An information security protection method in an internet financial service development process

Abstract

The application relates to the technical field of information security protection, in particular to an information security protection method in the process of carrying out an internet financial service, which comprises the following steps: acquiring a fingerprint image of a user, encoding various information categories of historical data of the user when the user carries out the internet financial service for multiple times to form a service information vector of each information category; acquiring each feature point; obtaining a feature evaluation value; screening each significant point; clustering all the significant points; obtaining information effectiveness, biological information coding and change significance; acquiring each feature category; obtaining a key by using a threshold secret sharing scheme; and protecting user information generated in the process of carrying out the internet financial service by using an encryption algorithm. The application avoids the disadvantage that user biological feature information is leaked and user information is cracked, and guarantees the information security of the user in the process of carrying out the internet financial service.

Description

Technical Field

[0001] This application relates to the field of information security protection technology, specifically to an information security protection method in the process of conducting internet finance business. Background Technology

[0002] With the rapid development of digital transformation and the internet finance industry, diversified financial businesses have emerged. The development of internet finance businesses not only involves a large amount of financial transaction information and user data, but also, due to their own openness and complexity, the security of user financial information has become one of the key obstacles to the development of internet finance businesses. In order to ensure the healthy and safe operation of the internet finance system and protect users' personal privacy and transaction data, information security protection measures are of paramount importance.

[0003] Because user biometrics are unique and stable, traditional methods that generate keys based on user biometrics have drawbacks such as insufficient accuracy and risk of leakage. Once a user's biometric information is leaked, it can lead to the cracking of the user's financial business information, and the user's information security cannot be guaranteed. Summary of the Invention

[0004] To address the aforementioned technical issues, this paper provides an information security protection method for the conduct of internet finance business, thereby resolving the existing problems.

[0005] The solution to the technical problem in this application is to provide an information security protection method during the operation of internet finance business, including the following steps:

[0006] The user's fingerprint image is obtained and encoded using historical data of various information categories from the user's multiple internet finance transactions to form a business information vector for each information category.

[0007] The edge contours in the fingerprint image are extracted and denoted as each ridge. Based on the distribution of the number of edge pixels in the neighborhood of each ridge, the edge pixels are filtered to obtain each feature point.

[0008] By taking into account the number of edge pixels on the ridge where each feature point is located, the distribution of the number of edge pixels in the neighborhood of each feature point, and the grayscale difference between each feature point and its neighborhood pixels, the feature evaluation value of each feature point is obtained.

[0009] Feature points are filtered based on the feature evaluation values ​​to obtain salient points; all salient points are clustered; the discrete distribution between each salient point and other salient points in the same cluster is analyzed, and the information validity of each salient point is determined in combination with the feature evaluation values. The information validity is encoded to obtain the bioinformatics code of each salient point.

[0010] Analyze the frequency characteristics of identical elements in the business information vector and the differences between different elements to determine the significance of changes in each information category. Filter the information categories to obtain each feature category. Using a threshold secret sharing scheme, combine the significance of changes in all feature categories to obtain the original secret. Shift and transcode the biometric information encoding to obtain the key. Use an encryption algorithm to encrypt and protect user information generated during the conduct of internet finance business.

[0011] Preferably, the further acquisition process of the business information vector for each information category is as follows:

[0012] The historical data of each information category for each user's financial transaction is converted into binary code. According to the time sequence of the user's financial transaction, all the binary codes of each information category are combined to form the business information vector of each information category.

[0013] Preferably, the further acquisition process of each feature point is as follows:

[0014] The fingerprint image is refined by refining all edge contours and denoted as each ridge. If any edge pixel on each ridge has three or more edge pixels in its eight-neighborhood, then that edge pixel is denoted as a feature point.

[0015] Preferably, obtaining the feature evaluation value of each feature point includes:

[0016] The number of all edge pixels in the eight neighborhood of each feature point on each ridge in the fingerprint image is used as the cross-connection factor of each feature point.

[0017] The sum of the differences in gray values ​​between each feature point and all pixels in its eight neighborhoods is denoted as the relative gray value difference.

[0018] Count the number of all edge pixels on the ridge line where each feature point is located in the fingerprint image;

[0019] The feature evaluation value is the normalized result of the product of the number, the cross-connection factor, and the relative gray level difference.

[0020] Preferably, the further acquisition process of each salient point is as follows:

[0021] A segmentation threshold is obtained for the feature evaluation values ​​of all feature points in the fingerprint image, denoted as the first segmentation threshold; feature points whose feature evaluation values ​​are greater than or equal to the first segmentation threshold are denoted as salient points.

[0022] Preferably, the clustering of all salient points includes:

[0023] Each salient point is treated as a graph node, and there is an edge between any two graph nodes. The distance between the coordinates of any two salient points in the fingerprint image is used as the edge weight between the corresponding two graph nodes to construct a weighted undirected graph.

[0024] Cluster the weighted undirected graph to obtain multiple clusters.

[0025] Preferably, determining the information validity of each salient point includes:

[0026] Calculate the Shannon entropy of each salient point and the position coordinates of all other salient points in its cluster;

[0027] The validity of the information is the normalized result of the product of the feature evaluation value of each salient point and the Shannon entropy.

[0028] Preferably, determining the significance of the change for each information category includes:

[0029] Count the number of times the same binary code appears in the business information vector for each information category; use the ratio of the highest frequency to the total number of binary codes in the business information vector as the concentration of user behavior for each information category.

[0030] Calculate the sum of edit distances between any two binary codes in the business information vector for each information category;

[0031] The significance of the change is the floor result of the product of the sum and the concentration of user behavior.

[0032] Preferably, the further acquisition process of each feature category is as follows:

[0033] Obtain the segmentation threshold of the change significance for all information categories, and denote it as the second segmentation threshold; denote the information categories whose change significance is greater than the second segmentation threshold as feature categories.

[0034] Preferably, the method for obtaining the key is as follows:

[0035] A polynomial function is constructed using the (t, n) threshold secret sharing scheme. Based on the change significance of each feature category, the secret of the (t, n) threshold secret sharing scheme is allocated and recovered to restore the original secret.

[0036] Using the original secret as the number of cyclic shifts, the bioinformatics codes of all salient points are cyclically shifted, and the bioinformatics codes of all salient points after cyclic shifts are combined to form a bioinformatics shift sequence.

[0037] The bioinformatics encoding of each significant point after cyclic shift is converted into a decimal number. The sum of the decimal number and all coefficients of the polynomial function is calculated. The value of each digit of the sum is used as a seed and input into the pseudo-random number generator to output a pseudo-random number sequence.

[0038] The pseudo-random sequences corresponding to all salient points are arranged in the order of the salient points in the bio-information shift sequence to form the key of the encryption algorithm.

[0039] This application has at least the following beneficial effects:

[0040] This application filters edge pixels in a user's fingerprint image to obtain feature points and their feature evaluation values. The advantage lies in identifying locations with richer fingerprint details and evaluating the fingerprint details contained within these feature points, thus further filtering to obtain salient points. This process effectively identifies salient points containing biometric details, providing more reliable biometric information for key generation. Secondly, a weighted undirected graph is constructed to cluster all salient points, determining the information validity of each salient point. This considers the stability of the biometric information contained in salient points, reducing reliance on individual salient points and mitigating the risk of low key security due to the loss of biometric features from a single salient point. Finally, the application determines the information category... The change saliency method filters all information categories and obtains each feature category. Its beneficial effect lies in considering the user's transaction behavior characteristics when conducting internet financial business. It reflects the user's transaction behavior through the dynamic saliency characteristics of changes in the user's financial business information. It combines the user's biometric information with the user's transaction behavior characteristics to generate a key. The encryption algorithm is used to encrypt and protect the user information generated during the conduct of internet financial business. Its beneficial effect is that the user's transaction behavior will change at different times, and the generated key will also change. By combining the user's biometric information and the user's financial transaction behavior characteristics to generate a key, it avoids the drawback of the user's information being cracked after the user's biometric information is leaked. It increases the multi-factor complexity of the key and can ensure the information security of users during the conduct of internet financial business. Attached Figure Description

[0041] The following section, in conjunction with the accompanying drawings, provides a more detailed explanation of an information security protection method for conducting internet finance business, as described in this application.

[0042] Figure 1 A flowchart illustrating the steps of an information security protection method during the conduct of internet finance business, as provided in this application embodiment;

[0043] Figure 2 A flowchart illustrating the steps of a method for obtaining the validity of information about each significant point provided in an embodiment of this application. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description, in conjunction with the accompanying drawings and implementation examples, provides a method for information security protection during the conduct of internet finance business. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0045] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0046] Please see Figure 1 The document illustrates a flowchart of an information security protection method for conducting internet finance business, according to an embodiment of this application. The method includes the following steps:

[0047] Step 1: Obtain the user's fingerprint image and encode it using historical data of various information categories from the user's multiple internet finance transactions to form a business information vector for each information category.

[0048] When a user conducts internet financial transactions, the fingerprint information of the user is obtained through a fingerprint sensor and converted into an image to obtain the user's fingerprint image. In order to prevent the fingerprint image from being severely affected by noise interference and thus affecting subsequent analysis, the fingerprint image is denoised.

[0049] In this embodiment, the median filtering algorithm is used to denoise the fingerprint image. The median filtering algorithm is a well-known technology and will not be described in detail here. As another implementation method, the implementer may use other methods such as the mean filtering algorithm. This embodiment does not impose any special restrictions on this.

[0050] Secondly, obtain various types of financial information data from the internet finance management platform when users conduct internet finance business each time in a historical period, record them as each information category, and obtain historical data of each information category when users conduct financial business each time.

[0051] In this embodiment, financial information data of the user during each internet finance transaction within the past two months is obtained from the internet finance management platform. Secondly, the financial information data includes: transaction amount, transaction date, transaction channel, deposit information, wealth management product amount, fund flow, wealth management profit, credit score, loan term, repayment record, etc.

[0052] The historical data of each information category when a user conducts financial transactions are converted into binary codes. According to the time sequence of the user's financial transactions, all the binary codes of each information category are combined to form a business information vector for each information category.

[0053] In this embodiment, the historical data is converted into binary code using the UTF-8 encoding method. The UTF-8 encoding method is a well-known technology and will not be described in detail here.

[0054] At this point, we have obtained the user's fingerprint image and the business information vector for each information category.

[0055] Step 2: Extract the edge contours from the fingerprint image and denote them as each ridge. Based on the distribution of the number of edge pixels in the neighborhood of each ridge, filter the edge pixels to obtain each feature point. By combining the number of edge pixels on the ridge where each feature point is located, the distribution of the number of edge pixels in the neighborhood of each feature point, and the grayscale difference between each feature point and its neighborhood pixels, obtain the feature evaluation value of each feature point.

[0056] Traditional fingerprint image feature analysis typically involves extracting ridges or valleys from the fingerprint image. However, this method has certain drawbacks. When a user's fingerprint experiences natural wear, aging, or injury, the ridges in the fingerprint image may become blurred or broken, or even the feature information of the ridges and valleys may be missing. This results in the inability to effectively extract biometric information, leading to reduced security and insufficient reliability in the subsequent process of combining biometric information with the user's financial transaction behavior to generate keys.

[0057] When generating a key by combining a user's biometric information with their financial transaction behavior, the selected biometric information should possess higher uniqueness and stability. Specifically, a higher contrast between the feature point in the fingerprint image and its surrounding environment indicates that the feature point provides richer, more detailed features, making it more difficult for attackers to recover the complete fingerprint image using only partial features. Furthermore, stronger integrity of the ridges containing the feature point indicates a less significant impact of biometric loss on key generation, and the high information entropy of dense feature points enables the generation of more secure keys.

[0058] Therefore, by extracting the edge contours from the fingerprint image and analyzing the distribution characteristics of the number of edge pixels in the local neighborhood of each edge pixel on the edge contour, each feature point is obtained, specifically:

[0059] An edge detection algorithm is used to extract all edge contours in the fingerprint image and refine all edge contours, which are then denoted as each ridge line.

[0060] In this embodiment, the Canny edge detection algorithm is used to extract all edge contours in the fingerprint image. The edge erosion method is then used to refine all edge contours to the width of one pixel. The Canny edge detection algorithm and the edge erosion method are well-known technologies and will not be described in detail here. As other implementation methods, implementers can use other methods of existing technology, such as the Sobel operator for edge detection. This embodiment does not impose any special restrictions on this.

[0061] If any edge pixel on each ridge line has three or more edge pixels in its eight-neighborhood, then any edge pixel is recorded as a feature point.

[0062] It should be noted that if any edge pixel has three or more edge pixels in its eight-neighborhood, it means that the edge pixel is located at the connection point between different ridges in the fingerprint image.

[0063] Furthermore, by analyzing the grayscale differences between each feature point and its eight neighboring pixels, as well as the number of edge pixels within each feature point's eight neighboring pixels, a feature evaluation value is calculated to reflect the uniqueness of the biometric information at each feature point's location. Specifically:

[0064] The number of all edge pixels in the eight neighborhood of each feature point on each ridge in the fingerprint image is used as the cross-connection factor of each feature point.

[0065] The sum of the differences in gray values ​​between each feature point and all pixels in its eight neighborhoods is denoted as the relative gray value difference.

[0066] In this embodiment, the sum of the absolute values ​​of the differences in grayscale values ​​between each feature point and all pixels in its eight neighborhoods is denoted as the relative grayscale difference.

[0067] The number of all edge pixels on the ridge where each feature point is located in the fingerprint image is counted. The normalized result of the product of the number, the cross-connection factor, and the relative gray level difference is used as the feature evaluation value of each feature point in the fingerprint image.

[0068] In this embodiment, the sigmoid function is used for normalization. The sigmoid function is a well-known technique and will not be described in detail here. As another implementation method, the implementer may use other methods of the prior art, such as the tanh function. This embodiment does not impose any special restrictions on this.

[0069] It should be noted that the cross-connection factor reflects the complexity of the ridge structure within the neighborhood of the feature point. The larger the cross-connection factor, the richer the fingerprint details at the location of the corresponding feature point. Secondly, the relative grayscale difference reflects the difference between the grayscale value of the feature point and the grayscale values ​​of the surrounding pixels. The larger the relative grayscale difference, the more different the fingerprint feature information contained in the corresponding feature point and the other pixels in its neighborhood. The number reflects the completeness of the ridge where the feature point is located. The larger the obtained feature evaluation value, the higher the uniqueness of the corresponding feature point, the richer the fingerprint details it contains, and the more reliable biometric information it can provide for key generation.

[0070] Thus, the feature evaluation values ​​of each feature point in the fingerprint image are obtained.

[0071] Step 3: Based on the feature evaluation value, the feature points are filtered to obtain each salient point; all salient points are clustered; the discrete distribution between each salient point and other salient points in the same cluster is analyzed, and the information validity of each salient point is determined in combination with the feature evaluation value. The information validity is encoded to obtain the bioinformatics code of each salient point.

[0072] The flowchart of the method for obtaining the validity of information about each salient point provided in the embodiments of this application is as follows: Figure 2 As shown.

[0073] First, based on the feature evaluation values, feature points are filtered to obtain salient points, specifically as follows:

[0074] A segmentation threshold is obtained for the feature evaluation values ​​of all feature points in the fingerprint image, denoted as the first segmentation threshold; feature points whose feature evaluation values ​​are greater than or equal to the first segmentation threshold are denoted as salient points.

[0075] In this embodiment, the method for obtaining the first segmentation threshold is as follows: calculating the average and standard deviation of the feature evaluation values ​​of all feature points in the fingerprint image; based on the average and standard deviation, obtaining the distribution interval using the Laida criterion; using the right endpoint of the distribution interval as the first segmentation threshold, wherein the Laida criterion is a well-known technique and will not be described in detail here; assuming that the average of the feature evaluation values ​​of all feature points in the fingerprint image is... Standard deviation is The distribution interval is then obtained as follows: Then the first segmentation threshold is .

[0076] It should be noted that the location of the salient point contains a significant amount of biological detail feature information.

[0077] Secondly, clustering is performed on the salient points, specifically as follows:

[0078] Each salient point is treated as a graph node, and there is an edge between any two graph nodes. The distance between the coordinates of any two salient points in the fingerprint image is used as the edge weight between the corresponding two graph nodes to construct a weighted undirected graph.

[0079] Cluster the weighted undirected graph to obtain multiple clusters;

[0080] In this embodiment, the Markov Cluster Algorithm (MCL) is used to cluster the weighted undirected graph. The Markov Cluster Algorithm is a well-known technique and will not be described in detail here.

[0081] It should be noted that by constructing a weighted undirected graph of salient points, the spatial distribution characteristics of feature points with significant biometric information in the fingerprint image are captured. This allows for the subsequent establishment of a connection between salient points and global biometric information. By leveraging the spatial distribution relationship among multiple salient points, the dependence on a single salient point is reduced, thereby mitigating the risk of low security in the generated key due to the loss of biometric information contained in a single salient point. For example, if the fingerprint at the location corresponding to a certain salient point is worn away and becomes unstable, but other salient points of the same class contain similar biometric information, the biometric information of these other similar salient points can be used to identify the fingerprint, thus reducing the risk of low security in the generated key due to the loss of biometric information contained in a single salient point.

[0082] Furthermore, in a weighted undirected graph, the denser the distribution of different salient points, the more complex the texture structure of the region containing the salient points, and the more stable and effective the user's biometric information. Therefore, by analyzing the distribution dispersion of salient points within the same cluster and combining it with the aforementioned feature evaluation values, the effectiveness of the information is calculated to reflect the effectiveness and stability of the biometric information contained in the corresponding salient points. Specifically:

[0083] Calculate the Shannon entropy of each salient point and the position coordinates of all other salient points in its cluster;

[0084] It should be noted that the calculation of Shannon entropy is a well-known technique and will not be elaborated upon here.

[0085] The normalized result of the product of the feature evaluation value and the Shannon entropy of each salient point is taken as the information validity of each salient point;

[0086] In this embodiment, the sigmoid function is used for normalization. The sigmoid function is a well-known technique and will not be described in detail here. As another implementation method, the implementer may use other methods of the prior art, such as the tanh function. This embodiment does not impose any special restrictions on this.

[0087] It should be noted that in fingerprint images, the richer the fingerprint details provided by salient points, the less impact there is on the loss of biometric features during subsequent key generation. Furthermore, the denser and more uniform the distribution of salient points within their neighborhood, the greater the Shannon entropy. This indicates that the texture structure of the region where the salient point is located is more complex, and the biometric information provided is more stable and effective. Therefore, a complex and uniform distribution means that the relative positional relationships between salient points are richer, providing more detailed information and thus improving the accuracy and reliability of fingerprint recognition. The greater the effectiveness of the obtained information, the higher the effectiveness of the biometric information that the corresponding salient point can represent.

[0088] Furthermore, based on the validity of the information, it is encoded to obtain the corresponding binary code, specifically as follows:

[0089] The validity of the information of each salient point in the fingerprint image is converted into binary code, which is denoted as the biometric code of each salient point.

[0090] In this embodiment, the UTF-8 encoding method is used to convert the validity of the information of each significant point in the fingerprint image into binary code. The UTF-8 encoding method is a well-known technology and will not be described in detail here.

[0091] Thus, the biometric encoding of each salient point in the fingerprint image is obtained.

[0092] Step 4: Analyze the frequency characteristics of the same elements in the business information vector and the differences between different elements to determine the significance of changes in each information category. Filter the information categories to obtain each feature category. Using a threshold secret sharing scheme, combine the significance of changes in all feature categories to obtain the original secret. Shift and transcode the biometric information code to obtain the key. Use an encryption algorithm to encrypt and protect the user information generated during the Internet finance business.

[0093] In the course of internet finance business, user transaction behavior is highly dynamic and flexible. Based on user transaction behavior during the development of internet finance business, the (t, n) threshold secret sharing scheme is adopted to generate keys. This can enhance the multi-factor complexity of the keys, making them more difficult for attackers to crack. Moreover, as user transaction behavior changes over a period of time, the keys can be updated in a timely manner to ensure user information security. It should be noted that the (t, n) threshold secret sharing scheme is a well-known technology and will not be elaborated here.

[0094] First, by analyzing the business information vectors of each information category of a user during the course of internet finance business operations, the significance of changes in each information category is calculated to reflect the concentration of user's internet finance business behavior or attributes corresponding to that information category. This illustrates the information transaction characteristics of the user's internet finance business under this information category, specifically:

[0095] Count the number of times the same binary code appears in the business information vector for each information category; use the ratio of the highest frequency to the total number of binary codes in the business information vector as the concentration of user behavior for each information category.

[0096] It should be noted that the greater the concentration of user behavior, the more concentrated the user's internet finance business transaction behavior or attributes are under that information category, and the better it reflects the transaction characteristics of the user's internet finance business under that information category.

[0097] Calculate the sum of edit distances between any two binary codes in the business information vector for each information category; take the floor result of the product of the sum and the user behavior concentration as the change significance for each information category.

[0098] In this embodiment, the edit distance is measured by the ED (Edit Distance) edit distance, which is a well-known technology and will not be described in detail here. As other implementations, implementers may use other methods of the prior art, such as Levenshtein edit distance, etc. This embodiment does not impose any special restrictions on this.

[0099] It should be noted that the larger the sum, the greater the difference between the two binary codes. The more significant the changes in the user's business information under this information category, the greater the fluctuation in the user's financial information data under this information category, i.e., the stronger the dynamics of the transaction behavior. The greater the significance of the changes, the more significant the changes in the user's business information under this information category and the higher the concentration of transaction behavior. This can more effectively represent the user's transaction behavior characteristics and has a higher value for key generation. The key generated through this information category enhances the complexity of the key, making it more difficult for attackers to crack.

[0100] The threshold for determining the significance of changes across all categories of user information is denoted as the second segmentation threshold.

[0101] In this embodiment, the Otsu threshold segmentation algorithm is used to obtain the segmentation threshold for the significance of the change in all information categories. The Otsu threshold segmentation algorithm is a well-known technology and will not be described in detail here.

[0102] The information categories whose change significance is greater than the second segmentation threshold are denoted as feature categories;

[0103] Furthermore, the (t, n) threshold secret sharing scheme is a cryptographic technique used to divide a secret S into n shares, requiring at least t shares to recover the original secret S. This scheme ensures that even if some shares of information are leaked, the original secret cannot be recovered as long as fewer than t shares are needed, thus improving the security of the secret.

[0104] Therefore, based on the significance of the change for each feature category, a random number generated by a pseudo-random number generator is obtained using a polynomial function constructed using the (t, n) threshold secret sharing scheme, and the key for the encryption algorithm is finally determined, specifically including:

[0105] The polynomial function constructed by the (t, n) threshold secret sharing scheme is: Where S is a random number generated by a pseudo-random number generator. The coefficients are randomly selected.

[0106] It should be noted that each share is a point of a polynomial function. The random number S represents the shared secret.

[0107] Substitute the significance of the changes in all feature categories into the polynomial function to obtain multiple points. At least t points must be selected to recover the random number S.

[0108] Using a random number S as the cyclic shift number, the bioinformatics codes of all salient points are cyclically shifted, and the cyclically shifted bioinformatics codes of all salient points are combined to form a bioinformatics shift sequence.

[0109] It should be noted that the cyclic shift process is a well-known technique and will not be described in detail here.

[0110] The bioinformatics encoding after cyclic shifting of each significant point is converted into a decimal number. The sum of all randomly selected coefficients in the polynomial function constructed by the (t, n) threshold secret sharing scheme is calculated. The value of each digit of the sum is used as a seed and input into the pseudo-random number generator to output a pseudo-random number sequence.

[0111] It should be noted that pseudo-random number generators are a well-known technology and will not be elaborated upon here.

[0112] The pseudo-random sequences corresponding to all salient points are arranged in the order of the salient points in the bio-information shift sequence to form the final key; the final key is used as the key of the encryption algorithm to encrypt the user's financial information data during the Internet finance business.

[0113] In this embodiment, the AES (Advanced Encryption Standard) encryption algorithm is used to encrypt the user's financial information data during the Internet financial business process. The AES encryption algorithm is a well-known technology and will not be described in detail here. As other implementation methods, implementers can use other methods of existing technology, such as the DES encryption algorithm, the MD5 encryption algorithm, etc. This embodiment does not impose any special restrictions on this.

[0114] It should be understood that, although Figure 1 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0115] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0116] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application. Therefore, any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of this application, without departing from the content of the technical solution of this application, shall fall within the protection scope of the technical solution of this application.

Claims

1. An information security protection method in an Internet financial service process, characterized in that, The method comprises the following steps: Obtaining a fingerprint image of a user, encoding various information categories of historical data of the user in multiple Internet financial service processes to form a service information vector of each information category; Extracting edge contours in the fingerprint image as each ridge line, screening edge pixel points based on the number distribution of edge pixel points in the neighborhood of each ridge line to obtain each feature point; Obtaining a feature evaluation value of each feature point by combining the number of edge pixel points on the ridge line where each feature point is located, the number distribution of edge pixel points in the neighborhood of each feature point, and the gray difference between each feature point and pixel points in the neighborhood thereof; Screening feature points based on the feature evaluation value to obtain each significant point, clustering all significant points, analyzing the discrete distribution between each significant point and the remaining significant points in the same cluster, combining the feature evaluation value to determine the information effectiveness of each significant point, encoding the information effectiveness to obtain biological information encoding of each significant point; Analyzing the frequency characteristics of the same elements and the difference of different elements in the service information vector to determine the change significance of each information category, screening information categories to obtain each feature category, using a threshold secret sharing scheme to combine the change significance of all feature categories to obtain an original secret, shifting and encoding the biological information encoding to obtain a key, and encrypting user information generated in the process of developing Internet financial services by using an encryption algorithm to protect the user information. The feature evaluation value of each feature point comprises: Taking the number of all edge pixel points in the eight-neighborhood of each feature point on each ridge line as a cross-connection factor of each feature point; Taking the sum of the gray value difference between each feature point and all pixel points in the eight-neighborhood thereof as a relative gray difference; Counting the number of all edge pixel points on the ridge line where each feature point is located in the fingerprint image; The feature evaluation value is the normalized result of the product of the number, the cross-connection factor and the relative gray difference.

2. The method of claim 1, wherein the method further comprises: The further obtaining process of the service information vector of each information category comprises: Converting historical data of each information category of the user in each financial service process into binary encoding, and composing all binary encodings of each information category into a service information vector of each information category according to the time sequence of the financial service process of the user.

3. The method of claim 1, wherein the method further comprises: The further obtaining process of each feature point comprises: ​ Thinning all edge contours in the fingerprint image as each ridge line, and taking any edge pixel point on each ridge line as a feature point if there are three or more edge pixel points in the eight-neighborhood of the edge pixel point.

4. The method of claim 1, wherein the method further comprises: The further obtaining process of each significant point comprises: ​ Obtaining a segmentation threshold value of the feature evaluation value of all feature points in the fingerprint image as a first segmentation threshold value, and taking a feature point with a feature evaluation value greater than or equal to the first segmentation threshold value as a significant point.

5. The method of claim 1, wherein the method further comprises: The clustering of all significant points comprises: ​ Each salient point is taken as a graph node, and there is an edge between any two graph nodes, and the distance between the position coordinates of any two salient points in the fingerprint image is taken as the edge weight between the two graph nodes, to construct a weighted undirected graph; The weighted undirected graph is clustered to obtain a plurality of clustering clusters.

6. The method of claim 1, wherein the method further comprises: The information validity of each salient point is determined, including: ​ The Shannon entropy of each salient point and the position coordinates corresponding to all the remaining salient points in the clustering cluster is calculated; The information validity is the normalized result of the product of the feature evaluation value of each salient point and the Shannon entropy.

7. The method of claim 1, wherein the method further comprises: The change saliency of each information category is determined, including: ​ The number of times of occurrence of the same binary code in the service information vector of each information category is counted; and the ratio of the highest number to the number of all binary codes in the service information vector is taken as the user behavior concentration of each information category. The cumulative sum of the edit distance between any two binary codes in the service information vector of each information category is calculated. The change saliency is the down rounding result of the product of the cumulative sum and the user behavior concentration.

8. The method of claim 1, wherein the method further comprises: The further acquisition process of each feature category is: ​ A segmentation threshold of the change saliency of all information categories is obtained, denoted as a second segmentation threshold; and the information category with the change saliency greater than the second segmentation threshold is denoted as a feature category.

9. The method of claim 1, wherein the method further comprises: The key acquisition method is: ​ A polynomial function is constructed through a (t, n) threshold secret sharing scheme, the secret of the (t, n) threshold secret sharing scheme is distributed and recovered based on the change saliency of each feature category, and the original secret is recovered; The original secret is taken as a cyclic shift bit number, and the biological information codes of all salient points are cyclically shifted respectively, and the cyclically shifted biological information codes of all salient points are combined to form a biological information shift sequence; The cyclically shifted biological information code of each salient point is converted into a decimal number, the sum of the decimal number and all coefficients of the polynomial function is calculated, and each digit value of the sum is respectively input into a pseudo-random number generator as a seed to output a pseudo-random number sequence; The pseudo-random sequences corresponding to all salient points are arranged according to the order of the salient points in the biological information shift sequence to form a key of an encryption algorithm.

Images