed on the payment account corresponding to the identity information of the target object to generate the payment result. That is, the payment account corresponding to the identity information of the target object can be processed to obtain the payment result.

[0129] Here, the payment verification mode corresponding to the second level is that no additional verification operation is performed on the cross-region payment object. When the payment risk control level of the terminal is the second level and the terminal is in the first type of scene, that is, the terminal is in a scene where the cross-region payment proportion is greater than the preset proportion threshold. At this time, even if the target object cross-region payment, no additional verification operation is required, and the payment operation is directly performed on the payment account corresponding to the identity information of the target object to generate a payment result. If the payment account is successfully debited, a payment success result is obtained. If the payment account fails to debit, a payment failure result is obtained.

[0130] According to the scene recognition result of the terminal, the payment risk control level can be dynamically adjusted, so that when the terminal is in a specific first type of scene such as a highway service area or a gas station, the payment risk control level can be reduced to the second level, thereby reducing the verification proportion of any one or more verification methods such as mobile phone verification. In this way, while ensuring payment accuracy and security, the payment process and operation complexity are simplified, thereby improving the user's payment experience.

[0131] In some embodiments, in step S1052, based on the adjusted payment risk control level and the identity information of the target object, a payment operation corresponding to the payment request is performed to generate a payment result. It can also be implemented in the following way: first, when the adjusted payment risk control level is the third level, the similarity between the image features corresponding to each object in the full-amount feature library and the biological feature recognition information corresponding to the to-be-recognized image features is obtained; then, when there are at least two target image features, the reserved verification information of the target object is obtained, where the target image feature is an image feature whose similarity with the to-be-recognized image feature is greater than a preset similarity threshold; then, a verification request is sent to the terminal, wherein the terminal will display a verification interface in response to the verification request after receiving the verification request; then, the terminal receives the to-be-verified information input by the target object through the verification interface, and sends the to-be-verified information to the server. After receiving the to-be-verified information sent by the terminal, the server judges the to-be-verified information and the reserved verification information. If the to-be-verified information is the same as the reserved verification information, a payment operation is performed on the payment account corresponding to the identity information of the target object, and a payment result is generated, thereby completing the payment processing process corresponding to the payment request.

[0132] Here, the payment verification mode corresponding to the third level can be an additional verification operation on only part of the cross-region payment object. When the payment risk control level of the terminal is the third level, and at this time the target object performs cross-region payment, when the identity information of the target object is matched through the full-amount feature library, if there are multiple target image features in the full-amount feature library that have a similarity with the to-be-identified image feature exceeding a preset similarity threshold, the reserved verification information of the target object is obtained. Then, based on the reserved verification information and the to-be-verified information input by the target object through the verification interface, the identity information of the target object is verified. If there is only one target image feature in the full-amount feature library that has a similarity with the to-be-identified image feature exceeding a preset similarity threshold, a deduction operation is directly performed on the payment account corresponding to the identity information of the object corresponding to the image feature.

[0133] When the payment risk control level is the third level, the application embodiment performs reserved verification information verification on part of the cross-region payment target object, thereby reducing the proportion of mobile phone verification, simplifying the payment process and operation complexity while ensuring payment accuracy and security, and improving user payment experience.

[0134] In the scenario where the target object uses the terminal to perform a payment operation, the application embodiment can first match the biometric feature recognition information sent by the terminal from the regional feature library to determine the identity information. If the matching fails, the biometric feature recognition information is matched from the full-amount feature library to further determine the identity information, wherein the regional feature library is a feature library determined based on the current geographical location information of the terminal. This feature library is a feature library that has a mapping relationship with the current location of the terminal, and it contains the feature information of the users in the current location rather than the feature information of all users. In this way, compared with directly using the full-amount feature library to match the identity information of the target object in the related art, the application embodiment pre-establishes a regional feature library for different regions, first matches the biometric feature recognition information of the target object using the regional feature library of the current region of the terminal, reduces the size of the feature library, and can more accurately identify the identity information of the target object, thereby reducing the probability of false brushing and improving the accuracy and security of payment. In addition, when the matching in the regional feature library fails, i.e., when the target object performs cross-region payment, the application embodiment further determines a scene recognition result based on the scene recognition information of the terminal, and performs a payment operation in combination with the identity information matched from the full-amount feature library and the scene recognition result, thereby further improving the accuracy and security of payment.

[0135] In some embodiments, referring to FIG. 6, FIG. 6 shows that before determining the regional feature library of the region where the terminal is currently located based on the geographical location information in step S102, the payment method provided by the application embodiment further includes the following steps S201 to S204:

[0136] In step S201, position data of each object in a preset object set in a preset historical time period when the object makes a payment is acquired.

[0137] Here, the preset historical time period can be set by the user, for example, one month, one year, etc. The preset object set can also be set by the user. In the embodiment of the present application, the preset object set can be all the objects registered in the payment account. The position data is the location where the object makes a payment, which can include the location name and address, etc. For example, the position data of the object registered in the payment account when making a payment in one year can be acquired.

[0138] In step S202, a target area of the object is determined from the position data of the object based on the number of times the object makes a payment in each position data.

[0139] Here, for any object, the area where the position data of the object making a payment most frequently is acquired, and the area is determined as the target area of the object. For example, object a makes a payment 100 times in one year, 80 times of which are in city A, 10 times in city B, and 10 times in city C. The target area of the object is city A.

[0140] In other embodiments, for any object, the position data of the object making a payment at different times can also be classified by a clustering algorithm, and the position data is divided into different location groups. Then, the location group containing the most position data is selected, and the location corresponding to the location group is determined as the target area of the object. The clustering algorithm is not limited in the embodiment of the present application, for example, it can be a K-means clustering algorithm.

[0141] In step S203, the image feature of the object is added to the area feature library of the target area in response to the existence of the area feature library of the target area.

[0142] For example, the target area of object a is city A, and the area feature library of city A already exists. Then, the image feature and identity information of object a are added to the area feature library of city A.

[0143] In step S204, the area feature library of the target area is constructed in response to the non-existence of the area feature library of the target area, and the image feature of the object is added to the area feature library of the target area.

[0144] For example, the target area of object a is city A, and if the area feature library of city A does not exist, the area feature library of city A is constructed, and then the image feature and identity information of object a are added to the area feature library of city A.

[0145] The embodiment of the application can reduce the size of the feature library used in matching the identity information of the target object, can more accurately identify the identity information of the target object, thereby reducing the probability of mistaken payment, and improving the accuracy and security of payment.

[0146] FIG. 7 is another optional flowchart of a payment method provided by the embodiment of the application, as shown in FIG. 7, the method comprises the following steps S301 to S313:

[0147] In step S301, the terminal receives an interactive operation of a user.

[0148] Here, the interactive operation can be a click payment operation, etc. The user can perform any kind of interactive operation through the terminal on the client of the payment application, for example, a purchase payment operation.

[0149] In step S302, the terminal collects biological feature recognition information of a target object in response to the interactive operation.

[0150] When the terminal receives the interactive operation of the user, it is determined that payment is currently required, therefore, in response to the interactive operation, the biological feature recognition information of the target object can be collected through a sensor, a camera or other biological feature collection device installed on or externally connected to the terminal device, and the collected biological feature recognition information is recorded.

[0151] In the embodiment of the application, different collection methods can be used to collect biological feature recognition information for different biological feature recognition information. For fingerprint information, an optical or capacitive sensor can be used to collect a fingerprint image, then the fingerprint ridges and branch points are extracted to obtain fingerprint feature recognition information; for face information, a two-dimensional or three-dimensional image of the face can be captured through a camera, then the relative positions of the eyes, nose and mouth are extracted to obtain face feature recognition information; for iris information, the texture features of the iris can be captured by an infrared camera, then the unique patterns in the iris texture are extracted to obtain iris feature recognition information; for vein information, the vein distribution map of the finger, palm or back of the hand can be obtained through an infrared camera, then the feature points in the vein distribution map are extracted to obtain vein feature recognition information; for voiceprint information, a microphone can be used to capture the voice of the user.

[0152] The collection technology and equipment used when collecting the biometric information includes, but is not limited to, any one of the following: an optical sensor, an infrared light source, and intelligent positioning technology. The optical sensor, such as a Charge-Coupled Device (CCD) or a Complementary Metal-Oxide-Semiconductor (CMOS), is used to capture images. The infrared light source is used for iris and vein recognition to improve the clarity of the features. The intelligent positioning technology can automatically adjust the position and angle of the collection equipment to ensure that a high-quality biometric signal is collected.

[0153] In step S303, the terminal generates a payment request.

[0154] After the biometric information of the target object is collected, the terminal encapsulates the biometric information of the target object into the payment request, and the payment request is used to request the server to perform a deduction operation on the payment account of the target object.

[0155] In the palm swipe payment scenario, there can also be a palm swipe payment device specially used for palm swipe payment. The palm swipe payment device is used to collect an image of the palm of the target object, and then encapsulate the collected palm image into the payment request. Alternatively, the terminal can be implemented as a palm swipe payment device. In response to an interaction operation, the terminal can start the camera to collect an image of the palm of the target object.

[0156] In step S304, the terminal sends the payment request to the server.

[0157] In some embodiments, the terminal is a palm swipe payment device. The palm swipe payment device can have a network module and can send the payment request to the server through the network module. In other embodiments, the palm swipe payment device can also be connected to a terminal, and the terminal runs an application program of a specific service (such as a payment service). After collecting the image of the palm of the target object, the palm swipe payment device can send the image of the palm of the target object to the terminal. The terminal encapsulates the image of the palm of the target object as the biometric information of the target object into the payment request by running the application program, and sends the payment request to the server. In some embodiments, the terminal can use HTTP or Web Socket protocol to send the payment request.

[0158] In step S305, the server acquires geographic location information of the location where the terminal is currently located in response to the payment request sent by the terminal.

[0159] Here, the terminal currently located geographical location information can be encapsulated with the target object biometric information together to the payment request. Or, the server in response to the terminal sent payment request, can again send a terminal geographical location query instruction, the terminal receives the geographical location query instruction, through the positioning module (terminal built-in map application function module) to obtain the geographical location information, and sends the geographical location information to the server.

[0160] Step S306, the server determines the terminal based on geographical location information, the area characteristics of the region where the library.

[0161] Here, for the specific implementation of determining the terminal based on geographical location information, the area characteristics of the region where the library can be seen in the above step S102 description, here will not be repeated.

[0162] Step S307, the server through the area characteristics of the library for biometric information matching.

[0163] Here, through the area characteristics of the library for biometric information matching, refers to determine whether there is a corresponding feature information from the biometric information of the characteristics of the information. With the corresponding biometric information of the characteristics of the information refers to the similarity between the biometric information is greater than or equal to the similarity threshold value of the feature information. When there is at least one similarity greater than or equal to the similarity threshold value of the feature information, then the feature information in the area characteristics of the library and the biometric information matching success.

[0164] Step S308, when the biometric information in the area characteristics of the library matching success, the identity information of the target object, the server in the target object identity information corresponding to the payment account to perform payment operation, get the payment result.

[0165] In the present application embodiment, the similarity greater than or equal to the similarity threshold value, and the maximum similarity of one image feature can be screened out, and the identity information of the object corresponding to the image feature is determined as the identity information of the target object. Then, the identity information of the target object corresponding to the payment account can be obtained, and the payment request corresponding to the payment account can be executed, and the payment result can be generated.

[0166] Step S309, when the biometric information in the area characteristics of the library matching fails, the server through the full feature library for biometric information matching, get the identity information of the target object.

[0167] When the similarity between the image feature of each object in the region feature library and the to-be-identified image feature is less than the preset similarity threshold, the image feature in the region feature library fails to match the biometric feature recognition information, indicating that the image feature of the target object does not exist in the region feature library.

[0168] The full-amount feature library includes region feature libraries of multiple regions. For the specific implementation of matching the biometric feature recognition information by using the full-amount feature library to obtain the identity information of the target object, refer to the description of step S103 above, which is not repeated here.

[0169] In step S310, the server acquires the scene recognition information of the terminal and determines the scene recognition result of the terminal based on the scene recognition information.

[0170] For the specific implementation of acquiring the scene recognition information of the terminal and determining the scene recognition result of the terminal based on the scene recognition information, refer to the description of step S104 above, which is not repeated here.

[0171] In step S311, the server performs a payment operation based on the scene recognition result and the identity information of the target object.

[0172] For the specific implementation of performing the payment operation based on the scene recognition result and the identity information of the target object, refer to the description of step S105 above, which is not repeated here.

[0173] In step S312, the server sends the payment result to the terminal.

[0174] In step S313, the terminal displays the payment result on the current interface.

[0175] In the scenario where the target object uses the terminal to perform a payment operation, in the embodiments of this application, a region feature library is established in advance for different regions. In this way, the server can first match the biometric feature recognition information sent by the terminal from the region feature library of the region where the terminal is currently located, so as to determine the identity information of the target object. If the matching fails in the region feature library, the biometric feature recognition information will be further matched from the full-amount feature library, so as to further determine the identity information of the target object. Such an identity authentication process, since the biometric feature recognition information of the target object is first matched by using the region feature library of the region where the terminal is currently located, the information of the region feature library is less than that of the full-amount feature library, that is, the size of the feature library is reduced, and the identity information of the target object can be more accurately identified, thereby reducing the probability of false brushing and improving the accuracy and security of payment. Moreover, if the matching fails in the region feature library, that is, the target object is currently making a cross-region payment, the scene recognition result will be further determined based on the scene recognition information of the terminal, and the payment processing will be performed in combination with the scene recognition result and the identity information matched from the full-amount feature library, so as to ensure the security during cross-region payment, and thus improve the accuracy and security of the entire payment process.

[0176] In the following, an exemplary application of the embodiments of the present application in an actual application scenario will be described.

[0177] The embodiments of the present application provide a payment method, which is a dynamic palm brushing cross-region risk control interception payment method based on geographic location and scene detection, and can be applied to a palm brushing payment scenario. In order to reduce the false brushing rate, a fixed city dimension risk control strategy is adopted, when the user uses across cities, the mobile phone number verification will be triggered, and the user experience will be reduced. In the scenarios such as highway service stations and gas stations where the proportion of cross-city users is high, the above shortcomings are more obvious. In summary, the related technical solutions mainly have the following shortcomings: first, since the related technology is mainly based on a fixed geographic location risk control strategy, when the user uses across cities, the system is easy to misjudge, resulting in the user needing to perform mobile phone number verification, thereby reducing the user experience: when performing risk control interception, the device location information is mainly relied on, but the specific scene where the user is located cannot be accurately judged, so in specific scenarios such as highway service stations and gas stations, the risk control strategy effect is poor, which easily leads to false brushing and low user experience. Secondly, it is not possible to optimize the processing for specific scenarios such as highway service stations and gas stations; finally, the risk control strategy is usually fixed, and cannot be dynamically adjusted according to the scene where the user is located, so the effect may not be good when dealing with complex scenarios.

[0178] The embodiment of the present application adopts an algorithm strategy of device location information combined with scene detection to accurately determine a cross-city scene, that is, to determine a scene such as a gas station and a highway service station where a proportion of cross-city users is greater than a preset proportion threshold. If located in the scene, detection is performed from a full-amount feature library, thereby reducing the proportion of mobile phone verification and improving the accuracy and security of palm payment and the palm payment experience. Specifically, first, the embodiment of the present application establishes a regional feature library for different cities through commonly used palm payment. This regional feature library based on geographic location information can reduce the size of the library and more accurately identify the palm features of the user, thereby reducing the probability of false brushing and improving the user experience. Second, the embodiment of the present application can more accurately determine the scene where the user is located through a dynamic risk control strategy combining scene detection and device location information, thereby dynamically adjusting the risk control strategy according to the identified scene, reducing the proportion of mobile phone verification, and improving the user experience. Finally, the embodiment of the present application can identify whether the user is in a specific scene such as a highway service station and a gas station, and perform detection from a full-amount feature library in these scenes to reduce the proportion of mobile phone verification and improve the user experience.

[0179] To facilitate understanding, the palm payment process and the mobile phone verification process are described below. FIGS. 8a and 8b are schematic diagrams of different palm payment devices provided by the embodiment of the present application. FIG. 9 is a flowchart of the palm payment provided by the embodiment of the present application. Referring to FIG. 9, after the user performs a payment operation on the palm payment device 900, the circular area 901 on the palm payment device 900 senses the user's palm and collects a palm image, and the palm image is sent to a server. The palm payment device 900 displays "payment in progress" on the current interface. After the server receives the palm image, the corresponding user can be found according to the palm image. If the user is found, the corresponding payment operation can be performed based on the payment account of the user, and a payment result representing the result corresponding to the payment operation is generated. The server feeds back the generated payment result to the palm payment device 900. The palm payment device 900 displays the payment result 902 on the current interface, such as payment success and the specific payment amount. FIG. 10 is a schematic diagram of a mobile phone verification interface provided by the embodiment of the present application. Referring to FIG. 10, when mobile phone verification is triggered, the server finds the user according to the palm image, obtains the reserved mobile phone number of the payment account, sends a mobile phone verification request to the palm payment device, and the palm payment device presents a mobile phone verification interface. After the user inputs the last four digits of the mobile phone number, the palm payment device sends the information input by the user to the server for comparison and verification with the reserved mobile phone number.

[0180] The specific implementation process of the payment method provided by the embodiment of the present application is described below.

[0181] First, the geographic location frequently used by the user is reported, and a regional feature library of city dimension is established. The geographic location information collection can be realized by the following way: when the user performs the palm swiping operation, the geographic location information (including latitude and longitude and POI) is collected through the map SDK module of the palm swiping payment device. The geographic location information can be collected by means of GPS positioning technology, base station positioning and WIFI positioning and the like. The user frequently used place recognition can be realized by the following way: the user's frequently used place is recognized by analyzing the geographic location data of the user in a period of time. For example, the places visited by the user at different times can be classified by a clustering algorithm, so as to find out the frequently used place of the user (which can be a city, corresponding to the region in the above embodiment). The regional feature library of city dimension can be realized by the following way: according to the recognized frequently used place of the user, a regional feature library is established for each city, and the palm swiping feature information (corresponding to the image feature in the above embodiment) of the user in the city is stored in the regional feature library. The regional feature library can be stored and managed by using database and file system and the like. When the regional feature library needs to be updated, the regional feature library can be updated regularly along with the change of the geographic location information of the user and the addition of new users. For example, a time period can be set, such as monthly or quarterly, to update the regional feature library. In the updating process, the new palm swiping feature information needs to be added to the regional feature library of the corresponding city, and the expired or invalid feature information needs to be deleted.

[0182] In some embodiments, the palm swiping feature information of the same user can be stored in one regional feature library or in multiple regional feature libraries. For example, for a user whose frequently used place is two cities, the palm swiping feature information of the user can be stored in the regional feature libraries of the two cities.

[0183] When the user performs the palm swiping operation, the corresponding city dimension regional feature library can be selected according to the geographical location information of the user to perform palm swiping recognition. By using the city dimension regional feature library, the accuracy of palm swiping recognition can be improved, and the probability of false swiping can be reduced. If the matching user cannot be searched in the city dimension regional feature library, the full-amount feature library is used for retrieval, and the last four digits of the mobile phone number are used for risk control strategy. Then, when the matching user cannot be searched in the city dimension regional feature library, but the user is searched through the full-amount feature library, the POI information of the palm swiping payment device can be reported, and it is determined whether the current interest point is located in a highway service area or a gas station and the like. The POI information of the device can be collected by using a map SDK to query one or more interest point information near the geographical location, and the interest point information usually includes name, type, address and the like. Then, it is determined whether the current interest point is located in a highway service area or a gas station and the like. After obtaining the interest point information, the interest point information can be traversed to check whether the name or type of the interest point contains the keywords such as “highway service area” or “gas station”. When there is at least one interest point information containing the keywords, it means that the current device is located in a specific scene (corresponding to the first type of scene in the above embodiment) such as a highway service area or a gas station.

[0184] In some embodiments, a scene picture (corresponding to the background image in the above embodiment) can also be captured by the camera after each startup of the palm swiping payment device, and when the matching user cannot be searched in the city dimension regional feature library, but the user is searched through the full-amount feature library, the scene picture is input into the trained model to determine whether the palm swiping payment device is in a scene such as a gas station or a highway. The surrounding scene picture is automatically captured by the camera at each startup of the palm swiping payment device. Since the camera on the palm swiping payment device cannot move, the captured scene picture may not have features that can be identified by the scene. At this time, other scene recognition methods can also be used for scene recognition. The captured scene picture is uploaded to the server, and the scene picture can be compressed during the uploading process to reduce transmission time and bandwidth consumption. After the server receives the uploaded scene picture, the scene picture is input into the trained model for scene recognition to obtain a scene recognition result. The model can identify specific scenes such as highway service areas and gas stations. The server returns the scene recognition result to the palm swiping payment device. According to the scene recognition result, the palm swiping payment device can perform corresponding operations, such as adjusting the risk control strategy or reminding the user and the like.

[0185] The embodiment of the application can use a convolutional neural network model (CNN) to implement the scene recognition process described above. The convolutional neural network model is a deep learning model and is widely used in image recognition, speech recognition, natural language processing and other fields. In the field of image recognition, the convolutional neural network model has become one of the mainstream models. The convolutional neural network model can be trained using a machine learning platform TensorFlow.

[0186] FIG. 11 is a schematic diagram of a convolutional neural network structure provided by the embodiment of the application. Referring to FIG. 11, the core of the convolutional neural network model is a convolutional layer and a pooling layer (down-sampling layer). The convolutional layer extracts the features of an image through convolution operation, and the pooling layer (down-sampling layer) reduces the dimension of the features through down-sampling operation, thereby reducing the amount of calculation and the amount of parameters. After the convolutional layer and the pooling layer (down-sampling layer), a fully connected layer and an activation function are usually added to perform classification or regression tasks. For an image recognition task, the convolutional neural network model usually adopts a structure of multiple convolutional layers and pooling layers (down-sampling layers) to extract multi-level features of the image. In the training process, the convolutional neural network model updates the model parameters through a back propagation algorithm to minimize the loss function. In the testing process, the convolutional neural network model performs prediction through a forward propagation algorithm and outputs the class or probability distribution of the image. The convolutional layer enhances the input signal through convolution operation and reduces noise. The input of the convolutional layer is connected to the receptive field 1101 of the upper layer to extract features and obtain a feature map 1102, and at the same time, the positional relationship between the features is obtained. The calculation method of the features output by the convolutional layer satisfies the following formula (5).

[0187] wherein “*” represents convolution operation; is the output of the jth neuron of the lth layer after convolution operation; represents the output of the ith neuron of the (l-1)th layer, i.e., the input data of the lth layer; represents a filter; is a bias value; f(·) represents a nonlinear activation function. The activation functions commonly used in neural networks are mainly sigmoid function, tanh function and ReLU function. Compared with the sigmoid function and the tanh function, the ReLU function has a faster training speed and higher accuracy for training the convolutional neural network.

[0188] The downsampling layer is also called the pooling layer. There is a downsampling layer after each convolutional layer in the convolutional neural network. Because the amount of feature data obtained through the convolutional layer is too large, if it is directly used for training, it will increase the complexity of the network model. Therefore, the dimension of the feature can be reduced by pooling the local feature. At the same time, the pooling processing can improve the overfitting problem of the convolutional neural network to a certain extent. After two times of feature extraction, the convolutional neural network can have strong distortion tolerance to the input image.

[0189] The model training method of the convolutional neural network based on the machine learning platform TensorFlow will be described below.

[0190] When training the convolutional neural network, first, the data set is prepared. After the first start of the camera of the brush palm payment device, the scene picture is captured. The camera of the terminal device is used to take pictures of the surrounding scene, and the data of the pictures is uploaded to the server for training of the scene model. After the brush palm payment device is deployed to the offline environment, the camera is used to take pictures at the first start, and a scene picture is uploaded to the server. The server stores the images uploaded by each brush palm payment device separately, and obtains ten thousand pictures of the whole scene taken by ten thousand brush palm payment devices for model training. Before training, the geographical position information can be combined to perform manual labeling. The standard of manual labeling includes whether it is in a gas station, a highway service area, and other scenes with a high cross-city ratio (manual labeling only labels specific scenes or non-specific scenes). Among them, 8000 pictures are used as a training set, and 2000 pictures are used as a test set for model training.

[0191] After the data set is prepared, the model is constructed. In the machine learning platform TensorFlow, the convolutional neural network model CNN can be constructed using tf.keras. Here, tf.keras is a Keras API that can be seamlessly integrated with the machine learning platform TensorFlow. Keras is an open source neural library written in the programming language python. Existing network structures can be selected, such as the deep convolutional neural network (Visual Geometry Group Network, VGG) model, the deep residual network (Residual Network, ResNet), or the neural network Inception, etc. The network structure can also be customized. When constructing the convolutional neural network model, the input layer, the convolutional layer, the pooling layer, and the fully connected layer of the convolutional neural network model need to be defined. Techniques such as random inactivation (Dropout) or batch normalization (Batch Normalization) can be used to prevent overfitting.

[0192] After the model is built, model training is performed. After the convolutional neural network model is built, the prepared data set can be used to train the convolutional neural network model. Model training can be performed using the fit method in tf.keras, where the fit method refers to training the model for a fixed number of rounds (number of iterations on the data set). During the training process, the convolutional neural network model needs to be optimized to improve the accuracy and generalization ability of the convolutional neural network model. Different optimization algorithms can be selected, such as Stochastic Gradient Descent (SGD) or Adaptive Moment Estimation (Adam). At the same time, the convolutional neural network model needs to be tuned to obtain the best model performance.

[0193] After the model training is completed, model evaluation is performed. After the model training is completed, the convolutional neural network model needs to be evaluated to determine the accuracy and generalization ability of the convolutional neural network model. The validation set can be used to evaluate the convolutional neural network model, or the test set can be used to evaluate the convolutional neural network model. Model evaluation can be performed using the evaluate method in tf.keras. Common evaluation metrics include accuracy, precision, recall, and F1 value.

[0194] After the model evaluation is completed, model application is performed. The trained and evaluated convolutional neural network model can be used for image scene detection. For a new image, it can be input into the convolutional neural network model, and the convolutional neural network model will output a probability distribution representing the probability of the image belonging to different scene categories. The category with the highest probability can be selected as the scene category of the image.

[0195] After obtaining the scene recognition result of the palm swipe payment device, the geographical location information and the scene recognition result can be combined to determine whether to perform risk control degradation for the palm swipe payment device. In the implementation process, the server determines whether the palm swipe payment device is in a specific scene, such as a highway service area or a gas station, according to the geographical location information and the scene recognition result uploaded by the palm swipe payment device. If one of them is determined to be in a special scene, risk control degradation is performed. For the adjustment strategy of risk control degradation, the server can determine whether to perform risk control degradation for the device according to the judgment result. For example, if the palm swipe payment device is in a specific scene such as a highway service area or a gas station, the server can reduce the proportion of mobile phone verification and improve the user's payment experience.

[0196] The payment method provided in the embodiments of the present application can more accurately identify the palm swiping information of a user by combining the geographic position information of the device and the scene recognition result, thereby improving the accuracy of palm swiping recognition. According to the geographic position information of the palm swiping device and the scene recognition result, the risk control strategy can be dynamically adjusted. For example, when the palm swiping device is in a specific scene such as a highway service area or a gas station, the proportion of mobile phone number verification can be reduced, thereby improving the user experience. By optimizing the risk control strategy, the operation complexity of the user in a specific scene can be reduced, thereby improving the payment experience of the user. The embodiments of the present application can improve the intelligent degree of the payment system by using deep learning and machine learning technologies for scene recognition, so that the payment system can better adapt to various complex scenes. The embodiments of the present application can effectively reduce the mis-swiping rate by improving the accuracy of palm swiping recognition and optimizing the risk control strategy, thereby improving the stability and reliability of the payment system.

[0197] It should be noted that in addition to the above-provided interest point information and the method of using the model to classify the image captured by the device to obtain the scene recognition result, other scene recognition methods can also be used in the payment method provided in the embodiments of the present application, for example, scene recognition based on WiFi information. Specifically, the WiFi information around the palm swiping device can be collected to perform scene recognition. According to the name and signal strength of the WiFi hotspot and other information, it can be inferred whether the palm swiping device is in a specific scene, such as a gas station or a highway service area. Scene recognition can also be based on Bluetooth beacons (Beacon). For example, Bluetooth beacon devices are deployed in specific scenes (such as gas stations or highway service areas). When the palm swiping device of the user receives the beacon signal, it can be inferred that the palm swiping device is in a specific scene (as long as the Bluetooth name is searched, it can be determined as a specific scene). The advantage of this method is that the beacon device can provide more accurate position information, thereby improving the accuracy of scene recognition.

[0198] The following continues to illustrate an example structure of the payment device 455 provided by the embodiments of the present application, which is implemented as a software module. In some embodiments, as shown in FIG. 2, the software module stored in the payment device 455 of the memory 450 can include: a position acquisition module 4551, configured to acquire geographic position information of a current position of a terminal in response to a payment request sent by the terminal; the payment request includes biological feature identification information of a target object; a feature library acquisition module 4552, configured to determine a regional feature library of an area where the terminal is currently located based on the geographic position information; a matching module 4553, configured to, in response to a failure of the biological feature identification information to match in the regional feature library, match the biological feature identification information through a full-amount feature library to obtain identity information of the target object; the full-amount feature library includes regional feature libraries of multiple areas; a scene recognition module 4554, configured to acquire scene recognition information of the terminal and determine a scene recognition result of the terminal based on the scene recognition information; and a payment module 4555, configured to perform payment processing based on the scene recognition result and the identity information of the target object.

[0199] In some embodiments, the biological feature identification information includes a biological feature identification image; the matching module 4553 is further configured to perform feature extraction on the biological feature identification image to obtain to-be-recognized image features; determine a similarity between the to-be-recognized image features and image features of each object in the full-amount feature library; and determine identity information of an object corresponding to image features with the maximum similarity as the identity information of the target object.

[0200] In some embodiments, the scene recognition module 4554 is further configured to acquire a plurality of point-of-interest information from the scene recognition information; the point-of-interest information includes scene information within a preset range of the current position of the terminal; and in response to searching for a first keyword in at least one point-of-interest information, determine that the scene recognition result of the terminal is that the terminal is currently located in a first type of scene.

[0201] In some embodiments, the scene recognition module 4554 is further configured to acquire network information of a network currently accessed by the terminal from the scene recognition information; determine a network name and a signal strength of the network currently accessed from the network information; and determine the scene recognition result of the terminal based on the network name and the signal strength.

[0202] In some embodiments, the scene recognition module 4554 is further configured to, in response to searching for a preset second keyword from the network name and the signal strength being greater than a preset strength threshold, determine that the scene recognition result of the terminal is that the terminal is currently located in a first type of scene.

[0203] In some embodiments, the scene recognition module 4554 is further configured to acquire, from the scene recognition information, a wireless connection signal received by the terminal; determine a to-be-recognized signal name of the wireless connection signal; and in response to the to-be-recognized signal name being the same as a target signal name corresponding to a beacon signal, determine that the terminal is currently located in a first type of scene as the scene recognition result of the terminal, wherein the beacon signal is a signal transmitted by a wireless connection beacon pre-set in the first type of scene.

[0204] In some embodiments, the scene recognition module 4554 is further configured to acquire, from the scene recognition information, a background image collected by the terminal in a current scene in which the terminal is located; determine, by using a pre-trained scene detection model, probabilities that image content in the background image belongs to different preset scenes; and determine a scene recognition result of the terminal based on the probabilities that the image content in the background image belongs to the different preset scenes, wherein the scene recognition result is a preset scene corresponding to a maximum probability.

[0205] In some embodiments, the payment module 4555 is further configured to adjust a payment risk control level of the terminal based on the scene recognition result; and perform a payment process corresponding to the payment request based on the adjusted payment risk control level and identity information of the target object, to obtain a payment result.

[0206] In some embodiments, the scene recognition result includes a first scene recognition result and a second scene recognition result, and scene recognition information corresponding to the first scene recognition result is different from scene recognition information corresponding to the second scene recognition result; the payment module 4555 is further configured to, in response to the first scene recognition result and the second scene recognition result both indicating that the terminal is currently located in a first type of scene, adjust a payment risk control level of the terminal from a first level to a second level; in response to any one of the first scene recognition result and the second scene recognition result indicating that the terminal is currently located in the first type of scene, adjust the payment risk control level from the first level to a third level; wherein the first level, the second level, and the third level correspond to different payment verification manners respectively; in response to the first scene recognition result and the second scene recognition result both indicating that the terminal is currently located in a second type of scene, maintain the payment risk control level as the first level; and the second type of scene is any scene other than the first type of scene.

[0207] In some embodiments, a flow complexity of a first payment verification manner corresponding to the first level is higher than a flow complexity of a second payment verification manner corresponding to the second level; a flow complexity of a third payment verification manner corresponding to the third level is lower than the flow complexity of the first payment verification manner, and the flow complexity of the third payment verification manner is higher than the flow complexity of the second payment verification manner.

[0208] In some embodiments, the payment module 4555 is further configured to, in response to the adjusted payment risk control level being the first level, obtain reservation verification information of the target object; send a verification request to the terminal, wherein the terminal, after receiving the verification request, displays a verification interface in response to the verification request; receive to-be-verified information input by the terminal through the verification interface; and in response to the to-be-verified information being the same as the reservation verification information, perform payment processing on a payment account corresponding to the identity information of the target object to obtain a payment result.

[0209] In some embodiments, the payment module 4555 is further configured to, in response to the adjusted payment risk control level being the second level, perform payment processing on a payment account corresponding to the identity information of the target object to obtain the payment result.

[0210] In some embodiments, the payment module 4555 is further configured to, in response to the adjusted payment risk control level being the third level, obtain a similarity between an image feature of each object in the full-amount feature library and a to-be-identified image feature of the biometric information; in response to there being at least two target image features, obtain reservation verification information of the target object; the target image feature is an image feature having a similarity greater than a preset similarity threshold with the to-be-identified image feature; send a verification request to the terminal, wherein the terminal, after receiving the verification request, displays a verification interface in response to the verification request; receive to-be-verified information input by the terminal through the verification interface; and in response to the to-be-verified information being the same as the reservation verification information, perform payment processing on a payment account corresponding to the identity information of the target object to obtain a payment result.

[0211] In some embodiments, the payment device 455 further includes a feature library construction module configured to, before determining the area feature library of the region where the terminal is currently located based on the geographic location information, obtain position data of each object in a preset object set when the object makes a payment in a preset historical time period; determine a target region of the object from a plurality of position data of the object based on a number of times the object makes a payment at each position data; in response to there being an area feature library of the target region, add the image feature of the object to the area feature library of the target region; and in response to there being no area feature library of the target region, construct the area feature library of the target region and add the image feature of the object to the area feature library of the target region.

[0212] The embodiment of the present application provides a computer program product, which comprises a computer program or computer executable instructions stored in a computer readable storage medium. The processor of the electronic device reads the computer executable instructions from the computer readable storage medium, and the processor executes the computer executable instructions, so that the electronic device executes the payment method provided by the embodiment of the present application.

[0213] The embodiment of the present application provides a computer readable storage medium, which stores computer executable instructions or computer programs. When the computer executable instructions or computer programs are executed by the processor, the processor executes the payment method provided by the embodiment of the present application, for example, the payment method shown in FIG. 3.

[0214] In some embodiments, the computer readable storage medium can be RAM, ROM, flash memory, magnetic surface memory, optical disc or compact disc read-only memory (CD-ROM) and the like, and can also be various devices comprising one or any combination of the above storage.

[0215] In some embodiments, the computer executable instructions can be in the form of programs, software, software modules, scripts or codes, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and can be deployed in any form, including being deployed as independent programs or being deployed as modules, components, subroutines or other units suitable for use in a computing environment.

[0216] As an example, the computer executable instructions can but not necessarily correspond to files in a file system, can be stored in a part of a file storing other programs or data, for example, stored in one or more scripts in a hyper text markup language (HTML) document, stored in a single file dedicated to the program in question, or stored in multiple cooperative files (for example, files storing one or more modules, subroutines or code parts).

[0217] As an example, the computer executable instructions can be deployed to execute on one electronic device, or on multiple electronic devices located in one place, or on multiple electronic devices distributed in multiple places and interconnected through a communication network.

[0218] In summary, through the embodiment of the present application, the accuracy and security of palm payment can be improved, the intelligent degree of the payment system can be improved, and the risk control strategy can be optimized, so that the false brushing rate is reduced and the user experience is improved.

[0219] The above merely provides an example of the present application, but is not intended to limit the protection scope of the present application. Any modification, equivalent replacement, and improvement made within the spirit and scope of the present application shall be included in the protection scope of the present application.

Claims

1. A payment method, the method being performed by an electronic device, the method comprising: obtaining geographical position information of a location where a terminal is currently located, in response to a payment request sent by the terminal; the payment request comprising biological feature identification information of a target object; determining a region feature library of a region where the terminal is currently located, based on the geographical position information; performing matching on the biological feature identification information through a full-amount feature library, to obtain identity information of the target object, in response to a failure of the biological feature identification information in matching in the region feature library; the full-amount feature library comprising region feature libraries of multiple regions; obtaining scene identification information of the terminal, and determining a scene identification result of the terminal based on the scene identification information; performing payment processing based on the scene identification result and the identity information of the target object.

2. The method of claim 1, wherein, the biological feature identification information comprises a biological feature identification image; the performing matching on the biological feature identification information through the full-amount feature library, to obtain the identity information of the target object, comprises: performing feature extraction on the biological feature identification image, to obtain image features to be identified; determining similarity between the image features to be identified and image features of each object in the full-amount feature library; determining identity information of an object corresponding to image features with maximum similarity as the identity information of the target object.

3. The method of claim 1 or 2, wherein, the obtaining the scene identification information of the terminal, and determining the scene identification result of the terminal based on the scene identification information, comprises: obtaining multiple point of interest information from the scene identification information; the point of interest information comprising scene information within a preset range where the terminal is currently located; determining the scene identification result of the terminal as the terminal being currently located in a first type of scene, in response to searching for a first keyword in at least one point of interest information.

4. The method according to any one of claims 1 to 3, wherein, the obtaining the scene identification information of the terminal, and determining the scene identification result of the terminal based on the scene identification information, comprises: obtaining network information of a network currently accessed by the terminal from the scene identification information; determining a network name and signal strength of the network currently accessed from the network information; determining the scene identification result of the terminal based on the network name and the signal strength.

5. The method according to any one of claims 1 to 4, wherein, the determining the scene identification result of the terminal based on the network name and the signal strength, comprises: determining the scene identification result of the terminal as the terminal being currently located in a first type of scene, in response to searching for a preset second keyword from the network name and the signal strength being greater than a preset strength threshold.

6. The method according to any one of claims 1 to 5, wherein, the obtaining the scene identification information of the terminal, and determining the scene identification result of the terminal based on the scene identification information, comprises: obtaining a wireless connection signal received by the terminal from the scene identification information; determining a signal name to be identified of the wireless connection signal; determining the scene identification result of the terminal as the terminal being currently located in a first type of scene, in response to the signal name to be identified being identical to a target signal name corresponding to a beacon signal; the beacon signal being a signal sent by a wireless connection beacon preset in the first type of scene.

7. The method according to any one of claims 1 to 6, wherein, The acquiring the scene recognition information of the terminal, determining a scene recognition result of the terminal based on the scene recognition information, comprises: acquiring a background image collected by the terminal in a current scene from the scene recognition information; determining a probability that image content in the background image belongs to different preset scenes by using a pre-trained scene detection model; determining the scene recognition result of the terminal based on the probability that the image content in the background image belongs to different preset scenes; the scene recognition result is that the terminal is located in a preset scene corresponding to the maximum probability.

8. The method according to any one of claims 1 to 7, wherein, The payment processing based on the scene recognition result and the identity information of the target object comprises: adjusting the payment risk control level of the terminal based on the scene recognition result; performing payment processing corresponding to the payment request based on the adjusted payment risk control level and the identity information of the target object to obtain a payment result.

9. The method according to any one of claims 1 to 8, wherein, The scene recognition result comprises a first scene recognition result and a second scene recognition result, and the scene recognition information corresponding to the first scene recognition result is different from the scene recognition information corresponding to the second scene recognition result. The adjusting the payment risk control level of the terminal based on the scene recognition result comprises: in response to the first scene recognition result and the second scene recognition result both being that the terminal is currently located in a first type of scene, adjusting the payment risk control level of the terminal from a first level to a second level; in response to any one of the first scene recognition result and the second scene recognition result being that the terminal is currently located in the first type of scene, adjusting the payment risk control level from the first level to a third level; wherein the first level, the second level and the third level correspond to different payment verification methods respectively; in response to the first scene recognition result and the second scene recognition result both being that the terminal is currently located in a second type of scene, maintaining the payment risk control level as the first level; the second type of scene is any scene other than the first type of scene.

10. The method according to any one of claims 1 to 9, wherein, The flow complexity of a first payment verification method corresponding to the first level is higher than the flow complexity of a second payment verification method corresponding to the second level; The flow complexity of a third payment verification method corresponding to the third level is lower than the flow complexity of the first payment verification method, and the flow complexity of the third payment verification method is higher than the flow complexity of the second payment verification method.

11. The method according to any one of claims 1 to 10, wherein, The payment processing based on the adjusted payment risk control level and the identity information of the target object to obtain a payment result comprises: in response to the adjusted payment risk control level being the first level, acquiring reserved verification information of the target object; sending a verification request to the terminal, wherein the terminal displays a verification interface in response to the verification request after receiving the verification request; receiving the to-be-verified information input through the verification interface and sent by the terminal; In response to the to-be-verified information being the same as the reservation verification information, performing payment processing on a payment account corresponding to the identity information of the target object to obtain a payment result.

12. The method according to any one of claims 1 to 11, wherein, The payment processing corresponding to the payment request is performed based on the adjusted payment risk control level and the identity information of the target object to obtain a payment result, including: In response to the adjusted payment risk control level being the second level, performing payment processing on a payment account corresponding to the identity information of the target object to obtain the payment result.

13. The method according to any one of claims 1 to 12, wherein, The payment processing corresponding to the payment request is performed based on the adjusted payment risk control level and the identity information of the target object to obtain a payment result, including: In response to the adjusted payment risk control level being the third level, obtaining a similarity between an image feature of each object in the full-amount feature library and a to-be-identified image feature of the biometric information; In response to there being at least two target image features, obtaining reservation verification information of the target object; the target image feature is an image feature with a similarity greater than a preset similarity threshold to the to-be-identified image feature; Sending a verification request to the terminal, wherein the terminal displays a verification interface in response to the verification request after receiving the verification request; Accepting to-be-verified information input by the terminal through the verification interface; In response to the to-be-verified information being the same as the reservation verification information, performing payment processing on a payment account corresponding to the identity information of the target object to obtain a payment result.

14. The method according to any one of claims 1 to 13, wherein, Before determining the regional feature library of the region where the terminal is currently located based on the geographic location information, the method further includes: Obtaining location data of each object in a preset object set when the object makes a payment in a preset historical time period; Determining a target region of the object from the multiple location data of the object based on the number of times the object makes a payment at each location data; In response to there being a regional feature library of the target region, adding the image feature of the object to the regional feature library of the target region; In response to there being no regional feature library of the target region, constructing a regional feature library of the target region and adding the image feature of the object to the regional feature library of the target region.

15. A payment device, the device comprising: A location acquisition module configured to acquire geographic location information of a location where a terminal is currently located in response to a payment request sent by the terminal; The payment request includes biometric information of a target object; The feature library acquisition module is configured to determine a regional feature library of a region where the terminal is currently located based on the geographic location information; The matching module is configured to, in response to the biometric information failing to match in the regional feature library, match the biometric information through a full-amount feature library to obtain identity information of the target object; the full-amount feature library includes regional feature libraries of multiple regions; The scene recognition module is configured to acquire scene recognition information of the terminal and determine a scene recognition result of the terminal based on the scene recognition information; The payment module is configured to perform payment processing based on the scene recognition result and the identity information of the target object.

16. An electronic device, comprising: a memory configured to store computer-executable instructions or computer programs; a processor configured to implement the payment method of any one of claims 1 to 14 when executing the computer-executable instructions or computer programs stored in the memory.

17. A computer-readable storage medium storing computer-executable instructions or computer programs, which, when executed by a processor, implement the payment method of any one of claims 1 to 14.

18. A computer program product comprising computer-executable instructions or computer programs, which, when executed by a processor, implement the payment method of any one of claims 1 to 14.

Images