Identity authentication method, system and device based on multi-factor behavior characteristics and medium
Through a two-level authentication method based on multi-factor behavioral characteristics, identity authentication is performed using behavioral characteristic models and handwriting judgment models, which solves the problems of memory inconvenience and insufficient privacy caused by password reliance in existing technologies, and achieves more secure and convenient identity authentication.
Patent Information
- Application Number
- CN202510769091.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-10
- Publication Date
- 2025-09-16
AI Technical Summary
Existing identity authentication technology relies on password verification, which makes it difficult for users to remember, has a high risk of leakage, and lacks privacy.
It adopts an identity authentication method based on multi-factor behavioral characteristics, performs two-level authentication by training behavioral characteristic models and handwriting judgment models, and uses the user's unique behavioral characteristics and handwriting characteristics for identity authentication without the need for password input.
The privacy of identity authentication is improved. Through multi-factor authentication based on behavioral characteristics and handwriting characteristics, the security and convenience of user identity authentication are enhanced, and the risk of password leakage is reduced.
Smart Images
Figure CN120658447A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of computer security technology, and in particular to an identity authentication method, system, device and medium based on multi-factor behavioral characteristics. Background Art
[0002] Today, the world has fully entered the digital age. Identity authentication technology is present in every aspect of every citizen's life, such as logging into communication software, gaming accounts, and banking accounts. Passwords, as "information known to the user," represent a user's identity. Multiple passwords set by a user can be related or even identical, making passwords private. Existing authentication technologies often require users to set a password for login, making it difficult for users to remember passwords and jeopardizing their privacy if the password is leaked. Summary of the Invention
[0003] In view of this, the purpose of the embodiments of the present invention is to provide an identity authentication method, system, device and medium based on multi-factor behavioral characteristics, which does not require password verification and improves privacy.
[0004] In one aspect, an embodiment of the present invention provides an identity authentication method based on multi-factor behavioral characteristics, which is applied to a server and includes:
[0005] Obtaining an identity authentication request, reading a client identification code according to the identity authentication request, and retrieving whether a behavioral feature model corresponding to the client is mature according to the identification code;
[0006] If the behavioral feature model is maturely trained, the behavioral feature information of the current object is obtained, and the behavioral feature information is input into the maturely trained behavioral feature model for first-level authentication. If the first-level authentication is passed, the client is allowed to enter the application system;
[0007] If the behavioral feature model is not mature or the first-level authentication fails, several characters of the current object are obtained and input into the trained handwriting judgment model for second-level authentication. If the second-level authentication passes, the client is allowed to enter the application system. If the second-level authentication fails, the login request is rejected.
[0008] Optionally, the method further includes:
[0009] Obtaining an identity registration request, and obtaining the client's identifiable code, handwriting sample data, and behavioral feature sample data of the target object according to the identity registration request;
[0010] A handwriting judgment model of the client corresponding to the recognizable code is trained according to the handwriting sample data, and a behavior feature model of the client corresponding to the recognizable code is trained according to the behavior feature sample data.
[0011] Optionally, the handwriting judgment model is trained by the following method:
[0012] Acquiring handwriting data of a target object and other objects, performing a first standardization process and labeling on the handwriting data to form handwriting sample data of the target object;
[0013] The first preset model is trained according to the handwriting sample data of the target object to determine the model parameters of the first preset model; the first preset model includes a convolution layer, a pooling layer, a fully connected layer and an output layer.
[0014] Optionally, the behavior feature model is trained by the following method:
[0015] Obtaining behavioral characteristic data of the target object; the behavioral characteristic data includes activity trajectory data, walking gait data, and screen operation behavior data;
[0016] Preprocessing the behavior feature data of the target object to obtain behavior feature sample data of the target object;
[0017] The second preset model is trained according to the behavioral feature sample data of the target object to determine the model parameters of the second preset model; the second preset model includes a convolutional layer, a pooling layer, a fully connected layer and an output layer.
[0018] Optionally, preprocessing the behavior feature data of the target object includes:
[0019] Performing coordinate transformation on the activity trajectory data to obtain a thermal density map, and generating a grayscale image from the thermal density map according to kernel density estimation;
[0020] Performing Fourier transform on the walking gait data to generate a time-frequency spectrum;
[0021] Encoding the screen operation behavior data into a trajectory vector diagram according to the touch trajectory coordinates and pressure values;
[0022] The grayscale image, the time-frequency spectrum diagram and the trajectory vector diagram are subjected to a second normalization process and feature fusion.
[0023] Optionally, the behavioral feature information is input into a well-trained behavioral feature model for first-level authentication, including:
[0024] Inputting the behavioral feature information into a well-trained behavioral feature model to obtain the probability of similarity between the behavioral features of the current user and the target user;
[0025] If the similarity probability of the behavioral characteristics is less than the first threshold, the first-level authentication fails;
[0026] If the probability of the behavioral feature similarity is greater than or equal to the first threshold and less than the second threshold, the first-level authentication is passed and the number of authentications is recorded; if the number of authentications exceeds the preset number, the first-level authentication fails;
[0027] If the behavior feature similarity probability is greater than or equal to the second threshold, the first-level authentication is passed.
[0028] Optionally, a plurality of the characters are input into a trained handwriting judgment model for secondary authentication, including:
[0029] Inputting a plurality of the characters into a trained handwriting judgment model to obtain a handwriting similarity probability value;
[0030] If the handwriting similarity probability value is greater than or equal to the third threshold, the second-level authentication is passed;
[0031] If the handwriting similarity probability value is less than the third threshold, the secondary authentication fails.
[0032] On the other hand, an embodiment of the present invention provides an identity authentication device based on multi-factor behavioral characteristics, applied to a server, comprising:
[0033] at least one processor;
[0034] at least one memory for storing at least one program;
[0035] When the at least one program is executed by the at least one processor, the at least one processor implements the above method.
[0036] On the other hand, an embodiment of the present invention provides a computer-readable storage medium storing a program executable by a processor. When the program is executed by the processor, it is used to perform the above method.
[0037] On the other hand, an embodiment of the present invention provides an identity authentication system based on multi-factor behavioral characteristics, including a client and a server connected to the client; wherein,
[0038] The client is configured to collect behavior feature information and character information of an object, and send the behavior feature information and character information to the server;
[0039] The server includes:
[0040] at least one processor;
[0041] at least one memory for storing at least one program;
[0042] When the at least one program is executed by the at least one processor, the at least one processor implements the above method.
[0043] The implementation of the embodiment of the present invention includes the following beneficial effects: This embodiment includes two-level authentication. If the behavioral feature model is maturely trained, the behavioral feature information of the current object is authenticated at the first level through the trained behavioral feature model. If the behavioral feature model is not maturely trained or the first level authentication fails, the characters of the current object are authenticated at the second level through the trained handwriting judgment model. During the identity authentication process, the behavioral feature information and / or characters of the current object are authenticated without the need for password verification. The behavioral feature information and characters of the current object are attribute characteristics unique to the user object and have a strong correlation with the user object, thereby improving privacy. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] Figure 1 This is a structural block diagram of an identity authentication system based on multi-factor behavioral characteristics provided by an embodiment of the present invention;
[0045] Figure 2 This is a flowchart of the steps of an identity authentication method based on multi-factor behavioral characteristics provided by an embodiment of the present invention;
[0046] Figure 3 This is a flowchart of the steps of an identity registration request method provided by an embodiment of the present invention;
[0047] Figure 4 This is a flowchart of the steps of a handwriting judgment model training method provided by an embodiment of the present invention;
[0048] Figure 5 This is a flowchart of the steps of a behavior feature model training method provided by an embodiment of the present invention;
[0049] Figure 6 This is a flowchart of another identity authentication method based on multi-factor behavioral characteristics provided by an embodiment of the present invention;
[0050] Figure 7 This is a structural block diagram of an identity authentication device based on multi-factor behavioral characteristics provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0051] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. The step numbers in the following embodiments are provided for ease of description only and do not limit the order of the steps. The order of execution of the steps in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.
[0052] It should be noted that, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flow chart, in some cases, the steps shown or described can be performed in a different order than the module division in the device or the order in the flow chart. The terms "first", "second", etc. in the specification and claims and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0053] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application pertains. The terms used herein are for the purpose of describing the embodiments of this application only and are not intended to limit this application.
[0054] Some technical terms in this embodiment are explained below.
[0055] Identity Authentication: The process of effectively determining the identity of a legitimate operator during the communication process of a computer network. It is used to ensure that the operator performing actual operations with a certain identity is the legitimate operator corresponding to that identity, thus achieving the unification of real and digital identities.
[0056] Biometric characteristics: A user's biometric characteristics refer to their unique fingerprints, voice, retina, behavioral habits, etc., among which behavioral habits include their handwriting.
[0057] Heatmap: A heatmap is a two-dimensional data visualization technique that represents the size of each value in a data set using color.
[0058] See Figure 1The server and the client are connected in communication. The server stores the application system. The client first sends a registration request for the application system to the server. The server collects the basic information and handwriting data of the user. After successful registration, the client first sends an identity authentication request for the application system to the server. After the identity authentication is passed, the client enters the application system of the server.
[0059] like Figure 2 As shown, an embodiment of the present invention provides an identity authentication method based on multi-factor behavioral characteristics, which is applied to a server and includes:
[0060] S100: Obtain an identity authentication request, read the client's identification code according to the identity authentication request, and retrieve whether the behavior feature model corresponding to the client is mature according to the identification code.
[0061] The client's identifiable code represents the client's identification information, and the client and the identifiable code have a one-to-one correspondence. The behavioral feature model is used to determine whether the client's current user is the target user. The current user refers to the user who is currently using the client device, and the target user refers to the client user registered on the server through the client. Specifically, the client sends an identity authentication request to the server, and at the same time encrypts the data using HTTPS to ensure that the data is not stolen during network transmission. The server reads the client's unique identifiable code based on the identity authentication request, and retrieves the client's corresponding behavioral feature model based on the identifiable code to determine whether it has been maturely trained. The server judges the maturity of the behavioral feature model through the model performance evaluation indicators accuracy and F1 score, and detects whether the behavioral feature model has been maturely trained.
[0062] S200: If the behavior feature model is maturely trained, obtain the behavior feature information of the current object, input the behavior feature information into the maturely trained behavior feature model for first-level authentication, and if the first-level authentication is passed, allow the client to enter the application system.
[0063] If the behavioral feature model is maturely trained, the server obtains the behavioral feature information of the current object through the client. The server inputs the behavioral feature information of the current object into the maturely trained behavioral feature model for first-level authentication. If the current object has a high degree of similarity with the target object, the first-level authentication is passed. If the first-level authentication is passed, the client is allowed to enter the application system.
[0064] S300. If the behavioral feature model training is immature or the first-level authentication fails, obtain several characters of the current object and input the several characters into the trained handwriting judgment model for second-level authentication. If the second-level authentication passes, the client is allowed to enter the application system. If the second-level authentication fails, the login request is rejected.
[0065] If the client receives information that the behavioral feature model is not yet capable of classification or if the first-level authentication fails, the server prompts the client to handwrite a few characters and then uses the handwriting recognition model to perform a second-level authentication to determine the handwriting of the current subject. If the second-level authentication passes, it indicates that the current subject has a high degree of similarity with the target subject, and the client is allowed to enter the application system. If the second-level authentication fails, it indicates that the similarity between the current subject and the target subject does not meet the requirements, and the login request is rejected.
[0066] Optionally, the method further includes:
[0067] S010, obtaining an identity registration request, and obtaining the client's identifiable code, handwriting sample data, and behavior feature sample data of the target object according to the identity registration request;
[0068] S020: training a handwriting judgment model of the client corresponding to the recognizable code based on the handwriting sample data, and training a behavior feature model of the client corresponding to the recognizable code based on the behavior feature sample data.
[0069] See Figure 3 The target subject sends an identity registration request to the server through the client. The registration process occurs when the target subject first registers with the system. The server collects the target subject's basic information: Upon receiving the target subject's identity registration request, the server returns a registration-approved status to the client. The client calls the mobile device API to obtain the device's unique identifier, IMEI, and transmits it to the server via a secure HTTPS channel to record the user's unique identifier code for the client's mobile device and prompt the target subject to enter their handwriting habits. The client uses the touchscreen sensor to collect the user's handwriting data, including stroke order, pressure, speed, and other characteristics, and stores it as time series data. After entering a sufficient amount of the user's handwriting habits, the client sends the corresponding device's unique identifier code and handwriting habit data to the server. After matching the new user, the server returns a successful handwriting habit collection status. The client notifies the user of successful registration: Upon receiving the successful handwriting habit collection status returned by the server, the client notifies the user of successful registration. The server continuously collects user behavioral characteristic data: After the user completes identity registration, the client will continue to collect user behavioral characteristic data and encrypt and send it to the server.
[0070] The server processes the collected behavioral feature data and handwriting data to generate training sample data. The server trains the handwriting recognition model for the client corresponding to the recognizable code based on the handwriting sample data, and trains the behavioral feature model for the client corresponding to the recognizable code based on the behavioral feature sample data.
[0071] Optionally, the handwriting judgment model is trained by the following method:
[0072] S030, obtaining handwriting data of the target object and other objects, performing first standardization processing and label recording on the handwriting data to form handwriting sample data of the target object;
[0073] S040. Train a first preset model according to the handwriting sample data of the target object to determine model parameters of the first preset model; the first preset model includes a convolutional layer, a pooling layer, a fully connected layer and an output layer.
[0074] Specifically, the collected handwriting data is first standardized and a first preset model is constructed, which includes but is not limited to a convolutional neural network (CNN) model, and then the convolutional neural network (CNN) model is trained.
[0075] In a specific embodiment, see Figure 4 , standardize the collected handwriting data: The server converts the user's handwriting data collected through a touch screen, handwriting tablet, or other input device into an image. This data is usually the user's handwritten numbers, letters, or signature. Then, the image is resized (for example, to 28*28 pixels), grayscaled (if the image is a color image), and pixel values are scaled to between 0 and 1 for normalization.
[0076] Construct a convolutional neural network (CNN) model: For the handwriting image after standardization, the server trains the handwriting judgment model based on the CNN model architecture. First, mark the processed user handwriting image as label 1. The CNN model architecture includes a convolutional layer, which performs a convolution operation on the image through multiple convolution filters to extract local features in the image. Assuming that the image size is W*H, where W and H are the width and height of the image respectively, after the convolution operation, the size of the output feature map is usually (W-F+1)*(H-F+1), where F is the size of the convolution kernel K. The convolution operation formula C(i,j) is:
[0077]
[0078] Where I is the input image, and (i, j) is the position of the pixel in the output feature map. The CNN model architecture includes a pooling layer, which is used to reduce the dimensionality of the convolutional layer output, reduce computational effort, and extract more abstract features. The most commonly used method is max pooling. The pooling operation formula, Pooled(i, j), is:
[0079] Pooled(i,j)=max(Region(i,j))
[0080] Here, Region represents the pooling window. The CNN model architecture also includes a fully connected layer, which combines the features extracted by the convolution and pooling layers at a high level and outputs the final classification result. Furthermore, each layer of a convolutional neural network typically includes an activation function, the most commonly used being the ReLU activation function to increase the network's nonlinear capabilities:
[0081] ReLU(x)=max(0,x)
[0082] Among them, x represents the output result of the previous layer. Finally, the CNN model architecture includes an output layer (OutputLayer), which uses the sigmoid activation function to map the output to the interval [0,1]:
[0083]
[0084] Here, z is the raw score input to the model, which is converted to a probability using the sigmoid activation function. Furthermore, the binary cross-entropy loss function is used to output the error of the activation function output:
[0085]
[0086] Among them, y i is the true label of the sample, that is, label 1 or label 0, p i is the probability predicted by the model, and N is the total number of samples.
[0087] Training a convolutional neural network (CNN) model: For processed user handwriting images, label the corresponding user's handwriting image as "consistent handwriting," i.e., label 1, and label all other images as "deviant handwriting," i.e., label 0. The CNN model is trained using the labeled handwriting image data, and the weights in the model network are optimized using the backpropagation algorithm.
[0088] Optionally, the behavior feature model is trained by the following method:
[0089] S050, obtaining behavioral characteristic data of the target object; the behavioral characteristic data includes activity trajectory data, walking gait data and screen operation behavior data;
[0090] S060: Preprocess the behavior feature data of the target object to obtain behavior feature sample data of the target object;
[0091] S070. Train the second preset model according to the sample data of the behavioral characteristics of the target object to determine the model parameters of the second preset model; the second preset model includes a convolutional layer, a pooling layer, a fully connected layer, and an output layer.
[0092] Specifically, first, the server continuously collects the behavioral characteristic data of the user. The server processes the collected user behavioral characteristic data, and then trains the second preset model, which includes but is not limited to a convolutional neural network (CNN) model.
[0093] In a specific embodiment, refer to Figure 5 , the server continuously collects the behavioral characteristic data of the user: after the user completes identity registration, the client will continuously collect the behavioral characteristic data of the user and encrypt it and send it to the server. These behavioral characteristic data include activity trajectory data, walking gait data, and screen operation behaviors. Among them, the activity trajectory data is collected through the GPS positioning module of the user's mobile device, recording information such as the user's common activity areas and movement routes. The walking gait data is collected using the acceleration sensor and gyroscope on the device, recording gait characteristics such as the user's steps and stride. The screen operation behaviors are collected by the touch screen of the device, recording operation data such as the sliding speed and sliding trajectory of the user's finger.
[0094] The server processes the collected user behavioral characteristic data: the server performs filtering and normalization processing on the collected behavioral data to remove noise and convert it into feature vectors. Use Fourier transform to extract frequency domain features from the time domain signal and convert them into numerical features, and extract some key features, such as: movement route, pace rhythm, sliding trajectory, etc.
[0095] Construct a convolutional neural network (CNN) model: input the processed user behavioral characteristic data into the convolutional neural network (CNN) model for training. The goal is to generate a personalized user behavioral characteristic model based on historical behavioral data. This model first converts the original behavioral characteristic data into a two-dimensional feature map that can be processed by CNN. The subsequent construction of the CNN model is the same as that in Embodiment 2, which are respectively a convolutional layer, a pooling layer, a fully connected layer, an output layer, and a loss function. The difference in the output layer from Embodiment 2 lies in the threshold. In this embodiment, the set thresholds are two parameters t1 and t2, and 0 < t1 < t2 < 1. When the probability output by the model is less than t1, the sample is judged as deviating from normal. When it is greater than or equal to t1 and less than t2, it is judged as having a normal probability. When it is greater than or equal to t2, it is judged as completely normal. <0,000213>Train the convolutional neural network (CNN) model: for the continuously collected user behavioral characteristic data, try to filter out several stable data from the numerous data, that is, the fluctuations in the time series are controllable, and directly consider these data as completely normal user behavioral characteristic data and use these data to train the model.
[0097] Optionally, preprocessing the behavior feature data of the target object includes:
[0098] S061. Performing coordinate transformation on the activity trajectory data to obtain a thermal density map, and generating a grayscale image from the thermal density map according to kernel density estimation;
[0099] S062. Performing Fourier transform on the walking gait data to generate a time-frequency spectrum;
[0100] S063, encoding the screen operation behavior data into a trajectory vector diagram according to the touch trajectory coordinates and pressure value;
[0101] S064: Perform a second normalization process and feature fusion on the grayscale image, the time-frequency spectrum diagram, and the trajectory vector diagram.
[0102] For activity trajectory data, the GPS coordinate sequence is converted into a heatmap, and kernel density estimation is used to generate a 256*256 grayscale image to quantify the user's resident area. For walking gait data, the accelerometer three-axis data X, Y, and Z are short-time Fourier transformed to generate a 128*128*3 time-frequency spectrum as a three-channel input. For screen operation behavior, the touch trajectory coordinates and pressure values are encoded into a 128*128*2 trajectory vector diagram, which contains two channels of position trajectory and pressure intensity. Data normalization is then performed to unify the image size to 256*256 pixels, and the pixels are normalized. Min-Max Scaling is used to compress the values of each channel to the [0,1] interval:
[0103]
[0104] Where x is the original eigenvalue, min(x) and max(x) are the minimum and maximum values of the feature in the training set, respectively. The three behavioral feature maps are concatenated into a 256*256*6 multi-channel input matrix. After data processing, the CNN model is constructed. The first input layer is a 256*256*6 feature fusion layer containing the heat density map, gait time spectrum map, and operation trajectory vector map.
[0105] Optionally, the behavioral feature information is input into a well-trained behavioral feature model for first-level authentication, including:
[0106] S210: Input the behavior feature information into a well-trained behavior feature model to obtain a probability of similarity between the behavior features of the current user and the target user;
[0107] S220: If the behavior feature similarity probability is less than the first threshold, the first-level authentication fails;
[0108] S230: If the probability of the behavioral feature similarity is greater than or equal to the first threshold and less than the second threshold, the first-level authentication is passed and the number of authentications is recorded; if the number of authentications exceeds a preset number, the first-level authentication fails;
[0109] S240: If the behavior feature similarity probability is greater than or equal to the second threshold, the first-level authentication is passed.
[0110] If the client receives information that the behavioral feature model already has classification capabilities, the client will collect the user's current behavioral features in real time and submit the user's recent historical behavioral features to the server for judgment. After receiving the data, the server inputs the data into the already trained behavioral feature model, which determines the similarity probability P between the current user's behavioral features and historical user behavioral features. If P is greater than or equal to t2, "behavior consistency" is considered, and the user enters the system directly; if P is less than t2 and greater than or equal to t1, the user is allowed to enter the system but will be recorded. If there are too many such records in a short period of time, further identity authentication measures will be triggered at the next login; otherwise, further identity authentication measures will be triggered directly.
[0111] Optionally, a plurality of the characters are input into a trained handwriting judgment model for secondary authentication, including:
[0112] S310, inputting a plurality of the characters into a trained handwriting judgment model to obtain a handwriting similarity probability value;
[0113] S320: If the handwriting similarity probability value is greater than or equal to the third threshold, the second-level authentication is passed;
[0114] S330: If the handwriting similarity probability value is less than the third threshold, the secondary authentication fails.
[0115] In a specific embodiment, when a user enters new handwriting, the client sends the collected handwriting data to the server and inputs it into the CNN model. After convolution, pooling, and fully connected layers, the CNN model outputs a probability value P(y=1|x). If the output probability value P(y=1|x) is greater than or equal to a preset threshold, it is judged as "handwriting consistency" and the user enters the system directly. If the output probability value P(y=1|x) is less than the threshold, it is judged as "handwriting deviation" and the server rejects the login request.
[0116] In a specific embodiment, see Figure 6 ,The identity authentication process based on multi-factor behavioral ,features is as follows.
[0117] Step 1: When the user registers his / her identity with the server, the server needs to record the unique identifiable code of the user's mobile device and enter the user's handwriting habits.
[0118] Step 2: The server needs to collect enough user handwriting data to train the handwriting model so that the model can perform two classifications on the newly input handwriting, namely "consistent handwriting" and "deviation handwriting".
[0119] Step 3: After the user completes identity registration, the server needs to collect the user's behavioral characteristics when the user normally uses the smart mobile device, including: the user's activity trajectory location information, the user's walking gait, and the user's screen sliding habits.
[0120] Step 4: The server needs to continuously collect the user's behavioral characteristics. During the collection process, the server continuously uses the collected data to train the behavioral characteristic model, so that the model can classify the behavioral characteristics collected in the future into two categories, namely "consistent behavior" and "inconsistent behavior".
[0121] Step 5: When the user authenticates to the server, the server needs to read the unique identification code of the mobile device and detect whether the behavioral feature model has been trained and matured.
[0122] Step 6: If the behavior feature model already has classification capabilities, that is, it is mature, the server collects the current user's behavior feature information for judgment.
[0123] Step 7: If the behavioral feature model does not yet have classification capabilities, that is, it is immature, the server prompts the user to write several characters and judges the user's handwriting.
[0124] Step 8: When the behavior feature model is mature, if the server's behavior feature model determines that the behavior is consistent, the user directly enters the application system.
[0125] Step 9: If the server's behavior feature model determines that the behavior is "inconsistent", the server prompts the user to write several characters and determines the user's handwriting through the handwriting model.
[0126] Step 10: If the behavior feature model is not yet mature or if the server's behavior feature model determines that the behavior is "inconsistent", if the server's handwriting model determines that the handwriting is "consistent", the user enters the system directly.
[0127] Step 11: If the behavior feature model is not yet mature or if the server's behavior feature model determines that the behavior is "inconsistent", if the server's handwriting model determines that the handwriting is "deviant", the server rejects the login request.
[0128] The implementation of the embodiment of the present invention includes the following beneficial effects: This embodiment includes two-level authentication. If the behavioral feature model is maturely trained, the behavioral feature information of the current object is authenticated at the first level through the trained behavioral feature model. If the behavioral feature model is not maturely trained or the first level authentication fails, the characters of the current object are authenticated at the second level through the trained handwriting judgment model. During the identity authentication process, the behavioral feature information and / or characters of the current object are authenticated without the need for password verification. The behavioral feature information and characters of the current object are attribute characteristics unique to the user object and have a strong correlation with the user object, thereby improving privacy.
[0129] See Figure 7 The embodiment of the present invention provides an identity authentication device based on multi-factor behavioral characteristics, which is applied to a server and includes:
[0130] at least one processor;
[0131] at least one memory for storing at least one program;
[0132] When the at least one program is executed by the at least one processor, the at least one processor implements the above method.
[0133] Among them, the memory is a non-transient computer-readable storage medium that can be used to store non-transient software programs and non-transient computer executable programs. The memory may include a high-speed random access memory and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory optionally includes a remote memory remotely arranged relative to the processor, and these remote memories can be connected to the processor via a network. Examples of the above-mentioned network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network and a combination thereof.
[0134] It can be seen that the contents of the above method embodiments are all applicable to the present device embodiments. The functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.
[0135] In addition, the embodiments of the present application further disclose a computer program product or computer program, which is stored in a computer-readable storage medium. The processor of a computer device can read the computer program from the computer-readable storage medium, and the processor executes the computer program, so that the computer device performs the above-mentioned method. Similarly, the contents of the above-mentioned method embodiment are all applicable to the present storage medium embodiment, and the functions specifically implemented by the present storage medium embodiment are the same as those of the above-mentioned method embodiment, and the beneficial effects achieved are also the same as those achieved by the above-mentioned method embodiment.
[0136] An embodiment of the present invention further provides a computer-readable storage medium, which stores a program executable by a processor. The program executable by the processor is used to implement the above method when executed by the processor.
[0137] It is understood that all or some steps, systems in the disclosed method above can be implemented as software, firmware, hardware and appropriate combinations thereof. Some physical components or all physical components can be implemented as software by a processor, such as a central processing unit, a digital signal processor or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application specific integrated circuit. Such software can be distributed on a computer-readable medium, and the computer-readable medium can include computer storage media (or non-transitory media) and communication media (or temporary media). As known to those of ordinary skill in the art, the term computer storage medium is included in any method or technology for storing information (such as computer-readable instructions, data structures, program modules or other data) and is volatile and non-volatile, removable and non-removable media. Computer storage media includes but is not limited to RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassette, magnetic tape, disk storage or other magnetic storage device, or can be used to store desired information and any other medium that can be accessed by a computer. Furthermore, as is well known to those skilled in the art, communication media typically embodies computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.
[0138] See Figure 1 The embodiment of the present invention provides an identity authentication system based on multi-factor behavioral characteristics, including a client and a server connected to the client; wherein,
[0139] The client is configured to collect behavior feature information and character information of an object, and send the behavior feature information and character information to the server;
[0140] The server includes:
[0141] at least one processor;
[0142] at least one memory for storing at least one program;
[0143] When the at least one program is executed by the at least one processor, the at least one processor implements the above method.
[0144] Specifically, the client may include at least one touch sensor, a locator and an angular velocity sensor, and the client may be specifically embodied as a smart mobile device, such as a mobile phone, a wearable device, etc.; and as for the computer device, it may be different types of electronic devices, including but not limited to desktop computers, laptops and other terminals.
[0145] It can be seen that the contents of the above method embodiments are all applicable to the present system embodiments. The functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.
[0146] It should be understood that in this application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.
[0147] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.
[0148] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.
[0149] The above is a specific description of the preferred implementation of the present invention, but the invention is not limited to the embodiments. Those skilled in the art can make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.
Claims
1. An identity authentication method based on multi-factor behavioral characteristics, characterized in that: Applicable to servers, including: Obtaining an identity authentication request, reading a client identification code according to the identity authentication request, and retrieving whether a behavioral feature model corresponding to the client is mature according to the identification code; If the behavioral feature model is maturely trained, the behavioral feature information of the current object is obtained, and the behavioral feature information is input into the maturely trained behavioral feature model for first-level authentication. If the first-level authentication is passed, the client is allowed to enter the application system; If the behavioral feature model is not mature or the first-level authentication fails, several characters of the current object are obtained and input into the trained handwriting judgment model for second-level authentication. If the second-level authentication passes, the client is allowed to enter the application system. If the second-level authentication fails, the login request is rejected.
2. The method according to claim 1, characterized in that The method further comprises: Obtaining an identity registration request, and obtaining the client's identifiable code, handwriting sample data, and behavioral feature sample data of the target object according to the identity registration request; A handwriting judgment model of the client corresponding to the recognizable code is trained according to the handwriting sample data, and a behavior feature model of the client corresponding to the recognizable code is trained according to the behavior feature sample data.
3. The method according to claim 1, characterized in that The handwriting judgment model is trained by the following method: Acquiring handwriting data of a target object and other objects, performing a first standardization process and labeling on the handwriting data to form handwriting sample data of the target object; The first preset model is trained according to the handwriting sample data of the target object to determine the model parameters of the first preset model; the first preset model includes a convolution layer, a pooling layer, a fully connected layer and an output layer.
4. The method according to claim 1, wherein The behavioral feature model is trained by the following method: Obtaining behavioral characteristic data of the target object; the behavioral characteristic data includes activity trajectory data, walking gait data, and screen operation behavior data; Preprocessing the behavior feature data of the target object to obtain behavior feature sample data of the target object; The second preset model is trained according to the behavioral feature sample data of the target object to determine the model parameters of the second preset model; the second preset model includes a convolutional layer, a pooling layer, a fully connected layer and an output layer.
5. The method according to claim 4, characterized in that Preprocessing the behavioral characteristic data of the target object includes: Performing coordinate transformation on the activity trajectory data to obtain a thermal density map, and generating a grayscale image from the thermal density map according to kernel density estimation; Performing Fourier transform on the walking gait data to generate a time-frequency spectrum; Encoding the screen operation behavior data into a trajectory vector diagram according to the touch trajectory coordinates and pressure values; The grayscale image, the time-frequency spectrum diagram and the trajectory vector diagram are subjected to a second normalization process and feature fusion.
6. The method according to claim 1, characterized in that Input the behavioral feature information into a well-trained behavioral feature model for first-level authentication, including: Inputting the behavioral feature information into a well-trained behavioral feature model to obtain the probability of similarity between the behavioral features of the current user and the target user; If the similarity probability of the behavioral characteristics is less than the first threshold, the first-level authentication fails; If the probability of the behavioral feature similarity is greater than or equal to the first threshold and less than the second threshold, the first-level authentication is passed and the number of authentications is recorded; if the number of authentications exceeds the preset number, the first-level authentication fails; If the behavior feature similarity probability is greater than or equal to the second threshold, the first-level authentication is passed.
7. The method according to claim 1, characterized in that Inputting a number of the characters into a trained handwriting recognition model for secondary authentication includes: Inputting a plurality of the characters into a trained handwriting judgment model to obtain a handwriting similarity probability value; If the handwriting similarity probability value is greater than or equal to the third threshold, the second-level authentication is passed; If the handwriting similarity probability value is less than the third threshold, the secondary authentication fails.
8. An identity authentication device based on multi-factor behavioral characteristics, characterized in that: Applicable to servers, including: at least one processor; at least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method according to any one of claims 1 to 7.
9. A computer-readable storage medium storing a program executable by a processor, characterized in that: The processor-executable program is configured to perform the method according to any one of claims 1 to 7 when executed by the processor.
10. An identity authentication system based on multi-factor behavioral characteristics, characterized in that: It includes a client and a server connected to the client; wherein, The client is configured to collect behavior feature information and character information of an object, and send the behavior feature information and character information to the server; The server includes: at least one processor; at least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method according to any one of claims 1 to 7.