Desktop cash register with fingerprint identification and payment verification and verification method

Abstract

The invention discloses a desktop cash register with fingerprint identification and payment verification and a verification method, and relates to the technical field of fingerprint identification, and the technical scheme is characterized in that the method comprises the following steps: obtaining fingerprint biological characteristic dynamic data and payment behavior associated data of a payment user, performing feature disassembly and recombination on the fingerprint biological feature dynamic data and the payment behavior associated data to form a multi-level verification feature system; calculating a reliability coefficient and association tightness of each level feature in the multi-level verification feature system to obtain a feature verification index; mining feature verification evolution rules in different payment scenes from historical verification data stored in a distributed manner; the method has the advantages that a serial process of data acquisition and single verification in information verification is avoided, parallel processing of verification links is achieved, total time consumption of settlement is further shortened, a basic verification process can be completed through fingerprint biological characteristics, dependence on a mobile phone terminal can be thoroughly eliminated, and transaction continuity is guaranteed.

Description

Technical Field

[0001] This invention relates to the field of fingerprint recognition technology, and more specifically, to a desktop cash register and verification method with fingerprint recognition and payment verification capabilities. Background Technology

[0002] With the booming development of the offline retail industry, the efficiency and security of point-of-sale (POS) settlement, as a core link in the transaction loop, directly affect the operational efficiency of merchants and the consumer experience of users. Currently, desktop POS machines have gradually evolved from traditional single-function settlement tools into integrated terminals that combine product scanning, billing processing, and payment verification. However, many problems still urgently need to be addressed in the payment verification process.

[0003] Traditional cash registers primarily rely on PIN entry and card swipes for payment verification, but these methods have significant limitations. PIN payments depend on user memory, making them susceptible to forgetting or leaking information, and the input process is time-consuming, potentially causing congestion at the checkout during peak hours. Card swipes, on the other hand, are vulnerable to card loss and fraud, and magnetic stripe card information is easily copied, resulting in weak security. Even with the introduction of newer methods like QR code payments, these still require a mobile phone to generate a payment code. If the user's phone is out of battery, has a network interruption, or is unfamiliar with the operation, transaction delays can still occur, impacting settlement efficiency. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the purpose of this invention is to provide a desktop cash register and verification method with fingerprint recognition and payment verification.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A verification method with fingerprint recognition and payment verification, the method comprising the following steps:

[0007] Obtain dynamic fingerprint biometric data and payment behavior-related data of payment users, and decompose and reassemble the dynamic fingerprint biometric data and payment behavior-related data to form a multi-level verification feature system;

[0008] The reliability coefficients and correlation tightness of features at each level in the multi-level verification feature system are calculated to obtain the feature verification index;

[0009] We mine the evolution patterns of feature verification under different payment scenarios from historical verification data in distributed storage, and construct an adaptive verification evaluation model based on the evolution patterns of feature verification. We input the feature verification indicators into the adaptive verification evaluation model to obtain the main verification qualification index and the auxiliary verification qualification index of payment users.

[0010] The overall qualification assessment value for payment verification is determined by calculating the dynamic equilibrium value of the primary verification qualification index and the secondary verification qualification index.

[0011] Based on the comprehensive qualification assessment value and the preset verification grading response rules, payment verification results and corresponding security control schemes are generated.

[0012] Preferably, the fingerprint biometric dynamic data includes dynamic deformation feature data during the fingerprint acquisition process, fingerprint ridge temporal change feature data, and fingerprint pressure distribution dynamic data; the payment behavior associated data includes payment account associated device data, payment habit preference data, and recent transaction flow feature data.

[0013] Preferably, the fingerprint biometric dynamic data and payment behavior-related data are decomposed and recombined to form a multi-level verification feature system, specifically including the following steps:

[0014] The basic features of fingerprint deformation are obtained by decomposing the deformation trajectory and deformation amplitude features of the dynamic deformation feature data of fingerprints.

[0015] Fingerprint temporal features are obtained by decomposing fingerprint ridge temporal variation feature data into ridge continuity and ridge variation features;

[0016] The dynamic features of fingerprint pressure are obtained by decomposing the dynamic data of fingerprint pressure distribution into pressure gradient and pressure concentration region features.

[0017] The device association characteristics of payment accounts are obtained by performing device type matching and device login frequency feature decomposition on the data of devices associated with payment accounts.

[0018] Payment habit characteristics are obtained by decomposing payment habit preference data into payment time period preference and payment amount range characteristics;

[0019] Transaction flow characteristics are obtained by decomposing recent transaction flow characteristic data into transaction frequency and transaction object association characteristics;

[0020] Based on the rules for classifying features according to their importance, the basic features of fingerprint deformation, the temporal features of fingerprint patterns, the dynamic features of fingerprint pressure, the features associated with accounts and devices, the features of payment habits, and the features of transaction flow are reorganized into the core verification layer, the auxiliary verification layer, and the supplementary verification layer, forming a multi-level verification feature system.

[0021] Preferably, the feature verification index is obtained by calculating the reliability coefficient and correlation tightness of features at each level in the multi-level verification feature system, specifically including the following steps:

[0022] The reliability coefficients of each level of features in the multi-level verification feature system are calculated to obtain the feature coefficient values;

[0023] The degree of feature correlation is obtained by judging the closeness of the correlation between features at different levels and between features within the same level.

[0024] The feature verification index is obtained by fusing the feature coefficient values ​​with the corresponding feature correlation.

[0025] Preferably, the evolutionary patterns of feature verification under different payment scenarios are mined from historical verification data in distributed storage, and an adaptive verification evaluation model is constructed based on these patterns. This specifically includes the following steps:

[0026] Extract valid verification samples from historical verification data stored across multiple nodes;

[0027] By determining the correlation and evolution trend between the changes of features over time in valid validation samples and the validation results, the evolution law of feature validation is obtained.

[0028] An initial adaptive verification and evaluation model is constructed based on the evolutionary rules of feature verification.

[0029] Preferably, the feature verification indicators are input into the adaptive verification evaluation model to obtain the primary verification qualification index and the secondary verification qualification index for the payment user, specifically including the following steps:

[0030] The feature verification indicators corresponding to the core verification layer in the multi-level verification feature system are determined as the main verification input parameters and input into the core evaluation unit of the adaptive verification evaluation model. The main verification qualification index, which reflects the qualification level of the core verification dimension, is judged and output through the core evaluation unit.

[0031] The feature verification indicators corresponding to the auxiliary verification layer and the supplementary verification layer are determined as auxiliary verification input parameters and input into the auxiliary evaluation unit of the adaptive verification evaluation model. Through the comprehensive judgment of the auxiliary evaluation unit, the auxiliary verification qualification index, which reflects the qualification level of the auxiliary verification dimension, is output.

[0032] Preferably, the comprehensive qualification assessment value for payment verification is determined by calculating the dynamic equilibrium value of the primary verification qualification index and the secondary verification qualification index, specifically including the following steps:

[0033] Based on the current security requirements of the payment scenario, dynamically configure the primary weight coefficient of the primary verification qualification index and the secondary weight coefficient of the secondary verification qualification index.

[0034] The degree of synergy matching is obtained by calculating the balance between the primary verification pass index and the secondary verification pass index.

[0035] The overall qualification rating of payment verification is obtained by integrating the primary verification qualification index with the primary weight coefficient, the secondary verification qualification index with the secondary weight coefficient, and combining them with the collaborative matching degree value.

[0036] Preferably, the payment verification result and corresponding security control plan are generated based on the comprehensive qualification assessment value and the preset verification grading response rules, specifically including the following steps:

[0037] If the overall qualification assessment value is higher than or equal to the excellent grade pass threshold, the excellent grade pass result is generated and the fast payment channel is activated to complete the settlement.

[0038] If the overall pass rating is between or equal to the excellent pass threshold and the normal pass threshold, a normal pass result is generated, and settlement is completed according to the regular payment process.

[0039] If the overall pass rating is between or equal to the verification threshold, the verification result will be generated and a prompt for supplementary account identity verification will be output.

[0040] If the overall qualification assessment value is between or equal to the verification threshold and the risk threshold, a verification restricted result is generated, the payment amount is limited, and verification abnormality information is recorded.

[0041] If the overall qualification assessment value is lower than the risk threshold, a verification failure result will be generated, and an emergency security control plan will be activated, including freezing the payment account and pushing risk warning information to users and merchants.

[0042] Preferably, it also includes dynamic feature calibration, specifically comprising the following steps:

[0043] Real-time monitoring of the stability of core verification layer features in a multi-level verification feature system;

[0044] If the stability of the core feature is lower than the preset stability threshold, the feature calibration mechanism will be activated, and the user's historical core feature data will be retrieved as the calibration benchmark.

[0045] By comparing features, the current core verification features are dynamically calibrated until the feature stability meets the verification requirements.

[0046] A desktop cash register with fingerprint recognition and payment verification includes:

[0047] Acquisition Module: Acquires dynamic fingerprint biometric data and payment behavior-related data of payment users, and decomposes and reassembles the dynamic fingerprint biometric data and payment behavior-related data to form a multi-level verification feature system;

[0048] Verification module: Calculates the reliability coefficient and correlation tightness of features at each level in the multi-level verification feature system to obtain feature verification indicators;

[0049] Mining Module: Mines feature verification evolution patterns under different payment scenarios from historical verification data in distributed storage, constructs an adaptive verification evaluation model based on feature verification evolution patterns, and inputs feature verification indicators into the adaptive verification evaluation model to obtain the main verification qualification index and auxiliary verification qualification index of payment users;

[0050] Calculation module: Determines the overall qualification assessment value for payment verification by calculating the dynamic equilibrium value of the primary verification qualification index and the secondary verification qualification index;

[0051] Control module: Generates payment verification results and corresponding security control schemes based on the comprehensive qualification assessment value and the preset verification grading response rules.

[0052] Compared with existing technologies, this invention has the following advantages: The process of acquiring dynamic fingerprint biometric data eliminates the need for users to remember passwords or operate mobile phones; core data acquisition can be completed simply by lightly touching the fingerprint acquisition module of the cash register. This reduces the time required for a single verification operation, improves payment efficiency, and effectively alleviates queuing pressure at cash registers during peak hours. Furthermore, the collection and processing of payment behavior-related data can be performed simultaneously with fingerprint recognition. The multi-level verification system formed through feature decomposition and recombination avoids the serial process of data collection and single verification in information verification, enabling parallel processing of the verification process and further shortening the total settlement time. The basic verification process can be completed using fingerprint biometrics, completely eliminating reliance on mobile terminals and ensuring transaction continuity. Attached Figure Description

[0053] Figure 1 This is a schematic diagram illustrating the steps of a verification method with fingerprint recognition and payment verification proposed in this invention;

[0054] Figure 2 This is a schematic diagram illustrating the steps in forming a multi-level verification feature system in the verification method with fingerprint recognition and payment verification proposed in this invention;

[0055] Figure 3 This invention presents a schematic diagram of a desktop cash register with fingerprint recognition and payment verification. Detailed Implementation

[0056] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0057] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0058] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single embodiment or an embodiment selectively excluded from other embodiments.

[0059] Reference Figures 1-3 As shown.

[0060] Example 1 further illustrates the desktop cash register and verification method with fingerprint recognition and payment verification proposed in this invention.

[0061] A verification method with fingerprint recognition and payment verification, the method comprising the following steps:

[0062] Obtain dynamic fingerprint biometric data and payment behavior-related data of payment users, and decompose and reassemble the dynamic fingerprint biometric data and payment behavior-related data to form a multi-level verification feature system;

[0063] The reliability coefficients and correlation tightness of features at each level in the multi-level verification feature system are calculated to obtain the feature verification index;

[0064] We mine the evolution patterns of feature verification under different payment scenarios from historical verification data in distributed storage, and construct an adaptive verification evaluation model based on the evolution patterns of feature verification. We input the feature verification indicators into the adaptive verification evaluation model to obtain the main verification qualification index and the auxiliary verification qualification index of payment users.

[0065] The overall qualification assessment value for payment verification is determined by calculating the dynamic equilibrium value of the primary verification qualification index and the secondary verification qualification index.

[0066] Based on the comprehensive qualification assessment value and the preset verification grading response rules, payment verification results and corresponding security control schemes are generated.

[0067] The fingerprint biometric dynamic data includes dynamic deformation feature data during fingerprint acquisition, fingerprint ridge temporal change feature data, and fingerprint pressure distribution dynamic data. The payment behavior associated data includes payment account associated device data, payment habit preference data, and recent transaction flow feature data.

[0068] During fingerprint acquisition, dynamic deformation feature data is recorded simultaneously. For example, during the process of pressing the finger against the acquisition device, the fingerprint edge contour undergoes subtle stretching or contraction due to pressure changes. This morphological data is unique to each individual, and differences in finger elasticity and pressing habits among different users will result in significant differentiation in deformation features. Simultaneously, fingerprint ridge temporal change feature data is also collected, i.e., the order in which the ridges appear from partial to complete during the fingerprint contact with the acquisition device, and the time interval between the appearance of ridges in different areas. For example, for some users, the core ridges of their fingerprints are recognized before the edge ridges, while for others, the edge ridges appear first. This temporal difference is a stable individual behavioral characteristic. Furthermore, dynamic fingerprint pressure distribution data is collected, showing how the pressure values ​​in different areas of the finger change over time during acquisition, such as the timing of pressure peaks in the fingertip and base areas, and the fluctuation range of pressure values. This data further supplements the static features of fingerprints with dynamic and unique information.

[0069] Payment account associated device data includes the matching between the device currently initiating the payment and the account's historically frequently used devices, such as device model, system version, and hardware identification information. If the current device is a frequently used device that the account has been bound to for a long time, the reliability of this feature will be correspondingly increased. Payment habit preference data covers the user's daily payment time and transaction type preferences. For example, if a user is accustomed to making small-amount daily consumption payments on weekday evenings, this habit will form a fixed behavioral characteristic. Recent transaction flow characteristic data refers to the transaction amount range, transaction frequency, and transaction object type information within the past week or month. For example, if a user has recently made many transactions with offline merchants under 100 yuan, this data can reflect the user's recent transaction patterns.

[0070] The process of decomposing and recombining dynamic fingerprint biometric data with payment behavior-related data to form a multi-level verification feature system includes the following steps:

[0071] The basic features of fingerprint deformation are obtained by decomposing the deformation trajectory and deformation amplitude features of the dynamic deformation feature data of fingerprints.

[0072] Fingerprint temporal features are obtained by decomposing fingerprint ridge temporal variation feature data into ridge continuity and ridge variation features;

[0073] The dynamic features of fingerprint pressure are obtained by decomposing the dynamic data of fingerprint pressure distribution into pressure gradient and pressure concentration region features.

[0074] The device association characteristics of payment accounts are obtained by performing device type matching and device login frequency feature decomposition on the data of devices associated with payment accounts.

[0075] Payment habit characteristics are obtained by decomposing payment habit preference data into payment time period preference and payment amount range characteristics;

[0076] Transaction flow characteristics are obtained by decomposing recent transaction flow characteristic data into transaction frequency and transaction object association characteristics;

[0077] Based on the rules for classifying features according to their importance, the basic features of fingerprint deformation, the temporal features of fingerprint patterns, the dynamic features of fingerprint pressure, the features associated with accounts and devices, the features of payment habits, and the features of transaction flow are reorganized into the core verification layer, the auxiliary verification layer, and the supplementary verification layer, forming a multi-level verification feature system.

[0078] The dynamic data of fingerprint biometrics is broken down into fine details. For the dynamic deformation feature data of fingerprints, the deformation trajectory and deformation amplitude are extracted. The deformation trajectory refers to the path of the fingerprint contour change when the finger is pressed on the acquisition device. For example, when the user presses, the fingerprint will deform along the upper left to lower right direction of the acquisition area, reflecting the finger's force habits. The deformation amplitude is the maximum stretching or contraction of the fingerprint contour during the deformation process. For example, the deformation amplitude of the fingerprint edge of the same user is basically stable in the range of 2% to 3% each time it is pressed. These sub-features together constitute the basic features of fingerprint deformation.

[0079] Fingerprint ridge temporal variation feature data is decomposed into ridge continuity and ridge variation features. Ridge continuity refers to the connection time and smoothness between each ridge segment during the process of fingerprint ridge acquisition and complete presentation. For example, some users' fingerprint ridges can be presented in a continuous temporal sequence without interruption, while the ridge presentation process of other users will have short pauses. Ridge variation features refer to the temporary changes in the ridge morphology during the acquisition process. For example, when the pressure fluctuates, some ridges may become temporarily blurred or deformed. These sub-features are integrated into fingerprint ridge temporal features.

[0080] The dynamic data of fingerprint pressure distribution is decomposed into pressure gradient and pressure concentration region features. The pressure gradient refers to the rate at which the pressure value in different areas of the fingerprint changes over time, such as the rate at which the pressure in the fingertip area rises from the initial value to the peak value. The difference in the force application rhythm of different users will make this rate show obvious differences. The pressure concentration region refers to the fingerprint area with the highest pressure value during the acquisition process. For example, the pressure of some users is concentrated in the core ridge area of ​​the fingerprint, while that of other users is concentrated in the edge area. These sub-features constitute the dynamic features of fingerprint pressure.

[0081] The payment account associated device data is broken down into device type matching and device login frequency features. Device type matching refers to the degree of overlap between the current payment device model, system information and the account's historically frequently used devices. For example, if the current device is a mobile phone that the account has been using continuously for the past six months, the matching degree is relatively high. Device login frequency refers to the number of times the device logs into the account in the past month. The more times, the stronger the stability of this feature. These sub-features constitute the account device association features.

[0082] Payment habit preference data is broken down into payment time preference and payment amount range characteristics. Payment time preference refers to the time period when users usually initiate payments, such as users mostly making payments between 6 pm and 8 pm every day. Payment amount range refers to the range of amounts in users' daily transactions, such as most transactions being concentrated between 50 yuan and 200 yuan. These sub-features are integrated into payment habit characteristics.

[0083] The recent transaction flow data was broken down into transaction frequency and transaction object association features. Transaction frequency refers to the number of transactions a user has made in the past week, such as an average of 2 to 3 transactions per day. Transaction object association features refer to the merchant type and counterparty account type information of recent transactions, such as transactions with offline catering merchants. These sub-features constitute the transaction flow features.

[0084] After all features are decomposed, they are reorganized according to the feature importance hierarchy. The basic fingerprint deformation features, fingerprint ridge temporal features, and fingerprint pressure dynamic features are assigned to the core verification layer. Since the fingerprint dynamic features have strong uniqueness, they serve as the core basis for verification. The account device association features and payment habit features are assigned to the auxiliary verification layer to supplement the verification dimensions of the core features. The transaction flow features are assigned to the supplementary verification layer to further refine the accuracy of verification. Finally, a verification feature system containing three layers—core, auxiliary, and supplementary—is formed, providing a structured feature foundation for subsequent reliability coefficient calculation and verification evaluation.

[0085] The reliability coefficients and correlation strengths of features at each level in a multi-level verification feature system are calculated to obtain feature verification indices. This process includes the following steps:

[0086] The reliability coefficients of each level of features in the multi-level verification feature system are calculated to obtain the feature coefficient values;

[0087] The degree of feature correlation is obtained by judging the closeness of the correlation between features at different levels and between features within the same level.

[0088] The feature verification index is obtained by fusing the feature coefficient values ​​with the corresponding feature correlation.

[0089] The reliability coefficients of each level of features are calculated, and the effectiveness of each decomposed sub-feature is evaluated to obtain the corresponding feature coefficient value. Taking the fingerprint deformation basic feature of the core verification layer as an example, the currently collected deformation trajectory is compared with the average data of the user's historical deformation trajectory, and the overlap between the two is calculated. If the overlap between the current deformation trajectory and the historical data reaches 95%, the reliability coefficient of the sub-feature can be set to 0.95. For the fingerprint ridge temporal feature, the deviation between the time interval of the current ridge continuity and the historical average interval is compared. If the deviation is within 5%, its reliability coefficient is 0.9. For the fingerprint pressure dynamic feature, the fluctuation amplitude of the current pressure gradient and the historical pressure gradient is compared. If the fluctuation amplitude is within a preset threshold, the reliability coefficient is 0.85. For the account device association feature of the auxiliary verification layer, if the current device is a frequently used device of the account and the login frequency in the past month is more than 20 times, its reliability coefficient is 0.8. In the payment habit feature, if the current payment time period and amount range both match historical habits, the reliability coefficient is 0.75. The transaction flow characteristics of the supplementary verification layer are evaluated. If the current transaction frequency and object type both conform to recent patterns, the reliability coefficient is 0.7. After the reliability coefficient of each sub-feature is calculated, they are summarized according to their respective levels to obtain the feature coefficient values ​​of each level. For example, the comprehensive feature coefficient value of the core verification layer can be obtained by weighted averaging of the reliability coefficients of each sub-feature. If the weight of the fingerprint deformation basic feature is 0.4, the weight of the ridge temporal feature is 0.3, and the weight of the pressure dynamic feature is 0.3, then its comprehensive feature coefficient value is 0.95×0.4+0.9×0.3+0.85×0.3=0.915.

[0090] The correlation between features is determined by assessing the degree of association between them, which in turn determines the matching degree between different features to reflect the consistency of the verification information. Looking at feature associations within the same layer, in the core verification layer, the correlation between the basic fingerprint deformation features and the dynamic fingerprint pressure features can be judged by whether the trends of their reliability coefficients are consistent. For example, when the deformation amplitude increases, do the values ​​in the pressure concentration area rise synchronously? If the consistency of their trends reaches 90%, then the correlation between these two features within the same layer is 0.9. In the auxiliary verification layer, the correlation between account device association features and payment habit features can be assessed by observing whether the payment habits corresponding to commonly used devices are stable. If the payment time deviation under this device is consistently within 10%, the correlation is 0.8. From the perspective of the correlation between features at different levels, the fingerprint pressure dynamic features of the core verification layer and the payment habit features of the auxiliary verification layer can determine whether the fingerprint pressure distribution is consistent with historical data of the same scenario in the transaction scenario corresponding to the payment habit. For example, in daily small-amount consumption scenarios, the concentrated area of ​​fingerprint pressure is mostly on the fingertips. If the current data conforms to this pattern, the correlation between the two is 0.75. The fingerprint temporal features of the core verification layer and the transaction flow features of the supplementary verification layer can be used to see whether the pattern of fingerprint temporal presentation is consistent with the history in recent high-frequency transaction periods. If the consistency reaches 80%, the correlation is 0.8.

[0091] The feature verification index is obtained by fusing feature coefficient values ​​and feature correlation. The coefficient value of each feature is weighted and integrated with its correlation. For example, for the basic fingerprint deformation feature of the core verification layer, its feature coefficient value is 0.95, the correlation of the fingerprint pressure dynamic feature of the same level is 0.9, and the correlation of the account device correlation feature of different levels is 0.8. Then the fusion value of this feature is 0.95×0.6+0.9×0.2+0.8×0.2=0.92. Then the fusion values ​​of each feature are summarized according to the level weight: the core verification layer has a weight of 0.5, the auxiliary verification layer has a weight of 0.3, and the supplementary verification layer has a weight of 0.2. If the average value of the core layer after fusion is 0.9, the auxiliary layer is 0.8, and the supplementary layer is 0.7, then the final feature verification index is 0.9×0.5+0.8×0.3+0.7×0.2=0.83. This value will be used as the core basis for subsequent verification and evaluation.

[0092] The evolutionary patterns of feature verification under different payment scenarios are mined from historical verification data in distributed storage. An adaptive verification evaluation model is then constructed based on these patterns. The specific steps include:

[0093] Extract valid verification samples from historical verification data stored across multiple nodes;

[0094] By determining the correlation and evolution trend between the changes of features over time in valid validation samples and the validation results, the evolution law of feature validation is obtained.

[0095] An initial adaptive verification and evaluation model is constructed based on the evolutionary rules of feature verification.

[0096] Valid verification samples are extracted from historical verification data stored across multiple nodes. This distributed historical data is scattered across multiple nodes, each recording verification information from different time periods and users. This data is first cleaned, and valid samples that meet certain criteria are selected. For example, verification records with missing data are removed, and records containing complete fingerprint dynamic features, payment behavior features, and the final verification result are retained. Taking offline small-amount consumption scenarios as an example, all verification data from the past six months for this scenario is extracted from the corresponding nodes. This includes the fingerprint deformation trajectory, pressure distribution, device information, payment time characteristics for each verification, and the results of verification success or rejection. Complete records constitute valid verification samples.

[0097] The evolutionary pattern of feature verification is obtained by judging the correlation between the changes of features over time in valid verification samples and the evolutionary trend of verification results. This pattern is then analyzed for different payment scenarios. For example, in offline small-amount consumption scenarios, the dynamic changes in fingerprint pressure distribution data are statistically analyzed. It is found that in the past three months, the fluctuation range of the concentrated area of ​​user fingerprint pressure in this scenario has decreased from 10% to 5%. Furthermore, when the fluctuation range is less than 5%, the probability of successful verification increases from 80% to 95%. The correlation between feature changes and verification results represents the evolutionary pattern of fingerprint pressure features in this scenario. If, in this scenario, the user's payment time period gradually expands from 6 PM to 8 PM to 5 PM to 9 PM, and if the payment amount remains within the original habitual range when initiating payment during the expanded time period, the probability of successful verification remains above 90%, then this represents the evolutionary pattern of payment habit features. For online large-amount transfer scenarios, we discovered the evolution trend of account device association features. In the past month, the login frequency threshold of frequently used devices has increased from 20 times / month to 25 times / month. When the device login frequency reaches more than 25 times, the matching degree of verification will increase by 10%. At the same time, the continuity deviation threshold of fingerprint ridge temporal features has tightened from 5% to 3%. When the deviation is less than 3%, the probability of verification is higher. The changes of these features over time and their correlation with the verification results together constitute the feature verification evolution law under different scenarios.

[0098] An initial adaptive verification evaluation model is constructed based on the evolutionary laws of feature verification. The evolutionary laws under different scenarios are transformed into training parameters for the model. For example, the law that the probability of successful verification increases when the fingerprint pressure fluctuation amplitude is ≤5% in offline small-amount consumption scenarios is transformed into a weight adjustment rule for this feature in the initial adaptive verification evaluation model. Similarly, the law that the weight of device-related features increases when the device login frequency is ≥25 times in online large-amount transfer scenarios is incorporated into the scenario adaptation logic of the initial adaptive verification evaluation model. During training, the initial adaptive verification evaluation model uses valid verification samples as input and the evolutionary laws of feature verification as constraints, allowing the model to learn the correspondence between feature combinations and verification results in different scenarios. For instance, in offline small-amount consumption scenarios, when the reliability coefficient of the basic fingerprint deformation feature is ≥0.9, the payment time is between 5 PM and 9 PM, and the payment amount is within the usual range, the model will output a higher verification pass index. Conversely, in online large-amount transfer scenarios, when the device login frequency is <25 times and the fingerprint pattern continuity deviation is >3%, the model will lower the verification pass index. Through training, the initial adaptive verification and evaluation model can dynamically adjust the evaluation logic of features according to the feature evolution patterns under different payment scenarios, thereby achieving accurate scenario-based verification.

[0099] The feature verification metrics are input into the adaptive verification evaluation model to obtain the primary verification qualification index and the secondary verification qualification index for payment users. This process includes the following steps:

[0100] The feature verification indicators corresponding to the core verification layer in the multi-level verification feature system are determined as the main verification input parameters and input into the core evaluation unit of the adaptive verification evaluation model. The main verification qualification index, which reflects the qualification level of the core verification dimension, is judged and output through the core evaluation unit.

[0101] The feature verification indicators corresponding to the auxiliary verification layer and the supplementary verification layer are determined as auxiliary verification input parameters and input into the auxiliary evaluation unit of the adaptive verification evaluation model. Through the comprehensive judgment of the auxiliary evaluation unit, the auxiliary verification qualification index, which reflects the qualification level of the auxiliary verification dimension, is output.

[0102] The main verification pass index is the feature verification metric of the core verification layer, calculated by the core evaluation unit of the adaptive verification evaluation model. The feature verification metrics corresponding to the core verification layer include verification metrics for basic fingerprint deformation features, temporal features of fingerprint ridges, and dynamic features of fingerprint pressure. Taking a user's offline small-amount payment as an example, the core verification input parameters are the verification metric values ​​for these features: for example, the verification metric for basic fingerprint deformation features is 0.95, indicating a high degree of matching between the current deformation trajectory and historical data; the verification metric for temporal features of fingerprint ridges is 0.9, indicating that the temporal pattern of the ridges is consistent with user habits; and the verification metric for dynamic fingerprint pressure features is 0.85, reflecting that the dynamic changes in pressure distribution conform to historical characteristics. The core evaluation unit will calculate these parameters by weighting them according to the evolution of feature verification in different payment scenarios. For example, in offline small-amount consumption scenarios, the weight of the basic fingerprint deformation feature is 0.4, the weight of the fingerprint ridge temporal feature is 0.3, and the weight of the fingerprint pressure dynamic feature is 0.3. Then the main verification qualification index is 0.95×0.4+0.9×0.3+0.85×0.3=0.915. This value directly reflects the qualification of the core verification dimension. The higher the value, the better the matching degree of the core biometric features.

[0103] The auxiliary verification pass index corresponds to the feature verification indicators of the auxiliary verification layer and the supplementary verification layer. It is comprehensively judged by the auxiliary evaluation unit of the adaptive verification evaluation model. The auxiliary verification input parameters cover the verification indicators of account device association features, payment habit features, and transaction flow features. Taking the user's offline small-amount consumption payment as an example, the account device association feature verification indicator of the auxiliary verification layer is 0.8, which means that the current device is the account's frequently used device and the login frequency meets the standard; the payment habit feature verification indicator is 0.75, which means that the current payment time and amount range are consistent with the user's daily preferences; and the transaction flow feature verification indicator of the supplementary verification layer is 0.7, which reflects the historical pattern of recent transaction frequency matching object type. The auxiliary evaluation unit will assign corresponding weights to these parameters based on the evolution of the scenario. For example, in offline small-amount consumption scenarios, the weight of account device association features is 0.4, the weight of payment habit features is 0.3, and the weight of transaction flow features is 0.3. The auxiliary verification qualification index is 0.8×0.4+0.75×0.3+0.7×0.3=0.765, which reflects the qualification level of auxiliary and supplementary verification dimensions. The higher the value, the stronger the consistency of payment behavior features.

[0104] By evaluating each unit separately, the adaptive validation assessment model can quantify the qualification level of core biometrics and auxiliary behavioral characteristics, providing a clear quantitative basis for the subsequent fusion of primary and auxiliary validation indices.

[0105] The overall qualification assessment value for payment verification is determined by calculating the dynamic equilibrium value of the primary verification qualification index and the secondary verification qualification index, specifically including the following steps:

[0106] Based on the current security requirements of the payment scenario, dynamically configure the primary weight coefficient of the primary verification qualification index and the secondary weight coefficient of the secondary verification qualification index.

[0107] The degree of synergy matching is obtained by calculating the balance between the primary verification pass index and the secondary verification pass index.

[0108] The overall qualification rating of payment verification is obtained by integrating the primary verification qualification index with the primary weight coefficient, the secondary verification qualification index with the secondary weight coefficient, and combining them with the collaborative matching degree value.

[0109] Based on the security requirements of the current payment scenario, the sovereign weight coefficient and auxiliary weight coefficient are dynamically configured. Different payment scenarios have vastly different risks, and the security requirement level directly determines the importance of core biometric features and auxiliary behavioral features in verification. A pre-defined rule for the correspondence between scenario security levels and weight coefficients is established. First, the type and risk attributes of the current payment scenario are identified, and then the corresponding weight configuration is matched. For example, online large-amount transfers belong to a high security requirement level. In this case, the dynamic fingerprint feature of the core verification layer is crucial for preventing impersonation. The sovereign weight coefficient is configured to 0.7, and the auxiliary weight coefficient is 0.3, thus highlighting the core verification role of biometric features. Offline small-amount convenience store consumption scenarios belong to a medium security requirement level. Users have higher requirements for payment convenience, so the sovereign weight coefficient can be adjusted to 0.6, and the auxiliary weight coefficient to 0.4, ensuring both security and efficiency. For low-security requirement scenarios such as online membership top-ups, the sovereign weight coefficient can be further reduced to 0.5, and the auxiliary weight coefficient increased to 0.5. Because the transaction amount is small and the frequency is high in these scenarios, the stability of behavioral features can provide sufficient support. The dynamic configuration method allows the weight coefficients to be accurately matched with the scenario risks, laying a reasonable foundation for subsequent calculations.

[0110] The degree of synergy is obtained by calculating the balance between the primary and secondary verification pass indices. The balance reflects the degree of fit between the primary and secondary verification pass indices. If both are at high or low levels, it indicates that the verification results of biometrics and behavioral characteristics are consistent, indicating good synergy. If one is extremely high and the other is extremely low, there is a verification contradiction, indicating poor synergy. The deviation rate between the two is used to quantify the balance. The formula for calculating the degree of synergy is: Synergy matching degree = 1 - |Primary verification pass index - Secondary verification pass index|. This value ranges from 0 to 1; the closer to 1, the better the synergy, and the closer to 0, the worse the synergy. Taking online large-amount transfer scenarios as an example, if a user's primary verification pass rate is 0.92 and the secondary verification pass rate is 0.88, the absolute value of the difference between the two is 0.04, and the degree of coordination is 0.96. This indicates that the verification results of fingerprint dynamic features and payment behavior features are highly coordinated and there is no obvious contradiction. If another user's primary verification pass rate is 0.95 and the secondary verification pass rate is 0.6 in the same scenario, the absolute value of the difference is 0.35, and the degree of coordination is 0.65. In this case, there may be a risk that the fingerprint information is misused but the behavioral features are abnormal, and the coordination is weak.

[0111] The primary and secondary verification pass indices are integrated with their corresponding weight coefficients, and combined with the degree of synergy matching value to obtain the comprehensive pass rating. The weighted results are then adjusted using the degree of synergy matching value, making verification results with good synergy more credible and results with poor synergy reasonably adjusted. The calculation formula is: Comprehensive pass rating = (Primary verification pass index × Primary weight coefficient + Secondary verification pass index × Secondary weight coefficient) × Degree of synergy matching value. For example, in a high-security online large-amount transfer scenario, User A's primary verification pass rate is 0.92, primary weight coefficient is 0.7, secondary verification pass rate is 0.88, secondary weight coefficient is 0.3, and collaborative matching degree value is 0.96. Their overall pass rate is (0.92×0.7+0.88×0.3)×0.96=0.872, indicating a high degree of credibility in the verification process. In the same scenario, User B's primary verification pass rate is 0.95, secondary verification pass rate is 0.6, and collaborative matching degree value is 0.65, resulting in an overall pass rate of (0.95×0.7+0.6×0.3)×0.65=0.549, indicating a lower degree of reliability and a potential risk of verification discrepancies. In offline small-amount consumption scenarios with medium security requirements, User C's primary verification pass rate is 0.85, primary weight coefficient is 0.6, secondary verification pass rate is 0.8, secondary weight coefficient is 0.4, and collaborative matching degree is 0.93. The overall pass rate is (0.85×0.6+0.8×0.4)×0.93=0.772, which meets the verification pass standard for this scenario. In contrast, in online membership recharge scenarios with low security requirements, User D's primary verification pass rate is 0.8, secondary verification pass rate is 0.82, both primary and secondary weight coefficients are 0.5, and collaborative matching degree is 0.98. The overall pass rate is (0.8×0.5+0.82×0.5)×0.98=0.794, meeting the verification requirements for convenient payment. The overall pass rate reflects both the differences in importance of features at different levels and the collaborative effectiveness between features, accurately reflecting the overall pass rate of the current payment verification and providing a scientific quantitative basis for generating subsequent verification results and security control schemes.

[0112] Based on the comprehensive qualification assessment value and the preset verification grading response rules, the payment verification result and corresponding security control plan are generated, which specifically includes the following steps:

[0113] If the overall qualification assessment value is higher than or equal to the excellent grade pass threshold, the excellent grade pass result is generated and the fast payment channel is activated to complete the settlement.

[0114] If the overall pass rating is between or equal to the excellent pass threshold and the normal pass threshold, a normal pass result is generated, and settlement is completed according to the regular payment process.

[0115] If the overall pass rating is between or equal to the verification threshold, the verification result will be generated and a prompt for supplementary account identity verification will be output.

[0116] If the overall qualification assessment value is between or equal to the verification threshold and the risk threshold, a verification restricted result is generated, the payment amount is limited, and verification abnormality information is recorded.

[0117] If the overall qualification assessment value is lower than the risk threshold, a verification failure result will be generated, and an emergency security control plan will be activated, including freezing the payment account and pushing risk warning information to users and merchants.

[0118] When the overall qualification assessment value is higher than or equal to the superior level pass threshold, a superior level pass result is generated and the fast payment channel is activated. The superior level pass threshold is the highest standard of verification credibility, usually set lower in low-security scenarios and extremely high in high-security scenarios. For example, online membership recharge is a low-security scenario, and the superior level pass threshold is preset to 0.85. User E's overall qualification assessment value in this scenario is calculated to be 0.88, which is higher than the threshold, so it is judged as superior level pass. At this time, the fast payment channel is activated, and the recharge amount is deducted and credited without user confirmation, thus improving convenience. On the other hand, online large-amount transfer is a high-security scenario, and the superior level pass threshold is preset to 0.95. Only when the user's fingerprint dynamic features and behavioral features match perfectly and have excellent coordination, such as user F's overall qualification assessment value reaching 0.96, will superior level pass be triggered. The fast payment channel skips the regular SMS verification process and directly completes the large-amount fund transfer, but will simultaneously generate an encrypted transaction certificate and push it to the user's mobile phone to ensure that the user is aware of the transaction.

[0119] If the overall pass rating is between or equal to the premium pass threshold and the normal pass threshold, a normal pass result is generated and the payment is settled according to the regular payment process. The normal pass threshold serves as the baseline for most normal payment scenarios, balancing security and efficiency. Taking a small-amount convenience store transaction as an example, the premium pass threshold is 0.8, the normal pass threshold is 0.7, and user G's overall pass rating is 0.75, falling between the two. Therefore, it is considered a normal pass, and the regular process is executed. After the user's fingerprint verification is successful, a transaction amount confirmation interface pops up. The user clicks to confirm and completes the payment. The POS machine prints a transaction receipt, and the transaction information is simultaneously synchronized to the user's payment app transaction record. For online shopping scenarios, the normal pass threshold is preset to 0.75. User H's overall pass rating is 0.75, which just meets the threshold. At this point, the user is required to enter an SMS verification code. After successful verification, the order is generated and the payment is deducted according to the regular process. The entire process conforms to the user's daily payment habits and does not create additional operational burdens.

[0120] When the overall qualification assessment value falls between or equal to the standard pass threshold and the verification pending threshold, a verification pending result is generated and an account identity supplementary verification prompt is output. Verification results within this range have slight uncertainty and require supplementary information to confirm the user's identity. For example, in offline restaurant payment scenarios, the standard pass threshold is 0.7, the verification pending threshold is 0.6, and user I's overall qualification assessment value is 0.65, falling within this range. A supplementary verification prompt will immediately pop up on the payment device. Common verification methods include entering the last four digits of the payment password, facial recognition, or answering preset security questions. Once the user enters the correct last four digits of the password, the system verifies the transaction and completes the payment. In online flight booking scenarios, the verification pending threshold is preset to 0.65. If user J's overall qualification assessment value is 0.65, the system will redirect to the identity supplementary verification page, requiring the user to upload photos of the front and back of their ID card for comparison. Only after successful comparison can the order proceed. If the user cannot provide valid supplementary information, the transaction will be temporarily frozen until the user contacts customer service for verification.

[0121] If the overall qualification assessment value falls between or equal to the verification threshold and the risk threshold, a verification restricted result is generated, limiting the payment amount and recording verification anomaly information. Payments within this range carry certain risks, requiring limit restrictions to mitigate potential losses, and retaining abnormal data for subsequent investigation. Taking online game recharge scenarios as an example, the verification threshold is 0.6, the risk threshold is 0.5, and user K's overall qualification assessment value is 0.55, falling within this range. Therefore, verification is restricted. First, the user's current single payment limit is limited to 500 yuan. If the user originally intended to recharge 1000 yuan, a limit restriction message is displayed, suggesting recharging in two installments. Simultaneously, detailed information about this verification anomaly is automatically recorded, including fingerprint feature deviations, abnormal IP addresses used for device login, differences between the payment time and historical habits, and this information is synchronized to the risk control backend for periodic analysis by risk control personnel. For large-amount offline supermarket purchases, the risk threshold is preset to 0.55. If user L's comprehensive qualification assessment value is 0.55, then the user's cumulative payment amount for the day will be limited to no more than 2,000 yuan. The user will also be sent a text message reminder that the limit has been limited and informed that the restriction can be lifted through customer service channels.

[0122] When the overall eligibility assessment value falls below the risk threshold, a verification failure result is generated and an emergency security control plan is initiated. The risk threshold is the bottom line for payment security; below this value, the payment carries extremely high risk, requiring strong measures to mitigate losses. For example, in online large-amount transfer scenarios, the risk threshold is preset to 0.5. User M's overall eligibility assessment value is 0.48, below the threshold, immediately indicating verification failure, and the emergency security control plan is activated simultaneously: first, the user's payment account is frozen, prohibiting any payment, transfer, or withdrawal operations; then, risk warning information is sent to the user via SMS, app push notifications, and telephone, clearly informing them of the abnormal payment attempt and the account freeze; simultaneously, a risk warning is pushed to the receiving merchant, informing them of the risk of the transaction and suggesting they suspend the provision of goods or services. For offline POS machine fraud scenarios, the risk threshold is preset to 0.45. When a payment's overall eligibility assessment value of 0.4 is detected, in addition to freezing the account and pushing warning information, the system will also synchronize the abnormal transaction information to the UnionPay risk control system to assist in tracing the flow of funds and minimize user losses.

[0123] Through a threshold-based response mechanism, precise processing solutions can be provided for verification results of different credibility levels, ensuring both strict control over high-risk transactions and convenient and efficient normal transactions, thus forming a complete closed loop for payment security.

[0124] It also includes dynamic feature calibration, which specifically includes the following steps:

[0125] Real-time monitoring of the stability of core verification layer features in a multi-level verification feature system;

[0126] If the stability of the core feature is lower than the preset stability threshold, the feature calibration mechanism will be activated, and the user's historical core feature data will be retrieved as the calibration benchmark.

[0127] By comparing features, the current core verification features are dynamically calibrated until the feature stability meets the verification requirements.

[0128] The system monitors the stability of core verification layer features in a multi-level verification feature system in real time. While performing payment verification, the system simultaneously activates the feature stability monitoring module. Using the fluctuation range of the user's historical core features as a benchmark, it calculates the stability value of the currently collected feature in real time. The calculation logic for the stability value is the proportion of the deviation between the current feature and historical features that falls within the historical normal fluctuation range; the higher the proportion, the stronger the stability. For example, the historical normal fluctuation range for the basic fingerprint deformation feature is ±3% deformation amplitude deviation. If the deviation of the currently collected deformation amplitude from the historical average is 2%, then the stability value of this feature is (3%-2%) / 3%×100%≈33.3%; if the deviation is 1%, the stability value increases to 66.7%. For fingerprint ridge temporal features, stability assessment focuses on the deviation of the time interval between ridge appearance and the historical average. The historical normal deviation range is ±0.2 seconds. A current deviation of 0.1 seconds results in a higher stability value, while a deviation of 0.3 seconds significantly reduces stability. The stability of fingerprint pressure dynamic features is evaluated by the deviation of the pressure gradient change rate. The historical normal deviation range is ±5% / second, and the smaller the current deviation, the higher the stability value. A uniform preset stability threshold is set for each core feature. When the stability value of any core feature falls below this threshold, the feature calibration mechanism is triggered.

[0129] If the stability of core features falls below a preset stability threshold, a feature calibration mechanism is activated, and the user's historical core feature data is retrieved as the calibration benchmark. Historical core feature data is extracted from distributed storage, representing the average value of valid core features for the user within the past three months in similar or identical payment scenarios. Feature data from abnormal transactions is excluded to ensure the reliability of the benchmark. For example, if a user makes a payment at a supermarket in cold weather, dry skin on their fingers can cause the stability value of the fingerprint pressure dynamic feature to drop to 52%, below the 60% stability threshold. The system immediately triggers calibration, retrieving 100 valid payment records from the distributed nodes for the user within the past three months in offline supermarket scenarios. The average rate of change of the fingerprint pressure gradient is calculated to be 8% / second, and the average coordinates of the pressure concentration area are used as the calibration benchmark for the fingerprint pressure dynamic feature. If a user's fingerprint ridge temporal feature stability drops to 55% due to a minor finger injury, ridge temporal data from all regular payment scenarios within the past three months is retrieved. The average time interval from partial to complete ridge appearance and the average order of appearance of each ridge are calculated as the calibration benchmark for the ridge temporal feature. During the retrieval process, the system will obtain data through an encrypted transmission channel to avoid the leakage of historical feature information.

[0130] Through feature comparison, the current core verification features are dynamically calibrated until the feature stability meets the verification requirements. Dynamic calibration uses historical calibration benchmarks as a reference, correcting deviations in the current features while preserving reasonable individual differences. During calibration, differentiated comparison and correction logic is employed for different core features. Taking the calibration of fingerprint deformation basic features as an example, the deviation between the current deformation trajectory and the historical benchmark trajectory is mainly concentrated in the fingerprint edge region. First, the deviation region is identified, and then, based on the deformation pattern of this region in the historical benchmark, the current deviation data is fine-tuned. If the deformation of the edge region in the historical benchmark is a smooth curve, while the current feature edge has sharp angles, the angled parts are corrected to a smooth curve conforming to historical patterns. After correction, the stability value is recalculated. If it is still below 60%, the correction is repeated until the requirements are met. For fingerprint ridge timing features, if the time interval between the current ridge appearances lags behind the historical baseline by 0.3 seconds, the lag is assessed based on the user's historical ridge timing trends to determine if it is within an acceptable range. If it is a temporary fluctuation, the current timing data is shifted and corrected according to the historical trend, reducing the time interval deviation to within 0.1 seconds and improving the stability value to over 65%. The calibration of fingerprint pressure dynamic features involves comparing the current pressure gradient change rate with the historical baseline. If the current rate is 12% / second, higher than the historical baseline of 8% / second, the current rate is corrected to 9% / second based on the pressure rate variation in historical data, making the pressure distribution more consistent with the user's typical characteristics. After calibration, the calibrated core features are re-input into the multi-level verification feature system to participate in subsequent reliability coefficient calculations and verification evaluations, ensuring that the final comprehensive pass / fail rating accurately reflects the user's actual verification performance.

[0131] The dynamic feature calibration mechanism enables the payment verification system to flexibly respond to temporary fluctuations in users' core features, which not only avoids the rejection of normal payments due to feature instability, but also improves the ability to identify abnormal features, further improving the security closed loop of payment verification.

[0132] A desktop cash register with fingerprint recognition and payment verification includes:

[0133] Acquisition Module: Acquires dynamic fingerprint biometric data and payment behavior-related data of payment users, and decomposes and reassembles the dynamic fingerprint biometric data and payment behavior-related data to form a multi-level verification feature system;

[0134] Verification module: Calculates the reliability coefficient and correlation tightness of features at each level in the multi-level verification feature system to obtain feature verification indicators;

[0135] Mining Module: Mines feature verification evolution patterns under different payment scenarios from historical verification data in distributed storage, constructs an adaptive verification evaluation model based on feature verification evolution patterns, and inputs feature verification indicators into the adaptive verification evaluation model to obtain the main verification qualification index and auxiliary verification qualification index of payment users;

[0136] Calculation module: Determines the overall qualification assessment value for payment verification by calculating the dynamic equilibrium value of the primary verification qualification index and the secondary verification qualification index;

[0137] Control module: Generates payment verification results and corresponding security control schemes based on the comprehensive qualification assessment value and the preset verification grading response rules.

[0138] It also includes: a rear scanning module, a front scanning module, an iButton module, a printer module, an NFC module, a card swiping module, a fingerprint module, a customer display module, and a four-in-one PCB. The front scanning module collects transaction product information, generates a payment bill, and displays it through the customer display module. The user obtains fingerprint information through the fingerprint acquisition module, which then transmits the fingerprint information to the four-in-one PCB. The four-in-one PCB retrieves a pre-stored fingerprint database of payment accounts and matches the collected fingerprint information with the database. If the match is successful, the four-in-one PCB sends a payment instruction to the payment system to complete the transaction and simultaneously triggers the printing module to output a transaction voucher. If the match fails, a verification failure message is output and the transaction is terminated. While the fingerprint acquisition module is acquiring fingerprint information, the four-in-one PCB simultaneously obtains the user's auxiliary payment information through the NFC module or card swiping module. The auxiliary payment information is used to verify the fingerprint information through the verification module. The rear scanning module is used to collect the user's payment code information to complete the payment.

[0139] During the transaction initiation phase, the front-end scanning module and the customer display module complete the collection of product information and the presentation of the bill. The front-end scanning module is usually integrated into the barcode scanner or fixed scanning platform at the cashier. It has the function of quickly recognizing barcodes and QR codes. The cashier scans the products selected by the customer one by one through this module. The product information is transmitted to the four-in-one PCB in real time. The calculation unit inside the four-in-one PCB summarizes and calculates the information, automatically generating a payment bill that includes a list of products, unit price, quantity, total price, and discount amount. Subsequently, the four-in-one PCB sends the bill data to the customer display module. As the display terminal facing the user, the customer display module clearly displays the complete bill information, making it convenient for the user to check the amount and avoid settlement errors.

[0140] After the user confirms the bill, the payment verification stage begins. The NFC module or card reader module acts as an auxiliary verification unit, jointly verifying identity and payment information. The user provides fingerprint information through the fingerprint acquisition module, which uses optical or capacitive acquisition technology to accurately capture the dynamic texture features of the fingerprint. The four-in-one PCB simultaneously triggers the auxiliary verification mechanism, obtaining the user's auxiliary payment information through the NFC module or card reader module: If the user chooses NFC payment, they simply need to bring their NFC-enabled mobile phone or bank card close to the NFC module, and the NFC module will read auxiliary data such as the payment account identifier and security chip information within the device or card; if the user chooses card payment, after inserting the bank card into the card slot of the card reader module, the card reader module will read the card's magnetic stripe or chip information and extract account association information as auxiliary verification data.

[0141] The 4-in-1 PCB plays a crucial role in data matching and verification decisions. Upon acquiring fingerprint and auxiliary payment information, the 4-in-1 PCB first retrieves its internally stored payment account fingerprint database. This database is linked to the user's payment account and stores the standard fingerprint feature data provided by the user during account opening, encrypted to ensure information security. Subsequently, the 4-in-1 PCB initiates a dual verification process: the first step is fingerprint information matching, precisely comparing the currently acquired fingerprint dynamic features with the standard features in the fingerprint database, including the overlap of ridge feature points and the consistency of the pressing dynamic trajectory; the second step is auxiliary information cross-verification, matching the payment account identifier obtained by the NFC module or card reader with the payment account associated with the fingerprint database to confirm that both belong to the same user.

[0142] The 4-in-1 PCB sends corresponding instructions to each module based on the results, completing the transaction or handling any exceptions. If verification is successful, the 4-in-1 PCB immediately sends a payment instruction to the payment system, which includes a verification pass identifier, payment account information, and transaction amount. Upon receiving the instruction, the payment system completes the fund deduction and settlement. Simultaneously, the PCB triggers the printer module to operate, quickly printing out a transaction voucher containing information such as the transaction serial number, time, product list, payment method, and amount for the user to keep.

[0143] The system also supports payment code payments via a rear-mounted scanning module, providing users with diverse options. The rear-mounted scanning module is typically installed on the user-facing side of the checkout counter for easy self-service. When a user chooses to present a payment code from their mobile payment app, they simply align the code with the scanning window of the rear-mounted module. The module quickly identifies the account information and payment authorization data within the payment code and transmits it to the four-in-one PCB. The four-in-one PCB parses and verifies the payment code information. Once the account status is confirmed to be normal, it directly sends a deduction instruction to the payment system to complete the transaction, simultaneously triggering the printer module to output a receipt. This makes it suitable for scenarios where fingerprint verification is inconvenient, such as when a user's fingers are wet, making fingerprint collection difficult. Users can quickly switch to payment code payments, improving transaction flexibility.

[0144] The iButton module, serving as an auxiliary security unit, primarily stores the device's encryption keys and security configuration information. During data transmission within the four-in-one PCB, the iButton module provides encrypted authentication, ensuring that sensitive data such as fingerprint information and payment account information are not tampered with or leaked during transmission, further enhancing the system's security capabilities. Through the coordinated scheduling of the four-in-one PCB, all modules achieve fully automated processes for product acquisition, billing presentation, dual verification, and transaction settlement, perfectly adapting to the high-frequency transaction scenarios of offline retail.

[0145] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A verification method with fingerprint recognition and payment verification, characterized in that, The method includes the following steps: Acquire dynamic fingerprint biometric data and payment behavior-related data of payment users, and decompose and reassemble the dynamic fingerprint biometric data and payment behavior-related data to form a multi-level verification feature system; The reliability coefficients and correlation tightness of features at each level in the multi-level verification feature system are calculated to obtain the feature verification index; We mine the evolution patterns of feature verification under different payment scenarios from historical verification data in distributed storage, and construct an adaptive verification evaluation model based on the evolution patterns of feature verification. We input the feature verification indicators into the adaptive verification evaluation model to obtain the main verification qualification index and the auxiliary verification qualification index of payment users. The overall qualification assessment value for payment verification is determined by calculating the dynamic equilibrium value of the primary verification qualification index and the secondary verification qualification index. Based on the comprehensive qualification assessment value and the preset verification grading response rules, payment verification results and corresponding security control schemes are generated.

2. The desktop cash register and verification method with fingerprint recognition payment verification according to claim 1, characterized in that, The fingerprint biometric dynamic data includes dynamic deformation feature data during fingerprint acquisition, fingerprint ridge temporal change feature data, and fingerprint pressure distribution dynamic data. The payment behavior associated data includes payment account associated device data, payment habit preference data, and recent transaction flow feature data.

3. The verification method with fingerprint recognition and payment verification according to claim 2, characterized in that, The process of decomposing and recombining dynamic fingerprint biometric data with payment behavior-related data to form a multi-level verification feature system includes the following steps: The basic features of fingerprint deformation are obtained by decomposing the deformation trajectory and deformation amplitude features of the dynamic deformation feature data of fingerprints. Fingerprint temporal features are obtained by decomposing fingerprint ridge temporal variation feature data into ridge continuity and ridge variation features; The dynamic features of fingerprint pressure are obtained by decomposing the dynamic data of fingerprint pressure distribution into pressure gradient and pressure concentration region features. The device association characteristics of payment accounts are obtained by performing device type matching and device login frequency feature decomposition on the data of devices associated with payment accounts. Payment habit characteristics are obtained by decomposing payment habit preference data into payment time period preference and payment amount range characteristics; Transaction flow characteristics are obtained by decomposing recent transaction flow characteristic data into transaction frequency and transaction object association characteristics; Based on the rules for classifying features according to their importance, the basic features of fingerprint deformation, the temporal features of fingerprint patterns, the dynamic features of fingerprint pressure, the features associated with accounts and devices, the features of payment habits, and the features of transaction flow are reorganized into the core verification layer, the auxiliary verification layer, and the supplementary verification layer, forming a multi-level verification feature system.

4. The verification method with fingerprint recognition and payment verification according to claim 3, characterized in that, The reliability coefficients and correlation strengths of features at each level in a multi-level verification feature system are calculated to obtain feature verification indices. This process includes the following steps: The reliability coefficients of each level of features in the multi-level verification feature system are calculated to obtain the feature coefficient values; The degree of feature correlation is obtained by judging the closeness of the correlation between features at different levels and between features within the same level. The feature verification index is obtained by fusing the feature coefficient values ​​with the corresponding feature correlation.

5. The verification method with fingerprint recognition and payment verification according to claim 4, characterized in that, The evolutionary patterns of feature verification under different payment scenarios are mined from historical verification data in distributed storage. An adaptive verification evaluation model is then constructed based on these patterns. The specific steps include: Extract valid verification samples from historical verification data stored across multiple nodes; By determining the correlation and evolution trend between the changes of features over time in valid validation samples and the validation results, the evolution law of feature validation is obtained. An initial adaptive verification and evaluation model is constructed based on the evolutionary rules of feature verification.

6. The verification method with fingerprint recognition and payment verification according to claim 5, characterized in that, The feature verification metrics are input into the adaptive verification evaluation model to obtain the primary verification qualification index and the secondary verification qualification index for payment users. This process includes the following steps: The feature verification indicators corresponding to the core verification layer in the multi-level verification feature system are determined as the main verification input parameters and input into the core evaluation unit of the adaptive verification evaluation model. The main verification qualification index, which reflects the qualification level of the core verification dimension, is judged and output through the core evaluation unit. The feature verification indicators corresponding to the auxiliary verification layer and the supplementary verification layer are determined as auxiliary verification input parameters and input into the auxiliary evaluation unit of the adaptive verification evaluation model. Through the comprehensive judgment of the auxiliary evaluation unit, the auxiliary verification qualification index reflecting the qualification level of the auxiliary verification dimension is output.

7. The verification method with fingerprint recognition and payment verification according to claim 6, characterized in that, The overall qualification assessment value for payment verification is determined by calculating the dynamic equilibrium value of the primary verification qualification index and the secondary verification qualification index, specifically including the following steps: Based on the current security requirements of the payment scenario, dynamically configure the primary weight coefficient of the primary verification qualification index and the secondary weight coefficient of the secondary verification qualification index. The degree of synergy matching is obtained by calculating the balance between the primary verification pass index and the secondary verification pass index. The overall qualification rating of payment verification is obtained by integrating the primary verification qualification index with the primary weight coefficient, the secondary verification qualification index with the secondary weight coefficient, and combining them with the collaborative matching degree value.

8. The verification method with fingerprint recognition and payment verification according to claim 7, characterized in that, Based on the comprehensive qualification assessment value and the preset verification grading response rules, the payment verification result and corresponding security control plan are generated, which specifically includes the following steps: If the overall qualification assessment value is higher than or equal to the excellent grade pass threshold, the excellent grade pass result is generated and the fast payment channel is activated to complete the settlement. If the overall pass rating is between or equal to the excellent pass threshold and the normal pass threshold, a normal pass result is generated, and settlement is completed according to the regular payment process. If the overall pass rating is between or equal to the verification threshold, the verification result will be generated and a prompt for supplementary account identity verification will be output. If the overall qualification assessment value is between or equal to the verification threshold and the risk threshold, a verification restricted result is generated, the payment amount is limited, and verification abnormality information is recorded. If the overall qualification assessment value is lower than the risk threshold, a verification failure result will be generated, and an emergency security control plan will be activated, including freezing the payment account and pushing risk warning information to users and merchants.

9. A verification method with fingerprint recognition and payment verification according to any one of claims 1-8, characterized in that, It also includes dynamic feature calibration, which specifically includes the following steps: Real-time monitoring of the stability of core verification layer features in a multi-level verification feature system; If the stability of the core feature is lower than the preset stability threshold, the feature calibration mechanism will be activated, and the user's historical core feature data will be retrieved as the calibration benchmark. By comparing features, the current core verification features are dynamically calibrated until the feature stability meets the verification requirements.

10. A desktop cash register with fingerprint recognition and payment verification, applied to the verification method with fingerprint recognition and payment verification as described in claims 1-9, characterized in that, include: Acquisition Module: Acquires dynamic fingerprint biometric data and payment behavior-related data of payment users, and decomposes and reassembles the dynamic fingerprint biometric data and payment behavior-related data to form a multi-level verification feature system; Verification module: Calculates the reliability coefficient and correlation tightness of features at each level in the multi-level verification feature system to obtain feature verification indicators; Mining Module: Mines feature verification evolution patterns under different payment scenarios from historical verification data in distributed storage, constructs an adaptive verification evaluation model based on feature verification evolution patterns, and inputs feature verification indicators into the adaptive verification evaluation model to obtain the main verification qualification index and auxiliary verification qualification index of payment users; Calculation module: Determines the overall qualification assessment value for payment verification by calculating the dynamic equilibrium value of the primary verification qualification index and the secondary verification qualification index; Control module: Generates payment verification results and corresponding security control schemes based on the comprehensive qualification assessment value and the preset verification grading response rules.

Images