Self-service storage interaction method and system based on shared information of bank financial system

Abstract

The invention discloses a self-service storage interaction method and system based on shared information of a bank financial system, and belongs to the technical field of self-service storage interaction. The system comprises a posture size parameter generation module, a posture verification module, an identity verification information generation module and an identity verification module. According to the invention, through an intelligent complement scheduling strategy based on comprehensive cost optimization, efficient execution of a complement service is realized, mechanical loss is reduced, a complement task is converted into a combinatorial optimization problem, an optimal page turning and printing sequence is calculated through an algorithm, and a mechanical motion path and time consumption are reduced, so that the service life of equipment is prolonged; the problems that in the prior art, bankbook self-service equipment is low in reregistration service efficiency and high in mechanical abrasion speed are solved.

Description

Technical Field

[0001] This invention relates to the field of self-service storage and interaction technology, and in particular to a self-service storage and interaction method and system based on shared information in a banking and financial system. Background Technology

[0002] With the development of information technology and self-service in the banking and financial system, passbook self-service machines have become an important part of financial services, widely used for deposits and withdrawals, balance inquiries, and transaction record updates. The level of intelligence of passbook self-service machines directly affects the efficiency of banking services and user experience.

[0003] Most existing self-service passbook machines perform page turning and printing operations in a linear order of passbook page numbers during the supplementary entry process. The machine starts from the first page, turns the pages one by one, and checks each page for any transaction records to be printed. If a record is found, it is printed; otherwise, the machine continues turning the pages until all supplementary entries are completed.

[0004] In existing technologies, because the transaction records to be supplemented may be distributed anywhere on the passbook, when the transaction records are scattered, the device needs to repeatedly flip through pages. This is especially true when the transactions to be supplemented are located on distant pages, significantly increasing the number of mechanical page flips and prolonging the operation time. Excessive page flipping accelerates the wear and tear on core components (such as the page-flipping mechanism), shortens the device's lifespan, and increases maintenance costs. This sequential processing method lacks optimization and cannot perform targeted operations based on the actual distribution of transaction records, resulting in overall low efficiency. Therefore, this sequential processing method suffers from low efficiency and rapid mechanical wear in passbook self-service equipment for supplementing transactions. Summary of the Invention

[0005] This application provides a self-service storage interaction method and system based on shared information in the banking and financial system. This solves the problems of low efficiency and rapid mechanical wear in the existing passbook self-service equipment, achieves efficient execution of the update business, reduces mechanical wear, reduces mechanical movement distance and time, and extends the service life of the equipment.

[0006] This application provides a self-service storage interaction method based on shared information of a banking financial system, including: S1, acquiring and monitoring the posture information of a user placing a passbook into a self-service device, processing the posture information based on a passbook posture recognition algorithm, and generating posture normalization parameters and size normalization parameters;

[0007] S2, determine whether the posture and size standardization parameters conform to the preset physical model standard. If they do, start the page turning and magnetic track information reading process to obtain the existing magnetic track data in the passbook. If they do not conform, reject the service and exit the passbook.

[0008] S3, compare the read track data with the authentication information entered by the user through the interactive interface, and generate a matching similarity that represents the correlation between the two based on an asymmetric cross-modal hash matching algorithm;

[0009] S4, Track Information Comparison and Verification: A dynamic security threshold is generated in real time based on the system risk assessment model. When the matching similarity is higher than the dynamic security threshold, the verification is successful. Based on the specific business type selected by the user, the account data in the bank's financial shared information system is called to execute the corresponding automated business operation process. When the matching similarity is not higher than the dynamic security threshold, the verification fails, the business is rejected, and the passbook is exited. The specific business types include: supplementary registration business, new account opening business, and passbook replacement business.

[0010] For supplementary registration, the process involves obtaining the transaction records to be supplemented and their corresponding target print page numbers from the bank's financial shared information system, analyzing the overall cost to determine the optimal execution sequence, and then executing it.

[0011] Furthermore, the step of generating the attitude normalization parameters includes:

[0012] At least two image sensors at different angles are used to acquire attitude images of the passbook, and a geometric feature vector V containing the passbook's deflection angle and centroid coordinate offset is extracted. img ;

[0013] For the geometric feature vector V img Probability density estimation is performed to calculate the posterior probability of belonging to the correct pose category, and this probability is used as the pose normalization parameter P. pose :

[0014]

[0015] Where K is the preset number of Gaussian components in the Gaussian mixture model, φ k The weight of the k-th Gaussian component is N(V). img |μ k ,Σ k ) indicates that the mean is μ k The covariance matrix is ​​Σ k Gaussian distribution on geometric eigenvector V img The probability density at that location.

[0016] Furthermore, the step of generating the size normalization parameters also includes:

[0017] The length l and width w of the passbook are measured from the posture information to form a size measurement vector V. dim =[l,w] T ;

[0018] Calculate the size measurement vector V dim With standard size vector μ dim =[l std ,w std ] T Mahalanobis distance D between M And use this distance as a size normalization parameter:

[0019]

[0020] Where S is the covariance matrix of the passbook size measurement, S -1 Let S be the inverse matrix;

[0021] The criterion is that when the attitude normalization parameter is greater than the preset probability threshold and the Mahalanobis distance is less than the preset distance threshold, it is determined to meet the physical model standard.

[0022] Furthermore, the step of generating a matching similarity that characterizes the correlation between the two includes:

[0023] Define the structured data read from the track as D mag The user-entered authentication information is D. user ;

[0024] Two different, pre-defined hash functions are used, including the track hash function f. mag and user input hash function f user ;

[0025] For structured data D respectively mag and authentication information D user The data is processed to generate corresponding fixed-length binary hash codes H. mag and H user H mag =f mag (D mag ,K1),H user =f user (D user K1 and K2 are encryption salts;

[0026] K1 and K2 are dynamically generated in the following manner:

[0027] K1 = HMAC-SHA256(device serial number, key seed for the day);

[0028] K2 = HMAC - SHA256(Session ID, User Operation Timestamp);

[0029] The differences are quantified by calculating the weighted Hamming distance between the two hash codes and then converted into matching similarity.

[0030] Furthermore, the step of generating dynamic security thresholds in real time based on the system risk assessment model includes:

[0031] Obtain users' historical transaction behavior information and the current system environment risk index, and analyze the dynamic security threshold T. dyn :

[0032]

[0033] Among them, T base Let N be a preset baseline safety threshold, α be a transaction behavior adjustment coefficient, and N be a threshold value. freq The frequency of transactions for this user within a specified time window, where N0 is the frequency normalization parameter, and σ is the frequency of transactions for this user within a specified time window. env Environmental risk index;

[0034] tanh is the hyperbolic tangent function, used to smoothly adjust the threshold;

[0035] σ env Obtain it through the following methods:

[0036] Furthermore, when a user selects the supplementary registration service, the steps for executing the corresponding automated business operation process include:

[0037] Retrieve the set of transaction records to be supplemented and their corresponding target print page numbers p1, p2, ..., p from the bank's financial shared information system. a ,...,p b ,...,p m m is the total number of pages to be printed, m≥1;

[0038] Establish a comprehensive cost function J(π) for page redirection and printing, and solve for its optimal execution sequence π. * The specific solution formula is as follows:

[0039]

[0040] Where π is the sequence of pages to be printed, p π(g) p is the target page number for the g-th page turn in the sequence, where g is the page turn operation number. π(j) Let C be the target page number for the j-th print operation in the sequence, where j is the print operation number. turn (p a ,p b ) is the passbook from the pth a Turn to page p b The mechanical time cost function of a page, where a is the current page, b is the target page, and C is the target page. print (p π(j) () represents the cost of time required to print the target page number in the j-th print in the sequence;

[0041] Based on the optimal execution sequence π * Perform the page turning and printing operations in sequence.

[0042] Furthermore, when a user selects the new account opening service, the steps for executing the corresponding automated business operation process include:

[0043] The user's identity is verified, then the new book is issued, the book is pushed to the page-turning area, and the first page is turned;

[0044] If page turning fails, the waste book recycling process will be initiated;

[0045] If the page is turned successfully, the adhesive silicone roller will feed the paper and separate the pages. The roller will then feed the book into the magnetic printing module and print the relevant information into the passbook's magnetic track.

[0046] After the magnetic stripe information is entered into the new passbook, a magnetic stripe verification operation is immediately performed to obtain the waveform of the read signal;

[0047] By performing a Fourier transform on the read-back signal, the effective power of the signal and the background noise power are separated, and the signal-to-noise ratio of the brushing quality is calculated.

[0048] If the signal-to-noise ratio is lower than the preset quality standard threshold, the magnetic stripping is deemed to have failed, and the waste paper recycling process is initiated.

[0049] If the magnetic stripe is successfully scanned, the account opening information on the first page will be printed. After printing, the account will be sent to the exit by the institution for the user to pick up.

[0050] Furthermore, the initiation of the waste recycling process includes:

[0051] When it is determined that the magnetic stripe printing failed or the new book page turning failed, the roller of the passbook delivery mechanism is controlled to roll in reverse, and the new passbook that failed to be made is delivered to a dedicated waste book recycling bin that is physically isolated from the old book recycling bin.

[0052] After the scrapped units are recycled, the system automatically initializes all mechanical mechanisms to standby mode and restarts the issuance of new units, without requiring user intervention.

[0053] Furthermore, when a user selects the passbook renewal service, the steps for executing the corresponding automated business operation process include:

[0054] After successfully reading the magnetic track information and verifying the identity of the old passbook, the old passbook is transported to the designated old passbook recycling bin, which is physically isolated from the old passbook recycling bin, by controlling the mechanical structure.

[0055] Subsequently, the new account opening process was seamlessly integrated, and a new passbook was issued and created for the user.

[0056] This application provides a self-service storage and interaction system based on shared information of a banking and financial system, which is used to implement the self-service storage and interaction method based on shared information of a banking and financial system, including: a posture size parameter generation module, a posture verification module, an identity verification information generation module, and an identity verification module;

[0057] The posture and size parameter generation module is used to acquire and monitor the posture information of the user placing the passbook into the self-service device, and process the posture information based on the passbook posture recognition algorithm to generate posture normalization parameters and size normalization parameters.

[0058] The posture verification module is used to determine whether the posture and size standardization parameters conform to the preset physical model standard. If they conform, the page turning and magnetic track information reading process is started to obtain the existing magnetic track data in the passbook. If they do not conform, the service is rejected and the passbook is exited.

[0059] The authentication information generation module is used to compare the read track data with the authentication information input by the user through the interactive interface, based on an asymmetric cross-modal hash matching algorithm, to generate a matching similarity that characterizes the correlation between the two.

[0060] The identity verification module is used for magnetic stripe information comparison verification: a dynamic security threshold is generated in real time according to the system risk assessment model. When the matching similarity is higher than the dynamic security threshold, the verification is successful. Based on the specific business type selected by the user, the account data in the bank's financial shared information system is called to execute the corresponding automated business operation process. When the matching similarity is not higher than the dynamic security threshold, the verification fails, the business is rejected, and the passbook is exited. The specific business types include: supplementary registration business, new account opening business, and passbook replacement business.

[0061] For supplementary registration, the process involves obtaining the transaction records to be supplemented and their corresponding target print page numbers from the bank's financial shared information system, analyzing the overall cost to determine the optimal execution sequence, and then executing it.

[0062] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0063] 1. By adopting an intelligent supplementary registration scheduling strategy based on comprehensive cost optimization, the supplementary registration business is executed efficiently, reducing mechanical wear and tear. The supplementary registration task is transformed into a combinatorial optimization problem. The algorithm calculates the optimal page turning and printing order, reducing the mechanical movement distance and time consumption. This solves the problems of low efficiency and rapid mechanical wear in traditional passbook self-service equipment, and extends the service life of the equipment.

[0064] 2. By using a passbook posture recognition algorithm based on Gaussian mixture model and Mahalanobis distance, accurate perception of passbook posture is achieved, improving the fault tolerance when the passbook is placed in the device. It abandons the traditional binary detection logic and, through probabilistic modeling and Mahalanobis distance calculation, can handle small errors in user operation, thus effectively solving the problem of low success rate of traditional passbook self-service equipment and significantly improving the user experience.

[0065] 3. By using an asymmetric cross-modal hash matching algorithm, data from different modalities are converted into a unified hash space for comparison, which improves the robustness and efficiency of verification. The weighted Hamming distance is used, which can tolerate minor differences in non-critical parts of the data while remaining sensitive to differences in core information, and at the same time ensures comparison efficiency. This effectively solves the problem of rigid information verification in traditional passbook self-service equipment and enhances security. Attached Figure Description

[0066] Figure 1 A flowchart illustrating a self-service storage and interaction method based on shared information in a banking and financial system, as provided in this application embodiment;

[0067] Figure 2 This is a structural diagram of a self-service storage and interaction system based on shared information in a banking and financial system, provided in an embodiment of this application. Detailed Implementation

[0068] This application provides a self-service storage interaction method and system based on shared information in the banking and financial system. It solves the problems of low efficiency and rapid mechanical wear in the existing passbook self-service equipment. By using an intelligent supplementary registration scheduling strategy based on comprehensive cost optimization, it achieves efficient execution of supplementary registration, reduces mechanical wear, and transforms the supplementary registration task into a combinatorial optimization problem. By calculating the optimal page turning and printing order through algorithms, it reduces the mechanical movement distance and time, thereby extending the service life of the equipment.

[0069] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0070] like Figure 1 As shown, this application provides a self-service storage interaction method based on shared information of a banking financial system. The method is applied to a self-service storage interaction system based on shared information of a banking financial system, including: S1, acquiring and monitoring the posture information of a user placing a passbook into a self-service device, processing the posture information based on a passbook posture recognition algorithm, and generating posture normalization parameters and size normalization parameters.

[0071] S2, determine whether the posture and size standardization parameters conform to the preset physical model standard. If they do, start the page turning and magnetic track information reading process to obtain the existing magnetic track data in the passbook. If they do not conform, reject the service and exit the passbook.

[0072] S3, compare the read track data with the authentication information entered by the user through the interactive interface, and generate a matching similarity that represents the correlation between the two based on an asymmetric cross-modal hash matching algorithm;

[0073] S4, Track Information Comparison and Verification: A dynamic security threshold is generated in real time based on the system risk assessment model. When the matching similarity is higher than the dynamic security threshold, the verification is successful. Based on the specific business type selected by the user, the account data in the bank's financial shared information system is called to execute the corresponding automated business operation process. When the matching similarity is not higher than the dynamic security threshold, the verification fails, the business is rejected, and the passbook is exited. The specific business types include: supplementary registration business, new account opening business, and passbook replacement business.

[0074] For supplementary registration, the process involves obtaining the transaction records to be supplemented and their corresponding target print page numbers from the bank's financial shared information system, analyzing the overall cost to determine the optimal execution sequence, and then executing it.

[0075] In this embodiment, when a user places their passbook into the self-service passbook machine, the machine monitors the passbook's posture. A passbook posture recognition algorithm checks if the posture and size of the passbook are within specifications. If correct, it is marked as correct; otherwise, it is marked as incorrect. If incorrect, the passbook is returned, and the user is notified that an error has occurred. If correct, the user continues filling in their identity information and uses a scroll wheel to advance the passbook to the page-turning area for the first page turn. If page turning fails, the scroll wheel descends to prepare to remove the passbook, and the passbook is returned. If page turning is successful, the passbook is sent to the printer for magnetic track information reading. The read magnetic track information is compared with the entered information. If they match, it is marked as successful; otherwise, it is marked as unsuccessful. If unsuccessful, the passbook is ejected, the scroll wheel rolls out of the passbook to the exit, and the user is notified that magnetic track information is missing or identity information does not match. If successful, the user performs the selected service.

[0076] When a user selects the reprint service, the magnetic stripe is read and the book is printed. The following steps are executed in sequence: the passbook is turned to the reprint start page in the page turning area, the book is fed into the printer, the printout is completed and the book is returned to the page turning area, it is determined whether the reprint task is completed. If not, the page turning continues; if the task is completed, the book is returned to the exit.

[0077] When a user selects the passbook replacement service, the magnetic stripe is read to complete the issuance of the new passbook. Then, the roller descends to prepare to collect the old passbook. The roller rolls to collect the passbook, the institution's position is initialized, and the process of initiating the issuance of the new passbook in the new account opening service is connected, and subsequent steps are executed.

[0078] When a user selects the new account opening service, the user's identity is verified, then the new booklet is issued, the booklet is pushed to the page-turning area, and the first page is turned.

[0079] If page turning fails, the scroll wheel descends to prepare for waste book collection. The scroll wheel rolls to collect the waste book, then the mechanism position is initialized, and new book issuance restarts.

[0080] If the page turning is successful, the adhesive silicone roller feeds the paper and separates the pages. The roller then feeds the book into the magnetic printing module, where the relevant information is printed onto the passbook's magnetic track. If the magnetic printing fails, the passbook is ejected, the roller descends to prepare for waste book collection, the roller rolls to collect the waste book, the mechanism is then initialized, and new passbook issuance is restarted. If the magnetic printing is successful, the first page of account opening information is printed. After printing, the passbook is sent to the exit by the organization for the user to pick up.

[0081] Furthermore, the step of generating the attitude normalization parameters includes:

[0082] At least two image sensors at different angles are used to acquire attitude images of the passbook, and the values ​​of the passbook's deflection angle θ and centroid coordinate offset (δ) are extracted. x ,δ y Geometric eigenvectors V, including img ;

[0083] The passbook deflection angle θ is obtained by detecting the straight line at the edge of the passbook using the Hough transform and calculating the angle between it and the device coordinate axis from the passbook image acquired by the image sensor. The centroid coordinate offset (δ) x ,δ y The OpenCV contour detection method with morphological processing is used to extract the passbook contour. The centroid coordinates are obtained by first moment, and the offset is the difference between the centroid and the standard center point (0,0) of the device.

[0084] For the geometric feature vector V img Probability density estimation is performed to calculate the posterior probability that it belongs to the "correct pose" category, and this probability is used as the pose normalization parameter P. pose :

[0085]

[0086] Where K is the preset number of Gaussian components in the Gaussian mixture model, φ k The weight of the k-th Gaussian component is N(V). img |μ k ,Σ k ) indicates that the mean is μ k The covariance matrix is ​​Σ k Gaussian distribution on geometric eigenvector V img The probability density at that location;

[0087] The number of Gaussian components is automatically determined using the Bayesian Information Criterion (BIC): K = arg min k [-2ln L+k lnM], where L is the maximum likelihood value of the Gaussian mixture model on the training sample set, and M is the number of samples in the training sample set, with at least 1000 sets of standard pose samples.

[0088] The μ k and Σ k It is obtained by training a Gaussian mixture model using the EM algorithm with at least 1000 sets of standard pose samples.

[0089] Furthermore, the step of generating the size normalization parameters also includes:

[0090] The length l and width w of the passbook are measured from the posture information to form a size measurement vector V. dim =[l,w] T ;

[0091] Calculate the size measurement vector V dim With standard size vector μ dim =[l std ,w std ] T Mahalanobis distance D between M And use this distance as a size normalization parameter:

[0092]

[0093] Where S is the covariance matrix of the passbook size measurement, representing the relationship between the length and width measurement errors. -1 Let S be the inverse matrix;

[0094] The S is obtained by performing at least 50 repeated measurements using a standard passbook and calculating the covariance matrix of the measurement error.

[0095] The judgment criterion is when the attitude normalization parameter P pose Greater than the preset probability threshold τ P And Mahalanobis distance D M Less than the preset distance threshold τ D When the time is right, it is determined to conform to the physical model standard.

[0096] probability threshold τ P Set according to the False Rejection Rate (FRR) requirement: τ P =Φ -1 (1-FRR), where Φ is the standard normal cumulative distribution function. Distance threshold τ D Based on the 3σ principle: λ maxLet be the largest eigenvalue of the covariance matrix S. Standard size l std ,w std The standard passbook size is used as stored in the database.

[0097] Standard size vector μ dim The covariance matrix S is obtained through a calibration process before the equipment leaves the factory, specifically including:

[0098] Fix the standard-sized passbook in the center of the device's imaging area;

[0099] The device monitors ambient light in real time through its built-in ambient light sensor, and maintains a specified illuminance with an adjustable LED supplemental lighting system. The illuminance range is 10-2000 lux. The image sensor collects at least 20 sets of dimensional measurement data under different lighting conditions.

[0100] Calculate the average of the length and width as μ. dim The covariance of the measurement error is calculated as S.

[0101] In one specific embodiment, two 30fps CMOS cameras are installed above the passbook entrance at a 45° angle to capture images of the passbook's edge, extract corner coordinates, obtain deflection angle and centroid offset, and calculate attitude normalization parameters and Mahalanobis distance. When the attitude normalization parameters are greater than a preset probability threshold and the Mahalanobis distance is less than a preset distance threshold, the posture is determined to be "correct".

[0102] In another specific embodiment, the posture information is processed based on a passbook posture recognition algorithm to generate posture normalization parameters and size normalization parameters. The method for determining whether the posture and size normalization parameters conform to a preset physical model standard also includes the following:

[0103] At least two sets of reflective photoelectric sensors are arranged along the passbook conveying direction at the page-turning station to detect whether the left and right edges of the passbook are obstructed, outputting two Boolean values: L (left) and R (right). If L=1 and R=1, both sides of the passbook are within the detection range, and it is determined to be in "correct posture". If either sensor is 0, the passbook is not aligned or exceeds the allowable area, and the service is rejected. The distance between the left and right edge sensors is used as a fixed reference dimension; if all sensors are 1, the passbook size is considered to meet the standard. No additional calculations are required; it is done directly through Boolean logic.

[0104] Specifically, image sensor-based pose recognition methods allow for setting more lenient thresholds, and slight misalignment or size fluctuations can be corrected by software, improving the first-pass yield. However, they require additional light sources and image processing units, resulting in high costs, complex algorithms, sensitivity to strong light or dirt, and more complicated troubleshooting.

[0105] The attitude recognition method based on photoelectric sensors can complete attitude and size determination in milliseconds. It has simple circuits, low cost, and strong resistance to dust interference. However, it can only give a binary result of "pass / fail" and cannot compensate for slight deviations or slight size differences. It will report an error as soon as the user places the object at a slight tilt, which is a bit awkward. Users can choose the corresponding method according to their actual needs.

[0106] Furthermore, the step of generating a matching similarity that characterizes the correlation between the two includes:

[0107] Define the structured data read from the track as D mag The user-entered authentication information is D. user ;

[0108] Two different, pre-defined hash functions are used, including the track hash function f. mag and user input hash function f user ;

[0109] For structured data D respectively mag and authentication information D user The data is processed to generate corresponding fixed-length binary hash codes H. mag and H user H mag =f mag (D mag ,K1),H user =f user (D user K1 and K2 are encryption salts;

[0110] K1 and K2 are dynamically generated in the following manner:

[0111] K1 = HMAC-SHA256(device serial number, key seed for the day);

[0112] K2 = HMAC - SHA256(Session ID, User Operation Timestamp);

[0113] Specifically, the track hash function f mag The processing steps include:

[0114] First, the track format is parsed to obtain structured data D. mag Remove start and end symbols to normalize the data; apply mean filtering to easily demagnetized areas (such as the expiration date field). To suppress noise, let I be the character position within the sliding window, i0-2 be the window start position, i0+2 be the window end position, and i0 be the current character position being processed, where 3 ≤ i0 ≤ I-2; and use cryptographic hash H. mag =BLAKE2s(D′) mag||K1||0xFACE) outputs a 256-bit binary hash code H mag .

[0115] User input hash function f user The optimization process for human input errors includes:

[0116] User input is the authentication information D user Full-width characters are converted to half-width characters, traditional characters to simplified characters, and all letters are standardized to uppercase to normalize the character set; easily confused characters are mapped (e.g., 1→I, 0→O) to improve encoding error tolerance; and entropy-increasing suffixes are generated based on the keystroke time interval Δt to obtain the encoded data D'. user Enhance timing; through cryptographic hash H user =SHA3-256(D') user ||Δt||K2) Outputs a 256-bit binary hash code H user .

[0117] The difference is quantified by calculating the weighted Hamming distance between the two hash codes and then converted into a matching similarity λ. match :

[0118]

[0119] Where N is the total length of the hash code, h mag,i and h user,i The respective track hash codes are H mag and H user The i-th bit, where i is the bit index of the hash code. For the XOR operation, w i These are preset weighting coefficients that assign higher weight to key information bits (such as ID numbers);

[0120] The w i The allocation is made in the following manner: the bit corresponding to the ID number (w) i =2.0, the name corresponds to the bit w i =1.5, other fields correspond to bit positions w i =0.8.

[0121] Furthermore, the step of generating dynamic security thresholds in real time based on the system risk assessment model includes:

[0122] Obtain users' historical transaction behavior information and the current system environment risk index, and analyze the dynamic security threshold T. dyn :

[0123]

[0124] Among them, T baseLet α be a preset baseline security threshold, and α be a transaction behavior adjustment coefficient obtained through a lookup table. The transaction behavior adjustment coefficient has preset values ​​in the database based on user type. For example: 0.5 for users with a history of default, 0.35 for users without default and whose deposits do not exceed the first deposit threshold, 0.25 for users without default and whose deposits exceed the first deposit threshold but not the second, and 0.15 for users without default and whose deposits exceed the second deposit threshold but not the third. freq The frequency of transactions for this user within a specified time window, where N0 is the frequency normalization parameter, and σ is the frequency of transactions for this user within a specified time window. env Environmental risk index;

[0125] tanh is the hyperbolic tangent function, used to smoothly adjust the threshold;

[0126] The frequency normalization parameter is calculated as follows: Where T is the 24-hour time window, t is the current time; σ env An environmental risk index, quantifying the current risk of cyberattacks or fraud and issued in real time by the banking and financial system, is obtained through the following methods:

[0127] Furthermore, when a user selects the supplementary registration service, the steps for executing the corresponding automated business operation process include:

[0128] Retrieve the set of transaction records R1, R2, ..., R to be supplemented from the bank's financial shared information system. n and their corresponding target print page numbers p1, p2, ..., p a ,...,p b ,...,p m m is the total number of pages to be printed, m≥1;

[0129] Establish a comprehensive cost function J(π) for page redirection and printing, and solve for its optimal execution sequence π. * To minimize mechanical operation time and losses, the specific solution formula is as follows:

[0130]

[0131] Where π is the sequence of pages to be printed, i.e., a permutation of length m, p π(g) Let p be the target page number for the g-th page turn in the sequence, where g is the page turn operation number, 1 ≤ g ≤ m-1. π(j) Let C be the target page number for the j-th print operation in the sequence, where j is the print operation number, 1 ≤ j ≤ m. turn (p a ,p b ) is the passbook from the pth a Turn to page pb The mechanical time cost function of a page, where a is the current page and b is the target page, that is, a and b are the specific page number values ​​of the actual page number of the passbook;

[0132] Mechanical time cost function and page number difference |p b -p a |Nonlinear correlation, specifically in the form: C turn (p a ,p b )=t0+γ·|p b -p a | 2 t0 and γ are determined through kinematic calibration experiments at the time of device delivery. t0 is the basic page-turning time, and γ is the page-turning distance penalty coefficient (unit: ms / page²). Different page number differences (Δp = |p_0|) are applied to the device using a standard passbook. b -p a The page-turning operation (|) is used to measure the actual page-turning time TΔ. p Record the page number difference versus time measurement data, and perform a quadratic polynomial regression on the measurement data: TΔ p =t0 + γ·(Δp) 2 γ is solved using the least squares method and written into the device firmware, C print (p π(j) () represents the cost of time required to print the target page number in the j-th print in the sequence;

[0133] The specific form is as follows: Where A page A represents the total area of ​​the page. blank (p k ) is the pth k The area of ​​blank space on the page, v print The printer's rated speed;

[0134] Based on the optimal execution sequence π * Perform the page turning and printing operations in sequence.

[0135] Furthermore, when a user selects the new account opening service, the steps for executing the corresponding automated business operation process include:

[0136] The user's identity is verified, then the new book is issued, the book is pushed to the page-turning area, and the first page is turned;

[0137] If page turning fails, the waste book recycling process will be initiated;

[0138] If the page is turned successfully, the adhesive silicone roller will feed the paper and separate the pages. The roller will then feed the book into the magnetic printing module and print the relevant information into the passbook's magnetic track.

[0139] After the magnetic stripe information is entered into the new passbook, a magnetic stripe verification operation is immediately performed to obtain the waveform of the read signal;

[0140] By performing a Fourier transform on the read-back signal, the effective power P of the signal can be separated. signal With background noise power P noise And calculate the signal-to-noise ratio ρ of the brushing quality. snr ;

[0141]

[0142] Where S(f) is the readback signal in the data frequency band [f low ,f high The effective frequency band [f] is the spectrum within the same frequency band, where N(f) is the noise spectrum within the same frequency band; N(f) is obtained by acquiring the background noise signal 200ms before each brushing operation and calculating its spectrum; low ,f high According to the track coding standard: for F2F coding, f low =0.8f base f high =1.2f base , where f base The fundamental frequency for track encoding;

[0143] If the signal-to-noise ratio ρ snr Below the preset quality standard threshold τ snr If the magnetic stripping fails, the waste recycling process will be initiated.

[0144] If the magnetic stripe is successfully scanned, the account opening information on the first page will be printed. After printing, the account will be sent to the exit by the institution for the user to pick up.

[0145] Furthermore, the initiation of the waste recycling process includes:

[0146] When it is determined that the magnetic stripe printing failed or the new book page turning failed, the roller of the passbook delivery mechanism is controlled to roll in reverse, and the new passbook that failed to be made is delivered to a dedicated waste book recycling bin that is physically isolated from the old book recycling bin.

[0147] After the scrapped units are recycled, the system automatically initializes all mechanical mechanisms to standby mode and restarts the issuance of new units, without requiring user intervention.

[0148] Furthermore, when a user selects the passbook renewal service, the steps for executing the corresponding automated business operation process include:

[0149] After successfully reading the magnetic track information and verifying the identity of the old passbook, the old passbook is transported to the designated old passbook recycling bin, which is physically isolated from the old passbook recycling bin, by controlling the mechanical structure.

[0150] Subsequently, the new account opening process was seamlessly integrated, and a new passbook was issued and created for the user.

[0151] like Figure 2 As shown, this application provides a self-service storage and interaction system based on shared information of a banking and financial system, which is used to implement the self-service storage and interaction method based on shared information of a banking and financial system, including: a posture size parameter generation module, a posture verification module, an identity verification information generation module, and an identity verification module;

[0152] The posture and size parameter generation module is used to acquire and monitor the posture information of the user placing the passbook into the self-service device, and process the posture information based on the passbook posture recognition algorithm to generate posture normalization parameters and size normalization parameters.

[0153] The posture verification module is used to determine whether the posture and size standardization parameters conform to the preset physical model standard. If they conform, the page turning and magnetic track information reading process is started to obtain the existing magnetic track data in the passbook. If they do not conform, the service is rejected and the passbook is exited.

[0154] The authentication information generation module is used to compare the read track data with the authentication information input by the user through the interactive interface, based on an asymmetric cross-modal hash matching algorithm, to generate a matching similarity that characterizes the correlation between the two.

[0155] The identity verification module is used for magnetic stripe information comparison verification: a dynamic security threshold is generated in real time according to the system risk assessment model. When the matching similarity is higher than the dynamic security threshold, the verification is successful. Based on the specific business type selected by the user, the account data in the bank's financial shared information system is called to execute the corresponding automated business operation process. When the matching similarity is not higher than the dynamic security threshold, the verification fails, the business is rejected, and the passbook is exited. The specific business types include: supplementary registration business, new account opening business, and passbook replacement business.

[0156] For supplementary registration, the process involves obtaining the transaction records to be supplemented and their corresponding target print page numbers from the bank's financial shared information system, analyzing the overall cost to determine the optimal execution sequence, and then executing it.

[0157] In this embodiment, after obtaining the transaction records to be updated and their corresponding target print page numbers from the bank's financial shared information system, the distribution of these transaction records in the passbook is analyzed first, rather than directly flipping through each page starting from the first page. A comprehensive cost function is constructed, including page-turning cost and printing cost, where page-turning cost is non-linearly related to the page number difference, and printing cost is related to the specific position of the print page. The optimal execution sequence of this comprehensive cost function is solved to find a page-turning and printing order that minimizes mechanical operation time and wear. The page-turning and printing operations are executed sequentially according to the optimal sequence, reducing unnecessary page-turning times and mechanical movement distances, improving update efficiency, and reducing equipment wear.

[0158] In summary, the embodiments of this application achieve efficient execution of supplementary registration services through an intelligent supplementary registration scheduling strategy based on comprehensive cost optimization, reducing mechanical wear and tear, transforming the supplementary registration task into a combinatorial optimization problem, and calculating the optimal page turning and printing order through algorithms to reduce mechanical movement distance and time consumption. This solves the problems of low efficiency and rapid mechanical wear in traditional passbook self-service equipment supplementary registration services, and extends the service life of the equipment.

[0159] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0160] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0161] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0162] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0163] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0164] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A self-service storage and interaction method based on shared information in a banking and financial system, characterized in that: Includes the following steps: S1, acquire and monitor the posture information of the user placing the passbook into the self-service device, process the posture information based on the passbook posture recognition algorithm, and generate posture normalization parameters and size normalization parameters. S2, determine whether the posture and size standardization parameters conform to the preset physical model standard. If they do, start the page turning and magnetic track information reading process to obtain the existing magnetic track data in the passbook. If they do not conform, reject the service and exit the passbook. S3, compare the read track data with the authentication information entered by the user through the interactive interface, and generate a matching similarity that represents the correlation between the two based on an asymmetric cross-modal hash matching algorithm; S4, Track Information Comparison and Verification: A dynamic security threshold is generated in real time based on the system risk assessment model. When the matching similarity is higher than the dynamic security threshold, the verification is successful. Based on the specific business type selected by the user, the account data in the bank's financial shared information system is called to execute the corresponding automated business operation process. When the matching similarity is not higher than the dynamic security threshold, the verification fails, the business is rejected, and the passbook is exited. The specific business types include: supplementary registration business, new account opening business, and passbook replacement business. For supplementary registration, the process involves obtaining the transaction records to be supplemented and their corresponding target print page numbers from the bank's financial shared information system, analyzing the overall cost to determine the optimal execution sequence, and then executing it.

2. The self-service storage and interaction method based on shared information in a banking financial system as described in claim 1, characterized in that, The steps for generating the attitude normalization parameters include: At least two image sensors at different angles are used to acquire attitude images of the passbook, and a geometric feature vector V containing the passbook's deflection angle and centroid coordinate offset is extracted. img ; For the geometric feature vector V img Probability density estimation is performed to calculate the posterior probability of belonging to the correct pose category, and this probability is used as the pose normalization parameter P. pose : Where K is the preset number of Gaussian components in the Gaussian mixture model, φ k Let N(V) be the weight of the k-th Gaussian component. img |μ k ,Σ k ) indicates that the mean is μ k The covariance matrix is ​​Σ k Gaussian distribution on geometric eigenvector V img The probability density at that location.

3. The self-service storage and interaction method based on shared information in a banking financial system as described in claim 2, characterized in that, The step of generating the size normalization parameters further includes: The length l and width w of the passbook are measured from the posture information to form a size measurement vector V. dim =[l,w] T ; Calculate the size measurement vector V dim With standard size vector μ dim =[l std ,w std ] T Mahalanobis distance D between M And use this distance as a size normalization parameter: Where S is the covariance matrix of the passbook size measurement, S -1 Let S be the inverse matrix; The criterion is that when the attitude normalization parameter is greater than the preset probability threshold and the Mahalanobis distance is less than the preset distance threshold, it is determined to meet the physical model standard.

4. The self-service storage and interaction method based on shared information in a banking financial system as described in claim 1, characterized in that, The step of generating a matching similarity that characterizes the correlation between the two includes: Define the structured data read from the track as D mag The user-entered authentication information is D. user ; Two different, pre-defined hash functions are used, including the track hash function f. mag and user input hash function f user ; For structured data D respectively mag and authentication information D user The data is processed to generate corresponding fixed-length binary hash codes H. mag and H user H mag =f mag (D mag ,K1),H user =f user (D user K1 and K2 are encryption salts; K1 and K2 are dynamically generated in the following manner: K1 = HMAC-SHA256(device serial number, key seed for the day); K2 = HMAC - SHA256(Session ID, User Operation Timestamp); The difference is quantified by calculating the weighted Hamming distance between the two hash codes and then converted into matching similarity.

5. The self-service storage and interaction method based on shared information in a banking financial system as described in claim 1, characterized in that, The step of generating dynamic security thresholds in real time based on the system risk assessment model includes: Obtain users' historical transaction behavior information and the current system environment risk index, and analyze the dynamic security threshold T. dyn : Among them, T base Let N be a preset baseline safety threshold, α be a transaction behavior adjustment coefficient, and N be a threshold value. freq The frequency of transactions for this user within a specified time window, where N0 is the frequency normalization parameter, and σ is the frequency of transactions for this user within a specified time window. env Environmental risk index; tanh is the hyperbolic tangent function, used to smoothly adjust the threshold; σ env Obtain it through the following methods:

6. The self-service storage and interaction method based on shared information in a banking and financial system as described in claim 1, characterized in that, When a user selects to complete the registration retroactively, the steps for executing the corresponding automated business operation process include: Retrieve the set of transaction records to be supplemented and their corresponding target print page numbers p1, p2, ..., p from the bank's financial shared information system. a ,...,p b ,...,p m m is the total number of pages to be printed, m≥1; Establish a comprehensive cost function J(π) for page redirection and printing, and solve for its optimal execution sequence π. * The specific solution formula is as follows: Where π is the sequence of pages to be printed, p π(g) p is the target page number for the g-th page turn in the sequence, where g is the page turn operation number. π(j) Let C be the target page number for the j-th print operation in the sequence, where j is the print operation number. turn (p a ,p b ) is the passbook from the pth a Turn to page p b The mechanical time cost function of a page, where a is the current page, b is the target page, and C is the target page. print (p π(j) () represents the cost of time required to print the target page number in the j-th print in the sequence; Based on the optimal execution sequence π * Perform the page turning and printing operations in sequence.

7. The self-service storage and interaction method based on shared information in a banking and financial system as described in claim 1, characterized in that, When a user selects the new account opening service, the steps for executing the corresponding automated business operation process include: The user's identity is verified, then the new book is issued, the book is pushed to the page-turning area, and the first page is turned; If page turning fails, the waste book recycling process will be initiated; If the page is turned successfully, the adhesive silicone roller will feed the paper and separate the pages. The roller will then feed the book into the magnetic printing module and print the relevant information into the passbook's magnetic track. After the magnetic stripe information is entered into the new passbook, a magnetic stripe verification operation is immediately performed to obtain the waveform of the read signal; By performing a Fourier transform on the read-back signal, the effective power of the signal and the background noise power are separated, and the signal-to-noise ratio of the brushing quality is calculated. If the signal-to-noise ratio is lower than the preset quality standard threshold, the magnetic stripping is deemed to have failed, and the waste paper recycling process is initiated. If the magnetic stripe is successfully scanned, the account opening information on the first page will be printed. After printing, the account will be sent to the exit by the institution for the user to pick up.

8. The self-service storage and interaction method based on shared information in a banking financial system as described in claim 7, characterized in that, The process for initiating waste recycling includes: When it is determined that the magnetic stripe printing failed or the new book page turning failed, the roller of the passbook delivery mechanism is controlled to roll in reverse, and the new passbook that failed to be made is delivered to a dedicated waste book recycling bin that is physically isolated from the old book recycling bin. After the scrapped units are recycled, the system automatically initializes all mechanical mechanisms to standby mode and restarts the issuance of new units, without requiring user intervention.

9. The self-service storage and interaction method based on shared information in a banking financial system as described in claim 7, characterized in that, When a user selects the passbook renewal service, the steps for executing the corresponding automated business operation process include: After successfully reading the magnetic track information and verifying the identity of the old passbook, the old passbook is transported to the designated old passbook recycling bin, which is physically isolated from the old passbook recycling bin, by controlling the mechanical structure. Subsequently, the new account opening process was seamlessly integrated, and a new passbook was issued and created for the user.

10. A self-service storage and interaction system based on shared information in a banking and financial system, used to implement the self-service storage and interaction method based on shared information in a banking and financial system as described in any one of claims 1-9, characterized in that, include: Attitude dimension parameter generation module, attitude verification module, identity verification information generation module, identity verification module; The posture and size parameter generation module is used to acquire and monitor the posture information of the user placing the passbook into the self-service device, and process the posture information based on the passbook posture recognition algorithm to generate posture normalization parameters and size normalization parameters. The posture verification module is used to determine whether the posture and size standardization parameters conform to the preset physical model standard. If they conform, the page turning and magnetic track information reading process is started to obtain the existing magnetic track data in the passbook. If they do not conform, the service is rejected and the passbook is exited. The authentication information generation module is used to compare the read track data with the authentication information input by the user through the interactive interface, based on an asymmetric cross-modal hash matching algorithm, to generate a matching similarity that characterizes the correlation between the two. The identity verification module is used for magnetic stripe information comparison verification: a dynamic security threshold is generated in real time according to the system risk assessment model. When the matching similarity is higher than the dynamic security threshold, the verification is successful. Based on the specific business type selected by the user, the account data in the bank's financial shared information system is called to execute the corresponding automated business operation process. When the matching similarity is not higher than the dynamic security threshold, the verification fails, the business is rejected, and the passbook is exited. The specific business types include: supplementary registration business, new account opening business, and passbook replacement business. For supplementary registration, the process involves obtaining the transaction records to be supplemented and their corresponding target print page numbers from the bank's financial shared information system, analyzing the overall cost to determine the optimal execution sequence, and then executing it.

Images