A SaaS-based smart store identity authentication method and system

Abstract

The present invention provides a SaaS-based smart store identity authentication method and system, which is applied to the server in the authentication system, and the method includes: receiving a first feature summary sent by a user end; determining the corresponding suspected user range according to the first feature summary, obtaining the standard authentication image of each suspected user, and processing the standard authentication image using an interference algorithm to obtain a plurality of transformed images; for each transformed image of each suspected user, extracting the suspected user features from the transformed image, and processing the suspected user features using an irreversible summary algorithm to obtain a second feature summary; comparing each second feature summary with the first feature summary, taking the suspected user corresponding to the second feature summary identical to the first feature summary as the login user, and allowing the client to log in. By applying the embodiments of the present invention, it is not easy to have authentication errors or authentication failures.

Description

Technical Field

[0001] The present invention relates to the field of identity authentication technology, and in particular to a SaaS-based smart store identity authentication method and system. Background Art

[0002] In SaaS (Software-as-a-Service), software vendors deploy application software uniformly on their own servers. Users can order the required application software services from software vendors through the Internet according to actual work needs, pay fees to software vendors according to the number of ordered service modules and the length of time, and obtain services provided by SaaS software vendors through the Internet. When users use the SaaS platform, they need to log in and authenticate. The existing authentication methods mainly include account password, fingerprint, and face recognition. The most commonly used method is account password login, but many software vendors have added fingerprint and face recognition authentication methods to user login authentication in order to improve the security of user accounts.

[0003] At present, the basic process of existing fingerprint and face recognition authentication methods is: use an image acquisition device to collect fingerprints or face images, then crop and compress the images and upload them to the server. The server will allow login after verification.

[0004] However, in the prior art, when an image is cropped and compressed, the features contained in the image are also cropped and compressed simultaneously, resulting in the loss of facial or fingerprint features contained in the image, which makes it easy for authentication errors or authentication failures to occur. Summary of the invention

[0005] The technical problem to be solved by the present invention is how to provide a SaaS-based smart store identity authentication method and system to reduce the authentication errors or authentication failures existing in the prior art.

[0006] The present invention solves the above technical problems through the following technical means:

[0007] The present invention provides a SaaS-based smart store identity authentication method, which is applied to a server in an authentication system, wherein the authentication system includes a server and a user, and the method includes:

[0008] Receiving a first feature summary sent by a user terminal, wherein the first feature summary is obtained by the user terminal using an irreversible summary algorithm to process a login user feature in an original image of the login user, and the original image includes: a face image, a fingerprint image, or a combination thereof;

[0009] Determine the corresponding suspected user range according to the first feature summary, obtain the standard authentication image of each suspected user, and process the standard authentication image using an interference algorithm to obtain a plurality of transformed images, wherein the interference algorithm includes: one or a combination of a noise algorithm and a perturbation algorithm;

[0010] For each transformed image of each suspected user, extracting suspected user features from the transformed image, and processing the suspected user features using an irreversible summary algorithm to obtain a second feature summary;

[0011] Each second feature summary is compared with the first feature summary, and the suspected user corresponding to the second feature summary that is the same as the first feature summary is taken as the login user, and the client is allowed to log in.

[0012] Optionally, the process of generating the first feature summary includes:

[0013] Use feature extraction algorithm to extract logged-in user features from the original image;

[0014] Performing a saliency evaluation on the logged-in user features in the original image according to a preset feature saliency evaluation model to obtain a first saliency score corresponding to each logged-in user feature;

[0015] sorting each logged-in user feature according to the first significance score to obtain a first feature sequence;

[0016] Extracting a first subsequence consisting of a first number of sequentially ordered logged-in user features having a larger first significance score from the first feature sequence;

[0017] The first subsequence is processed by using a hash algorithm to obtain a first feature summary.

[0018] Optionally, determining a corresponding suspected user range according to the first feature summary includes:

[0019] According to the IP address of the user terminal of the first feature summary, the nearest level operator gateway address corresponding to the IP address is obtained, and all system users who use the operator gateway address are regarded as suspected users.

[0020] Optionally, determining a corresponding suspected user range according to the first feature summary includes:

[0021] Partial features of the login user features are obtained, wherein the partial features are a second number of sequentially ordered login user features with larger first significance scores extracted by the user terminal from the first feature sequence, and the second number is smaller than the first number.

[0022] Optionally, extracting the suspected user features from the transformed image and processing the suspected user features using an irreversible summary algorithm to obtain a second feature summary includes:

[0023] Extract suspected user features from the transformed image using feature extraction algorithms;

[0024] Performing saliency evaluation on the suspected user features in the transformed image according to a preset feature saliency evaluation model to obtain a second saliency score corresponding to each suspected user feature;

[0025] Sort each suspected user feature according to the second significance score to obtain a second feature sequence;

[0026] Extracting from the second feature sequence a second subsequence consisting of a first number of sequentially ordered suspected user features having a larger second significance score;

[0027] The second subsequence is processed by using a hash algorithm to obtain a second feature digest.

[0028] Optionally, the step of extracting from the second feature sequence a second subsequence consisting of a first number of sequentially ordered suspected user features having a larger second significance score, comprises:

[0029] Extracting a third number of sequentially ordered suspected user features with larger second significance scores from the second feature sequence to obtain a feature set, wherein the third number is greater than the first number;

[0030] A first number of suspected user features are randomly extracted from the feature set, and are sorted according to second significance score values ​​to obtain a plurality of second subsequences.

[0031] Optionally, the step of extracting from the second feature sequence a second subsequence consisting of a first number of sequentially ordered suspected user features having a larger second significance score, comprises:

[0032] Extracting a third number of sequentially ordered suspected user features with larger second significance scores from the second feature sequence to obtain a feature set, wherein the third number is greater than the first number;

[0033] Sort the suspected user features in the feature set in descending order of the second significance score values ​​to obtain a third subsequence;

[0034] Taking the sequence consisting of the suspected user features in the first fourth number of sequential orders in the third subsequence as the fourth subsequence, wherein the fourth number is less than the first number;

[0035] Randomly select a fifth number of suspected user features from the other suspected user features in the third subsequence except the suspected user features in the fourth subsequence, and the fifth number is equal to the difference between the first number and the fourth number;

[0036] The fifth number of suspected user features extracted each time are arranged after the fourth subsequence in descending order of the second significance score values ​​to obtain a plurality of second subsequences.

[0037] Optionally, before generating the first feature summary, a preset feature word library is used to perform abstract processing on the login user feature to obtain an abstracted first subsequence, and then a hash algorithm is used to process the first subsequence to obtain the first feature summary;

[0038] Before generating the second feature summary, the suspected user feature is abstracted using a preset feature word library to obtain an abstracted second subsequence, and then the second subsequence is processed using a hash algorithm to obtain the second feature summary.

[0039] Optionally, when the number of the suspected users is two or more, the method further includes:

[0040] Returning an instruction for generating a temporary login account to the client, the client is configured to generate a temporary login account according to the instruction for generating a temporary login account, and upload the original image corresponding to the first feature summary to the server;

[0041] The original image returned by the client is received, the original image is compared with the standard authentication image one by one to obtain the corresponding login user, the login user authentication pass information is sent to the client and the client is allowed to log in.

[0042] The present invention also provides a SaaS-based smart store identity authentication system, the system comprising a user end and a server end that communicate using the Internet, wherein the server end is used to execute any of the methods described above.

[0043] The advantages of the present invention are:

[0044] By applying the embodiment of the present invention, the login user features are directly extracted from the original image, the login user features are combined and mapped using a hash algorithm, and the client sends a first feature summary to the server; the server performs a hash algorithm mapping on the suspected user features in the standard authentication image to obtain a second feature summary, and the client only sends the mapped character string when sending data, and the mapped character string comes from the features directly extracted from the original image, and no image compression is required, thereby reducing the loss in the feature extraction process, and it is not easy to have authentication errors or authentication failures; at the same time, the amount of character string sent is relatively small compared to the feature data, thereby improving the transmission speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A flowchart of a SaaS-based smart store identity authentication method provided by an embodiment of the present invention;

[0046] Figure 2 A schematic diagram of the structure of an identity authentication system based on a SaaS system provided by an embodiment of the present invention;

[0047] Figure 3 A schematic diagram of a process for generating a first feature summary in a SaaS-based smart store identity authentication method provided by an embodiment of the present invention;

[0048] Figure 4 A schematic diagram of a process for generating a second feature summary in a SaaS-based smart store identity authentication method provided by an embodiment of the present invention;

[0049] Figure 5 A schematic diagram of a flow chart of a second subsequence generation process in a SaaS-based smart store identity authentication method provided in an embodiment of the present invention;

[0050] Figure 6 A flowchart of another SaaS-based smart store identity authentication method provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in combination with the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] Example 1

[0053] It should be noted that Embodiment 1 of the present invention is preferably applicable to a server in an identity authentication system based on a SaaS system, in which a user terminal and a server terminal communicate via the Internet.

[0054] Figure 1 A flowchart of a SaaS-based smart store identity authentication method provided by an embodiment of the present invention is provided. Figure 2 A schematic diagram of the structure of an identity authentication system based on a SaaS system provided by an embodiment of the present invention is shown in FIG. Figure 1 and Figure 2 As shown, the method includes:

[0055] S101: Receive a first feature summary sent by a user terminal, wherein the first feature summary is obtained by the user terminal using an irreversible summary algorithm to process login user features in an original image of the login user, and the original image includes: one or a combination of a face image and a fingerprint image.

[0056] Figure 3 A flowchart of a process for generating a first feature summary in a SaaS-based smart store identity authentication method provided by an embodiment of the present invention is shown in FIG. Figure 3 As shown, in this step, the user terminal may have a camera for capturing the face image, iris image, etc. of the logged-in user, or may also have a fingerprint recognition device for acquiring the fingerprint image of the logged-in user. The face image, iris image, and fingerprint image acquired above may all be used as the original image of the logged-in user. It is understood that the original image includes but is not limited to the face image, iris image, and fingerprint image.

[0057] After the client collects the original image of the logged-in user, the first feature summary of the logged-in user is extracted using the following method:

[0058] S201: The user terminal uses an existing feature extraction algorithm such as a convolutional layer in a neural network to extract login user features from the original image. For example, the extracted login user features include:

[0059] Feature 1, feature 2, feature 3, feature 4, feature 5, feature 6. In practical applications, the logged-in user features may be the user's eyebrow length, eyebrow width, nose bridge length, eye center axis angle, bifurcation point, isolated point, short line length, direction, short line position, etc.

[0060] S202: Use the preset feature saliency evaluation model to perform saliency evaluation on the logged-in user features in the original image to obtain a first saliency score corresponding to each logged-in user feature:

[0061] First significance score 1, first significance score 2, first significance score 3, first significance score 4, first significance score 5, first significance score 6. It can be understood that the acquisition process of the feature significance evaluation model can be that the annotator annotates according to the significance of each feature in the training image, and then uses the annotated training image as a training sample to train the neural network model to obtain the trained model, and then uses the trained neural network model to evaluate each logged-in user feature to obtain the first significance score.

[0062] S203: The user terminal then sorts the features of each logged-in user in descending order according to the first significance score to obtain a first feature sequence. For example, the obtained first feature sequence is:

[0063] Feature 1, Feature 2, Feature 3, Feature 4, Feature 5, Feature 6.

[0064] S204: When the value of the first number is 4, a first subsequence consisting of the first four sequentially ordered login user features with the largest first significance scores is intercepted from the first feature sequence:

[0065] Feature 1, Feature 2, Feature 3, Feature 4.

[0066] S205: Process the first subsequence using a hash algorithm to obtain a first feature digest hash1.

[0067] The client sends the first feature digest hash1 to the server.

[0068] S102: Determine a corresponding suspected user range according to the first feature summary, obtain a standard authentication image of each suspected user, and process the standard authentication image using an interference algorithm to obtain a plurality of transformed images, wherein the interference algorithm includes: one or a combination of a noise algorithm and a perturbation algorithm.

[0069] In the first implementation of this step, the nearest level operator gateway address corresponding to the IP address can be obtained based on the IP address of the user terminal of the first feature summary. For example, if the nearest level operator gateway address corresponding to the IP address is a community Internet gateway, all system users communicating through the community Internet gateway will be regarded as suspected users. In the chain industry, the same community household may have more than one store. Therefore, there may be a phenomenon of staff secondment between adjacent stores, or a phenomenon of relatives introducing jobs. Therefore, it is possible that the staff of store A logs into the system of store B first, or clerk 1 logs into the system of store A, and clerk 2, a relative of clerk 1, logs into the system of store B. In addition, employees are seconded between stores. Therefore, treating all system users who use the operator gateway address as suspected users can avoid the probability of login failure.

[0070] In a second implementation of this step, a second number of features, such as two login user features, can be extracted from the first feature sequence feature 1, feature 2, feature 3, feature 4, feature 5, and feature 6 obtained in step S101, and the two login user features can be used as a basis to screen out suspected users. Using features with relatively obvious significance to match suspected users can improve the reliability of suspected user matching, reduce the probability of mismatching, and reduce the amount of calculation.

[0071] For example, the suspected users matched in this step are:

[0072] Suspected user 1, suspected user 2, suspected user 3, suspected user 4.

[0073] For each suspected user, an optical noise algorithm is used to interfere with the standard authentication image of the suspected user 1. A noise algorithm can also be used to interfere with the standard authentication image of the suspected user 1 to simulate the login environment of the logged-in user in actual applications to improve the accuracy of matching. Furthermore, the standard authentication image is a three-dimensional vector image scanned using holographic technology. When transforming the standard authentication image, the three-dimensional vector image can be first mapped into a three-dimensional space, and a simulated light source is set at a pre-selected position in the three-dimensional space to obtain a transformed image under the illumination of light sources at different positions, further improving the influence of light source illumination on the logged-in user, and improving the similarity of illumination between the transformed image and the original image, so as to imitate the login environment of the logged-in user in actual applications to improve the accuracy of matching.

[0074] In practical applications, the original image can be processed by superposition of multiple interference algorithms to obtain a transformed image.

[0075] By analogy, several transformed images of each suspected user are obtained.

[0076] S103: for each transformed image of each suspected user, extracting suspected user features from the transformed image, and processing the suspected user features using an irreversible summary algorithm to obtain a second feature summary.

[0077] Figure 4 A flowchart of a second feature summary generation process in a SaaS-based smart store identity authentication method provided by an embodiment of the present invention is shown in FIG. Figure 4 As shown, in the first implementation of this step, the following method can be used to extract the second feature summary:

[0078] S401: For each transformed image, first use the same feature extraction algorithm as in the process of extracting the logged-in user features on the user side in step S101 to extract the suspected user features from the transformed image. The extracted suspected user features are:

[0079] Suspected user feature 1, suspected user feature 2, suspected user feature 3, suspected user feature 4, suspected user feature 5, suspected user feature 6, suspected user feature 7, suspected user feature 8.

[0080] For example, when a first number of four of all logged-in user features are used to generate the first feature summary, the number of suspected user feature summaries used to generate the second feature summary should also be all second feature summaries, which should also be four.

[0081] S402: using the feature saliency evaluation model obtained in step S101 to obtain the first saliency score, perform saliency evaluation on the suspected user features in the transformed image to obtain a second saliency score corresponding to each suspected user feature;

[0082] S403: Sort each suspected user feature according to the second significance score to obtain a second feature sequence:

[0083] Suspected user feature 1, suspected user feature 2, suspected user feature 3, suspected user feature 4, suspected user feature 5, suspected user feature 6, suspected user feature 7, suspected user feature 8.

[0084] S404: A second subsequence consisting of a first number of sequentially ordered suspected user features with a larger second significance score is intercepted from the second feature sequence; when the value of the first number is 4, the second subsequence is: suspected user feature 1, suspected user feature 2, suspected user feature 3, suspected user feature 4,

[0085] S405: Process the second subsequence using a hash algorithm to obtain a first feature summary.

[0086] In actual applications, the type and number of login user features used when generating the first feature summary are consistent with the type and number of login user features used when generating the second feature summary.

[0087] In the second implementation of this step, the following method can be used to extract the second feature summary:

[0088] The suspected user features are extracted from the transformed image. The extracted suspected user features are sorted in descending order according to the second significance value to obtain the second feature sequence:

[0089] Suspected user feature 1, suspected user feature 2, suspected user feature 3, suspected user feature 4, suspected user feature 5, suspected user feature 6, suspected user feature 7, suspected user feature 8.

[0090] When the third number is 6, the suspected user features of the first 6 order of the second significance score can be firstly intercepted from the second feature sequence to obtain a feature set, as shown below:

[0091] Suspected user feature 1, suspected user feature 2, suspected user feature 3, suspected user feature 4, suspected user feature 5, suspected user feature 6.

[0092] Then, four suspected user features are randomly selected from the feature set and sorted according to the second significance score values ​​to obtain several second subsequences:

[0093] Suspected user feature 1, suspected user feature 2, suspected user feature 3, suspected user feature 4;

[0094] Suspected user feature 1, suspected user feature 2, suspected user feature 3, suspected user feature 5;

[0095] Suspected user feature 1, suspected user feature 2, suspected user feature 3, suspected user feature 6;

[0096] Suspected user feature 1, suspected user feature 3, suspected user feature 5, suspected user feature 6.

[0097] The 6 suspected user features in the feature set are freely combined to obtain a total of 15 second subsequences.

[0098] By applying the above-mentioned embodiment of the present invention, a first number of suspected user features are positively screened out from the suspected user features and freely combined, so as to obtain as many suspected user feature combinations as possible. Even if there are subtle differences in the feature significance rankings between the server and the user in the feature extraction algorithm, in the embodiment of the present invention, a feature order corresponding to the first feature summary can also be combined, thereby improving the accuracy of the comparison.

[0099] In the third implementation of this step, the following method can be used to extract the second feature summary: Figure 5 A schematic diagram of a flow chart of a second subsequence generation process in a SaaS-based smart store identity authentication method provided by an embodiment of the present invention, such as Figure 5 As shown,

[0100] S501: For example, when a first number of four of all logged-in user features is used when generating a first feature summary, the number of suspected user feature summaries used when generating a second feature summary should also be four.

[0101] The suspected user features are extracted from the transformed image, and the extracted suspected user features are sorted in descending order according to the second significance value to obtain the second feature sequence:

[0102] Suspected user feature 1, suspected user feature 2, suspected user feature 3, suspected user feature 4, suspected user feature 5, suspected user feature 6, suspected user feature 7, suspected user feature 8.

[0103] Therefore, when the third number is 6, the suspected user features in the first 6 order of the second significance score may be firstly intercepted from the second feature sequence to obtain a feature set.

[0104] S502: After sorting the first feature geometry in descending order according to the second significance value, the third subsequence obtained is as follows:

[0105] Suspected user feature 1, suspected user feature 2, suspected user feature 3, suspected user feature 4, suspected user feature 5, suspected user feature 6.

[0106] S503: Then, the fourth number is a value less than the first number, for example, it can be 3; select the three suspected user features with the largest second significance value from the feature set, and sort the first three suspected user features according to the extracted second significance value, and obtain a fourth subsequence after arranging them in order.

[0107] Suspected user feature 1, suspected user feature 2, suspected user feature 3.

[0108] S504: taking the other suspected user features in the third subsequence except the fourth subsequence as a residual set, and randomly selecting a fifth number, ie, 4-3=1 suspected user features from the residual set as suspected user feature 4.

[0109] S505: The selected one suspected user characteristic is combined and sorted with the fourth subsequence to obtain a plurality of second subsequences.

[0110] Suspected user feature 1, suspected user feature 2, suspected user feature 3, suspected user feature 4.

[0111] By applying the above-mentioned embodiment of the present invention, in actual applications, even if there are slight differences in the algorithm operation results, the extraction results of features with greater significance are relatively consistent. Therefore, it is ensured that the suspected user features with a larger second significance value remain unchanged, and then the suspected user features with a smaller second significance value are freely replaced. In this way, compared with the third embodiment, the number of combined second subsequences can be reduced, and the accuracy can also be guaranteed.

[0112] In a fourth implementation of this step, the following method may be used to extract the second feature summary: in order to avoid the difference between the first feature summary and the second feature summary caused by a slight difference between the length of the nose used in generating the first feature summary and the length of the nose used in generating the second summary, before generating the first feature summary, the embodiment of the present invention uses a preset feature word library to abstract the features of the logged-in user to obtain an abstracted first subsequence, and then uses a hash algorithm to process the first subsequence to obtain the first feature summary; before generating the second feature summary, uses a preset feature word library to abstract the features of the suspected user to obtain an abstracted second subsequence, and then uses a hash algorithm to process the second subsequence to obtain the second feature summary.

[0113] For example, in the embodiment of the present invention, the value is divided according to the length of the nose, and the nose is marked according to the length of the nose. The marking uses a preset feature word library such as super short nose, short nose, normal, long nose, and super long nose, and then the preset feature word library is used to describe the features respectively, that is, the corresponding preset feature word library is used to perform abstract processing of the login user's features, and a first subsequence composed of fields in the feature word library is obtained. It should be noted that the above setting of the feature word library according to the length of the nose is only a simple example. In actual applications, fields such as nose width, nose bridge height, number of nose bridge scars, and nose bridge scar color can also be added to the feature word library to achieve relatively accurate description.

[0114] By applying the above embodiment of the present invention, the feature word library is used to standardize the login user features and the intended user features when generating the first feature summary and the second feature summary, which can avoid errors caused by the algorithm and thus improve the comparison accuracy.

[0115] S104: Compare each second feature summary with the first feature summary, take the suspected user corresponding to the second feature summary that is the same as the first feature summary as the login user, and allow the client to log in.

[0116] This step is the same as the method of allowing login after authentication in the prior art, and the embodiment of the present invention will not be described in detail here.

[0117] Example 2

[0118] Figure 6 A flowchart of another SaaS-based smart store identity authentication method provided by an embodiment of the present invention is shown in FIG. Figure 6 As shown, when the number of the suspected users is two or more, it indicates that the server authentication is unsuccessful. In order to avoid repeated authentication of the user, Embodiment 2 of the present invention adds the following steps on the basis of Embodiment 1:

[0119] S105: The server returns an instruction for generating a temporary login account to the client, and the client generates a temporary login account according to the instruction for generating a temporary login account. When the customer has a settlement requirement, the client settles the customer in the temporary login account and uses the original image of the customer as identification information of the temporary login account. During the settlement process, the original image corresponding to the first feature summary and the information in the temporary login account are uploaded to the server;

[0120] S106: After receiving the original image returned by the client, the server uses an existing image similarity algorithm or a person recognition algorithm to compare the original image with the standard authentication image one by one to obtain the corresponding login user, sends the login user authentication pass information to the client and authorizes the client to log in; at the same time, merges the information in the temporary login account with the account information corresponding to the customer.

[0121] By applying the above-mentioned embodiment of the present invention, secondary authentication is performed during the client settlement process without affecting the client settlement, and user-free authentication is achieved.

[0122] Example 3

[0123] Corresponding to Embodiment 1 of the present invention, Embodiment 3 of the present invention provides an identity authentication device based on a SaaS system, which is applied to a server in an authentication system, wherein the authentication system includes a server and a user, and the device includes:

[0124] A receiving module, configured to receive a first feature summary sent by a user terminal, wherein the first feature summary is obtained by the user terminal using an irreversible summary algorithm to process a login user feature in an original image of the login user, and the original image includes: a face image, a fingerprint image, or a combination thereof;

[0125] an acquisition module, configured to determine a corresponding suspected user range according to the first feature summary, acquire a standard authentication image of each suspected user, and process the standard authentication image using an interference algorithm to obtain a plurality of transformed images, wherein the interference algorithm includes: one or a combination of a noise algorithm and a perturbation algorithm;

[0126] An extraction module, for extracting suspected user features from each transformed image of each suspected user, and processing the suspected user features using an irreversible summary algorithm to obtain a second feature summary;

[0127] The comparison module is used to compare each second feature summary with the first feature summary, take the suspected user corresponding to the second feature summary that is the same as the first feature summary as the login user, and allow the client to log in.

[0128] In a specific implementation of Example 3 of the present invention, the process of generating the first feature summary includes:

[0129] Use feature extraction algorithm to extract logged-in user features from the original image;

[0130] Performing a saliency evaluation on the logged-in user features in the original image according to a preset feature saliency evaluation model to obtain a first saliency score corresponding to each logged-in user feature;

[0131] sorting each logged-in user feature according to the first significance score to obtain a first feature sequence;

[0132] Extracting a first subsequence consisting of a first number of sequentially ordered logged-in user features having a larger first significance score from the first feature sequence;

[0133] The first subsequence is processed by using a hash algorithm to obtain a first feature summary.

[0134] In a specific implementation of Example 3 of the present invention, the acquisition module is used to:

[0135] According to the IP address of the user terminal of the first feature summary, the nearest level operator gateway address corresponding to the IP address is obtained, and all system users who use the operator gateway address are regarded as suspected users.

[0136] In a specific implementation of Example 3 of the present invention, the acquisition module is used to:

[0137] Partial features of the login user features are obtained, wherein the partial features are a second number of sequentially ordered login user features with larger first significance scores extracted by the user terminal from the first feature sequence, and the second number is smaller than the first number.

[0138] In a specific implementation of Example 3 of the present invention, the extraction module is used to:

[0139] Extract suspected user features from the transformed image using feature extraction algorithms;

[0140] Performing saliency evaluation on the suspected user features in the transformed image according to a preset feature saliency evaluation model to obtain a second saliency score corresponding to each suspected user feature;

[0141] Sort each suspected user feature according to the second significance score to obtain a second feature sequence;

[0142] Extracting from the second feature sequence a second subsequence consisting of a first number of sequentially ordered suspected user features having a larger second significance score;

[0143] The second subsequence is processed by using a hash algorithm to obtain a second feature digest.

[0144] In a specific implementation of Example 3 of the present invention, the extraction module is used to:

[0145] Extracting a third number of sequentially ordered suspected user features with larger second significance scores from the second feature sequence to obtain a feature set, wherein the third number is greater than the first number;

[0146] A first number of suspected user features are randomly extracted from the feature set, and are sorted according to second significance score values ​​to obtain a plurality of second subsequences.

[0147] In a specific implementation of Example 3 of the present invention, the extraction module is used to:

[0148] Extracting a third number of sequentially ordered suspected user features with larger second significance scores from the second feature sequence to obtain a feature set, wherein the third number is greater than the first number;

[0149] Sort the suspected user features in the feature set in descending order of the second significance score values ​​to obtain a third subsequence;

[0150] Taking the sequence consisting of the suspected user features in the first fourth number of sequential orders in the third subsequence as the fourth subsequence, wherein the fourth number is less than the first number;

[0151] Randomly select a fifth number of suspected user features from the other suspected user features in the third subsequence except the suspected user features in the fourth subsequence, and the fifth number is equal to the difference between the first number and the fourth number;

[0152] The fifth number of suspected user features extracted each time are arranged after the fourth subsequence in descending order of the second significance score values ​​to obtain a plurality of second subsequences.

[0153] In a specific implementation of Example 3 of the present invention, before generating the first feature summary, the login user feature is abstracted using a preset feature word library to obtain an abstracted first subsequence, and then the first subsequence is processed using a hash algorithm to obtain the first feature summary;

[0154] Before generating the second feature summary, the suspected user feature is abstracted using a preset feature word library to obtain an abstracted second subsequence, and then the second subsequence is processed using a hash algorithm to obtain the second feature summary.

[0155] In a specific implementation of Embodiment 3 of the present invention, when the number of the suspected users is two or more, the method further includes:

[0156] Returning an instruction for generating a temporary login account to the client, the client is configured to generate a temporary login account according to the instruction for generating a temporary login account, and upload the original image corresponding to the first feature summary to the server;

[0157] The original image returned by the client is received, the original image is compared with the standard authentication image one by one to obtain the corresponding login user, the login user authentication pass information is sent to the client and the client is allowed to log in.

[0158] Example 4

[0159] Corresponding to Embodiment 1 of the present invention, Embodiment 4 of the present invention provides an identity authentication system based on a SaaS system.

[0160] Figure 2 A schematic diagram of the structure of an identity authentication system based on a SaaS system provided by an embodiment of the present invention is shown in FIG. Figure 2 As shown, the system includes a user terminal 10 and a server terminal that communicate using the Internet, wherein the server terminal 20 is used to execute the method described in any one of Embodiments 1 or 2.

[0161] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A SaaS-based smart store identity authentication method, characterized in that: A server end used in an authentication system, wherein the authentication system includes a server end and a user end, and the method includes: Receiving a first feature summary sent by a user terminal, wherein the first feature summary is obtained by the user terminal using an irreversible summary algorithm to process a login user feature in an original image of the login user, and the original image includes: a face image, a fingerprint image, or a combination thereof; Determine the corresponding suspected user range according to the first feature summary, obtain the standard authentication image of each suspected user, and process the standard authentication image using an interference algorithm to obtain a plurality of transformed images, wherein the interference algorithm includes: one or a combination of a noise algorithm and a perturbation algorithm; For each transformed image of each suspected user, extracting suspected user features from the transformed image, and processing the suspected user features using an irreversible summary algorithm to obtain a second feature summary; Compare each second feature summary with the first feature summary, take the suspected user corresponding to the second feature summary that is the same as the first feature summary as the login user, and allow the client to log in; Determining a corresponding suspected user range according to the first feature summary includes: According to the IP address of the user terminal of the first feature summary, the nearest level operator gateway address corresponding to the IP address is obtained, and all system users who use the operator gateway address are regarded as suspected users.

2. According to the SaaS-based smart store identity authentication method of claim 1, it is characterized in that: The generation process of the first feature summary includes: Use feature extraction algorithm to extract logged-in user features from the original image; Performing a saliency evaluation on the logged-in user features in the original image according to a preset feature saliency evaluation model to obtain a first saliency score corresponding to each logged-in user feature; sorting each logged-in user feature according to the first significance score to obtain a first feature sequence; Extracting a first subsequence consisting of a first number of sequentially ordered logged-in user features having a larger first significance score from the first feature sequence; The first subsequence is processed by using a hash algorithm to obtain a first feature summary.

3. According to the SaaS-based smart store identity authentication method of claim 2, it is characterized in that: Determining a corresponding suspected user range according to the first feature summary includes: Partial features of the login user features are obtained, wherein the partial features are a second number of sequentially ordered login user features with larger first significance scores extracted by the user terminal from the first feature sequence, and the second number is smaller than the first number.

4. According to the SaaS-based smart store identity authentication method of claim 2, it is characterized in that: The step of extracting the suspected user features from the transformed image and processing the suspected user features using an irreversible summary algorithm to obtain a second feature summary includes: Extract suspected user features from the transformed image using feature extraction algorithms; Performing saliency evaluation on the suspected user features in the transformed image according to a preset feature saliency evaluation model to obtain a second saliency score corresponding to each suspected user feature; Sort each suspected user feature according to the second significance score to obtain a second feature sequence; Extracting from the second feature sequence a second subsequence consisting of a first number of sequentially ordered suspected user features having a larger second significance score; The second subsequence is processed by using a hash algorithm to obtain a second feature digest.

5. A SaaS-based smart store identity authentication method according to claim 4, characterized in that: The second subsequence consisting of a first number of sequentially ordered suspected user features with larger second significance scores is extracted from the second feature sequence, comprising: Extracting a third number of sequentially ordered suspected user features with larger second significance scores from the second feature sequence to obtain a feature set, wherein the third number is greater than the first number; A first number of suspected user features are randomly extracted from the feature set, and are sorted according to second significance score values ​​to obtain a plurality of second subsequences.

6. A SaaS-based smart store identity authentication method according to claim 4, characterized in that: The second subsequence consisting of a first number of sequentially ordered suspected user features with larger second significance scores is extracted from the second feature sequence, comprising: Extracting a third number of sequentially ordered suspected user features with larger second significance scores from the second feature sequence to obtain a feature set, wherein the third number is greater than the first number; Sort the suspected user features in the feature set in descending order of the second significance score values ​​to obtain a third subsequence; Taking the sequence consisting of the suspected user features in the first fourth number of sequential orders in the third subsequence as the fourth subsequence, wherein the fourth number is smaller than the first number; Randomly select a fifth number of suspected user features from the other suspected user features in the third subsequence except the suspected user features in the fourth subsequence, and the fifth number is equal to the difference between the first number and the fourth number; The fifth number of suspected user features extracted each time are arranged after the fourth subsequence in descending order of the second significance score values ​​to obtain a plurality of second subsequences.

7. A SaaS-based smart store identity authentication method according to any one of claims 3-6, characterized in that: Before generating the first feature summary, the login user feature is abstracted using a preset feature word library to obtain an abstracted first subsequence, and then the first subsequence is processed using a hash algorithm to obtain the first feature summary; Before generating the second feature summary, the suspected user feature is abstracted using a preset feature word library to obtain an abstracted second subsequence, and then the second subsequence is processed using a hash algorithm to obtain the second feature summary.

8. A SaaS-based smart store identity authentication method according to any one of claims 1-6, characterized in that: When the number of the suspected users is two or more, the method further includes: Returning an instruction for generating a temporary login account to the client, the client is configured to generate a temporary login account according to the instruction for generating a temporary login account, and upload the original image corresponding to the first feature summary to the server; The original image returned by the client is received, the original image is compared with the standard authentication image one by one to obtain the corresponding login user, the login user authentication pass information is sent to the client and the client is allowed to log in.

9. A SaaS-based smart store identity authentication system, characterized in that: The system comprises a user end and a server end for communicating using the Internet, wherein the server end is used to execute the method according to any one of claims 1 to 8.

Images