Login authentication methods, devices, non-volatile storage media, and electronic devices
By detecting the similarity between user input behavior features detected on the terminal device and benchmark behavior features, and using the access token of the cloud server for login verification, the problem of excessive reliance on master password security in existing technologies for login verification is solved, thereby improving the security of the login process.
Patent Information
- Application Number
- CN202611231537.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-13
- Publication Date
- 2026-11-10
AI Technical Summary
Current login verification technologies rely excessively on the security of the master password, and once the master password is leaked, the security of the login process cannot be guaranteed.
By detecting user input behavior characteristics on the terminal device and comparing them with the benchmark behavior characteristics stored locally, a token acquisition request is generated after verification. The access token returned by the cloud server is used for login, and dynamic security tokens are used to replace static passwords.
It enhances the security of login verification, reduces the risk of batch credential loss due to master password leakage, and adds a non-transferable behavioral feature verification factor to the credential acquisition process, thereby improving the security of the login process.
Smart Images

Figure CN123074780A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of information security, and more specifically, to a login authentication method, apparatus, non-volatile storage medium, and electronic device. Background Technology
[0002] In related technologies, when verifying user login, there is an over-reliance on the security of the master password. Once the master password is leaked, the user's account cannot be protected.
[0003] There is currently no effective solution to the above problems. Summary of the Invention
[0004] This application provides a login verification method, apparatus, non-volatile storage medium, and electronic device to at least solve the technical problem that the login process cannot be guaranteed due to excessive reliance on the security of the master password during login verification in related technologies.
[0005] According to one aspect of the embodiments of this application, a login verification method is provided, comprising: when a password input operation of a target object is detected, determining the input behavior characteristics of the target object; determining the similarity between the input behavior characteristics and the baseline behavior characteristics stored locally on a terminal device; if the password entered by the target object passes verification and the similarity is greater than a preset similarity threshold, generating a token acquisition request; sending the token acquisition request to a cloud server, and receiving an access token returned by the cloud server corresponding to the token acquisition request, wherein the access token is used to log in to a target website.
[0006] Optionally, before determining the input behavior characteristics of the target object, the method further includes: when the target object registers the target password for the corresponding target website, guiding the target object to repeatedly enter the target password, and repeatedly collecting the input behavior information of the target object when entering the target password during the process of the target object repeatedly entering the target password; and determining the baseline behavior characteristics based on the input behavior information.
[0007] Optionally, the input behavior information includes at least one of the following: the key delay time series when the target object enters the target password, the movement trajectory of the mouse pointer, the acceleration characteristics of the mouse pointer, the sliding trajectory on the touch screen, and the pressure change information of the touch screen.
[0008] Optionally, generating a token acquisition request includes: determining an input behavior context corresponding to the input behavior features, wherein the input behavior context includes at least one of the following: device identification information of the terminal device, timestamp, target service provider identification corresponding to the target URL, and login location information; generating an initial token acquisition request based on the input behavior context; and encrypting the initial token acquisition request using the hardware key of the terminal device to obtain the token acquisition request.
[0009] Optionally, the method further includes: if the target URL does not support a tokenized authentication protocol, the cloud server uses the password registered by the target object for logging into the target URL to simulate logging into the target URL, and captures the session cookie or session token returned by the target URL after successful login; after receiving a token acquisition request for the target URL sent by the terminal device, the cloud server sends the session cookie or session token of the target URL as an access token to the terminal device.
[0010] Optionally, the method further includes: determining feature summary information of the baseline behavioral features; and sending the feature summary information to a cloud server.
[0011] Optionally, after sending a token acquisition request to the cloud server, the method further includes: the cloud server verifying the token acquisition request and determining the risk level corresponding to the token acquisition request; if the risk level is low, sending an access token to the terminal device; if the risk level is medium or high, sending an authentication instruction to the terminal device, and sending an access token to the terminal device if the target object passes the authentication.
[0012] According to another aspect of the embodiments of this application, a login verification device is also provided, applicable to a terminal device, comprising: a first processing module, configured to determine the input behavior characteristics of the target object when a password input operation of the target object is detected; a second processing module, configured to determine the similarity between the input behavior characteristics and the baseline behavior characteristics stored locally on the terminal device; a third processing module, configured to generate a token acquisition request when the password entered by the target object passes verification and the similarity is greater than a preset similarity threshold; and a fourth processing module, configured to send the token acquisition request to a cloud server and receive an access token returned by the cloud server corresponding to the token acquisition request, wherein the access token is used to log in to the target website.
[0013] According to another aspect of the embodiments of this application, a non-volatile storage medium is also provided, wherein a program is stored in the non-volatile storage medium, and the program controls the device where the non-volatile storage medium is located to execute a login verification method when it runs.
[0014] According to another aspect of the embodiments of this application, an electronic device is also provided, including: a memory and a processor, wherein the processor is configured to run a program stored in the memory, wherein the program executes a login verification method during runtime.
[0015] According to another aspect of the embodiments of this application, a computer program product is also provided, including a computer program that, when executed by a processor, implements the steps of a login verification method.
[0016] In this embodiment, when a password input operation is detected, the input behavior characteristics of the target object are determined; the similarity between the input behavior characteristics and the baseline behavior characteristics stored locally on the terminal device is determined; if the password entered by the target object passes verification and the similarity is greater than a preset similarity threshold, a token acquisition request is generated; the token acquisition request is sent to the cloud server, and an access token corresponding to the token acquisition request is received from the cloud server, wherein the access token is used to log in to the target website. By coordinating user behavior feature authentication based on the client's local location with a dynamic security token managed in the cloud, a non-transferable behavior feature verification factor is added to the credential acquisition process, and a dynamic token is used to replace the static password for transmission and login. This achieves the technical effect of improving the security of the login verification process and reducing the risk of batch credential loss due to master password leakage. In turn, it solves the technical problem that the security of the login process cannot be guaranteed due to the over-reliance on master password security during login verification in related technologies. Attached Figure Description
[0017] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:
[0018] Figure 1 This is a schematic diagram of the structure of a computer terminal (or mobile device) according to an embodiment of this application;
[0019] Figure 2 This is a flowchart illustrating a login verification method provided according to an embodiment of this application;
[0020] Figure 3 This is a schematic diagram of the structure of a login verification system provided according to an embodiment of this application;
[0021] Figure 4 This is a schematic diagram of a login verification device provided according to an embodiment of this application. Detailed Implementation
[0022] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.
[0023] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.
[0024] To better understand the embodiments of this application, the technical terms involved in the embodiments of this application are explained below:
[0025] Cloud Password Manager: A software suite that runs on user terminal devices (such as mobile phones and computers) and cloud servers to securely store, manage, and populate user login credentials across different websites and applications.
[0026] Master authentication factor: The basic credential required for users to access the core functions of the cloud password manager (such as unlocking the vault and synchronizing data), usually a master PIN code set by the user or a biometric feature (such as fingerprint).
[0027] Behavioral characteristics: User operating patterns exhibited during specific interactions that possess individual uniqueness and relative stability. Behavioral characteristics may include typing rhythm: the time interval sequence between keystrokes when a user enters a master PIN code or a specific string; mouse / touch movement patterns: the movement trajectory and acceleration characteristics of the mouse pointer, or the sliding trajectory and pressure changes on the touchscreen, when a user unlocks the interface or performs an authorization operation; the digital model formed by collecting and processing the above characteristics is called a behavioral characteristic model.
[0028] Dynamic security token: A time-sensitive digital credential bound to a specific session or device, used to replace static passwords for authentication. It may include OAuth tokens: Access and refresh tokens issued by the authorization server (the cloud system of this invention) to the client (user terminal) following the OAuth 2.0 or OpenID Connect protocol, used to access third-party resource servers; and session tokens: Credentials issued by a third-party website server after a user logs in, used to maintain the current login session.
[0029] Credential Vault: A secure database on a cloud server used to encrypt and store all user account credentials (including dynamic tokens, encrypted static passwords that cannot be replaced by tokens, etc.).
[0030] Risk Adaptive Engine: A software module on the cloud server that analyzes the context information of login requests (such as source IP geolocation, time, device fingerprint, requested account type, etc.) and assesses the risk level according to preset policies, thereby dynamically adjusting authentication requirements.
[0031] Mainstream cloud password management solutions in related technologies (such as LastPass, 1Password, Bitwarden, etc.) typically adopt the following technical architecture:
[0032] Client-server model: Users install plugins or applications on their terminal devices (clients), and all website accounts and passwords are encrypted with a key derived from the user's master password and stored on a cloud server.
[0033] Authentication and Synchronization Process: When a user needs to access a website, the client plugin is activated. The user first enters their master password (or verifies their identity via device biometrics) to decrypt the locally cached credential store or retrieve encrypted data blocks from the cloud. After successful verification, the client automatically fills in the corresponding username and password into the website's login form.
[0034] Core security assumptions: The security of this scheme relies heavily on the strength and confidentiality of the master password, and assumes that the encryption algorithm cannot be cracked. All synchronization and recovery operations are based on the principle that "whoever knows the master password owns it."
[0035] The login verification methods in related technologies mainly suffer from the following technical problems:
[0036] Single point of failure risk: The master password becomes the single critical weakness of the entire system. Once the master password is compromised due to guessing, keylogging, phishing, or user input on an insecure device, attackers can unlock and obtain all stored passwords, leading to a complete security collapse.
[0037] Static credential synchronization risk: Even if the password is encrypted during transmission and storage, its final form (the decrypted plaintext password) still needs to appear in the client's memory for padding. What is synchronized in the cloud is the ciphertext of the static password. If the encryption algorithm is broken in the future, or the cloud service provider is subjected to an internal attack, there is still a risk of large-scale data leakage.
[0038] The authentication process relies solely on "what is known" (master password) or "all" (logged-in devices). It cannot detect whether the login action was performed by the user themselves, and it lacks utilization of additional factors such as "where" (context) and "what is" (behavioral characteristics), making it difficult to achieve a better balance between convenience and security.
[0039] To address the aforementioned issues, this application provides relevant solutions, which are detailed below.
[0040] According to an embodiment of this application, a method embodiment for login verification is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.
[0041] The methods and embodiments provided in this application can be executed on mobile terminals, computer terminals, or similar computing devices. Figure 1 A hardware structure block diagram of a computer terminal (or mobile device) for implementing a login authentication method is shown. Figure 1 As shown, the computer terminal 10 (or mobile device 10) may include one or more processors 102 (shown as 102a, 102b, ..., 102n in the figure) 102 (processor 102 may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, computer terminal 10 may also include... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.
[0042] It should be noted that the aforementioned one or more processors 102 and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be integrated, in whole or in part, into any other element within the computer terminal 10 (or mobile device). As involved in the embodiments of this application, the data processing circuits serve as a processor control mechanism (e.g., selection of a variable resistor termination path connected to an interface).
[0043] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the login verification method in this embodiment. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby implementing the login verification method described above. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the computer terminal 10 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.
[0044] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the computer terminal 10. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a Radio Frequency (RF) module, used for wireless communication with the Internet.
[0045] The display may be, for example, a touchscreen liquid crystal display (LCD) that allows the user to interact with the user interface of the computer terminal 10 (or mobile device).
[0046] Under the above operating environment, this application embodiment provides a login verification method, such as... Figure 2 As shown, the method includes the following steps:
[0047] Step S202: When a password input operation of the target object is detected, determine the input behavior characteristics of the target object;
[0048] In some embodiments of this application, before determining the input behavior characteristics of the target object, the method further includes: guiding the target object to repeatedly enter the target password multiple times when registering the target password for the corresponding target website, and repeatedly collecting the input behavior information of the target object when entering the target password during the process of the target object repeatedly entering the target password; and determining the baseline behavior characteristics based on the input behavior information. The target object includes the user.
[0049] As an optional implementation, the input behavior information includes at least one of the following: the key delay time sequence when the target object enters the target password, the movement trajectory of the mouse pointer, the acceleration characteristics of the mouse pointer, the sliding trajectory on the touch screen, and the pressure change information of the touch screen.
[0050] In some embodiments of this application, taking a user logging into an online office application (i.e., the target URL) that supports the OAuth 2.0 protocol using a laptop as an example, during the system initialization and user registration phases, when the user sets up the cloud password manager for the first time, they need to set an easy-to-remember master password (master PIN code). This master password is both the authentication factor for unlocking the manager and the basis for collecting behavioral characteristics. The client will prompt the user to enter the master password several times. At the same time, the background behavioral characteristic collection module silently records the typing rhythm of each input by the user, specifically capturing the millisecond-level time interval sequence between key operations. These time delay characteristics constitute the user's unique keyboard operation behavior pattern.
[0051] After obtaining sufficient time-series samples, the client employs machine learning algorithms suitable for processing time-series data, such as Hidden Markov Models, to train the collected inter-key delay data, generating a user-specific behavioral feature model, which is then established as the baseline behavioral feature. This baseline behavioral feature is immediately encrypted after generation and stored only in a secure local storage area on the terminal device, such as a hardware-isolated environment like a TPM (Trusted Platform Module) chip or a secure enclave. This ensures that the complete baseline behavioral feature data never leaves the local device and is never uploaded to the cloud, fundamentally avoiding the risk of feature data being intercepted or leaked during network transmission or cloud storage.
[0052] In some embodiments of this application, the method further includes: determining feature summary information of the baseline behavioral features; and sending the feature summary information to a cloud server.
[0053] Optionally, the terminal device can extract a partial feature summary from the complete baseline behavioral features. This summary information has a certain distinguishing ability but is not enough to reverse-engineer the complete behavioral model. After uploading it to the cloud server, it can be used for quick pre-screening when a login request is initiated, helping the cloud risk adaptive engine to initially judge the degree of abnormality of the request. The complete model always completes the final accurate comparison locally.
[0054] For example, when a terminal device initiates a token acquisition request, a corresponding feature summary can be extracted in real time from the user's current input behavior (i.e., using the same dimensionality reduction or hashing method as the baseline feature summary to extract simplified features of the current typing rhythm), and this real-time feature summary is uploaded to the cloud server along with the token acquisition request. Upon receiving the request, the cloud server compares the real-time feature summary with the pre-stored baseline behavioral feature summary for the device (e.g., by calculating the Euclidean distance or cosine similarity between feature vectors) to obtain a pre-screening score. This score serves as an input factor for the risk adaptive engine, which combines contextual information such as device identification, geographical location, and time to comprehensively determine the risk level of the request. If the summary similarity is too low, indicating a significant deviation between the current input behavior and the behavior pattern during registration, the engine will directly mark it as high-risk or require supplementary identity verification; if the similarity is within the normal range, the process proceeds to the next stage, where a local complete behavioral feature model performs the final accurate comparison. This two-layer judgment method—cloud coarse screening + local fine inspection—utilizes the low latency advantage of cloud pre-screening while ensuring that complete behavioral features do not leave the user's device, thus achieving a balance between security and efficiency.
[0055] In some embodiments of this application, after the behavioral feature baseline is established, the user begins adding an account to the online office application. The client guides the user through the standard OAuth 2.0 authorization process. After authorization, the token management module of the cloud server obtains the access token (dynamic security token) and refresh token from the application's authorization server and encrypts and stores these tokens in the user's credential vault (credential vault), without storing the user's original password throughout the process. Through the above process, the system completes the comprehensive collection of input behavioral information and the establishment of baseline behavioral features during the registration phase, providing a highly reliable reference for the real-time comparison of the collected behavioral features with the baseline behavioral features during subsequent login verification.
[0056] Step S204: Determine the similarity between the input behavior features and the baseline behavior features stored locally on the terminal device;
[0057] In some embodiments of this application, when a user visits the login page of a target website, the client browser extension is activated, and the extension interface pops up, prompting the user to enter the master password (master PIN code) to unlock. During the process of the user entering the master password, the client collects the typing rhythm of this input in real time as input behavior features, and compares the input behavior features with the baseline behavior features stored locally on the terminal device in real time to calculate a similarity score. At the same time, the client verifies the correctness of the master password entered by the user.
[0058] Step S206: If the password entered by the target object passes verification and the similarity is greater than a preset similarity threshold, generate a token acquisition request.
[0059] Optionally, during local authentication, the client (i.e., the terminal device) will collaboratively determine the master password (master PIN) verification result and the behavioral feature similarity score: only when the master password is verified correctly and the similarity score exceeds a preset similarity threshold will the local authentication be considered successful. If the similarity score does not reach the threshold, even if the master password is entered correctly, the client will determine that the local authentication has failed and may trigger a security alert or prompt the user to complete identity verification through secondary authentication methods such as biometrics on the terminal device.
[0060] In some embodiments of this application, generating a token acquisition request includes: determining an input behavior context corresponding to the input behavior features, wherein the input behavior context includes at least one of the following: device identification information of the terminal device, timestamp, target service provider identification corresponding to the target URL, and login location information; generating an initial token acquisition request based on the input behavior context; and encrypting the initial token acquisition request using the hardware key of the terminal device to obtain the token acquisition request.
[0061] Optionally, after successful local authentication, the client determines the input behavior context for this login request. This context includes the terminal device's device identification information (device ID), the current timestamp, the target service provider identifier (target SP identifier) corresponding to the target URL the user wants to log in to, and the login location information. Based on this input behavior context, the client generates an initial token acquisition request. It then calls the terminal device's built-in hardware key module (such as a TPM security chip) to encrypt (sign) this initial token acquisition request, thus obtaining the final token acquisition request. Finally, the client sends this token acquisition request to the cloud server to complete the subsequent token issuance process.
[0062] Step S208: Send a token acquisition request to the cloud server and receive an access token corresponding to the token acquisition request returned by the cloud server, wherein the access token is used to log in to the target website.
[0063] As an optional implementation, the method further includes: when the target URL does not support a tokenized authentication protocol, the cloud server uses the password registered by the target object for logging into the target URL to simulate logging into the target URL, and captures the session cookie or session token returned by the target URL after successful login; after receiving a token acquisition request for the target URL sent by the terminal device, the cloud server sends the session cookie or session token of the target URL as an access token to the terminal device.
[0064] Optionally, when the target URL does not support tokenized authentication protocols (such as OAuth), a dynamic token emulation scheme can be used to obtain an access token. Specifically, when a user adds an account to such a target URL, they can be required to provide a login password for that URL once. This password is securely encrypted and stored on a cloud server. After obtaining the password provided by the user, the cloud server uses this password to initiate a simulated login to the target URL. Upon successful login, it captures the session cookie or session token returned by the target URL and encrypts and stores it as an access token (dynamic security token) in a credential vault (credential vault). This access token has the same validity period as the original session on the target URL. When the user subsequently requests to log in to the target URL, the client, after completing behavioral feature verification and local authentication, sends a token retrieval request to the cloud server. After verifying the risk level, the cloud server retrieves the corresponding session cookie or session token from the credential vault, encrypts it with a session key, and issues it to the client as an access token. The client's browser extension then embeds this token into the browser, thereby bypassing the login page and achieving automatic login. In addition, the validity of the session cookie or session token will be continuously monitored. When the session expires, the cloud server will automatically use the previously stored password to simulate logging into the target website again, obtain new session credentials and update the stored access token. This process is transparent to the user and does not require the user to re-enter the password or perform any additional operations.
[0065] In some embodiments of this application, after sending a token acquisition request to the cloud server, the method further includes: the cloud server verifying the token acquisition request and determining the risk level corresponding to the token acquisition request; if the risk level is low, sending an access token to the terminal device; if the risk level is medium or high, sending an authentication instruction to the terminal device, and sending an access token to the terminal device if the target object passes authentication.
[0066] Optionally, upon receiving a token acquisition request, the cloud server first verifies the device signature. Then, it analyzes information in the request context, such as the terminal device identifier, timestamp, target service provider identifier, and login location, to determine whether the login was initiated from a frequently used device, common geographical location, and typical time, and accordingly determines the request's risk level. If the risk level is low, the cloud server directly approves the request; if it is medium or high risk, it sends an authentication instruction to the terminal device, such as pushing a secondary confirmation to the user's mobile phone, and approves the request only after the target device completes additional authentication. After the request is approved, the cloud server's token management module retrieves the access token (dynamic security token) corresponding to the target URL from the credential vault, encrypts it using the session key, and sends it to the client. The client decrypts the token to obtain the access token used to log in to the target URL.
[0067] By employing a method that determines the input behavior characteristics of a target object upon detecting password input, determines the similarity between these input behavior characteristics and baseline behavior characteristics stored locally on the terminal device, generates a token acquisition request when the target object's password passes verification and the similarity exceeds a preset similarity threshold, sends the token acquisition request to the cloud server, and receives an access token corresponding to the token acquisition request returned by the cloud server (where the access token is used to log in to the target website), this approach combines client-local user behavior characteristic authentication with a cloud-managed dynamic security token. This adds a non-transferable behavioral characteristic verification factor to the credential acquisition process and uses a dynamic token instead of a static password for transmission and login. This achieves the technical effect of improving the security of the login verification process and reducing the risk of batch credential loss due to master password leakage. Furthermore, it solves the technical problem of insufficient security in the login process caused by over-reliance on master password security in related technologies.
[0068] Furthermore, the method provided in this application embodiment enhances security: even if an attacker steals a user's master password (master PIN), they cannot pass local behavioral feature verification without a registered device and the ability to mimic the user's specific typing rhythm, thus the attack fails. The core credentials for cloud storage and synchronization are access tokens (dynamic security tokens), such as OAuth tokens or session cookies, rather than the original static password. These tokens are time-sensitive, revocable, and usually bound to a specific device or session. Even if the cloud vault is breached, the attacker obtains tokens that are difficult to use directly or expire quickly, making it impossible to recover the user's original password. This achieves the security effect of "password not being stored in the ground" or "password minimization retention." During the login transmission process, tokens or encrypted session data are transmitted, avoiding the exposure of passwords in plaintext and thus reducing the risk of network eavesdropping. While enhancing security, the user experience is also optimized: behavioral authentication is seamlessly integrated into the user's master password input process, requiring no additional steps; for applications supporting modern authentication protocols such as OAuth, the login process completely eliminates the need for username / password forms, enabling one-click login and a smoother workflow; when a user restores access on a new device, behavioral characteristics must first be collected and verified on that device, or authorization must be granted by a trusted device, thus ensuring the security of cross-device synchronization. Furthermore, the risk-adaptive engine can intelligently adjust security policies based on the login context, achieving a dynamic balance between security and convenience.
[0069] This application also provides a login verification system. Figure 3 This is a schematic diagram of the system structure. From Figure 3 As can be seen from this, the system includes a user terminal 30, a cloud server 32, and a third-party service provider 34.
[0070] exist Figure 3 In the login verification system shown, the user terminal 30 serves as the interaction entry point and is equipped with client software or a browser extension. This terminal has a behavior feature acquisition module 302 deployed within it. This module can silently and continuously capture the user's operating habits when interacting with the input device. For the keyboard, it can record the precise moment each key is pressed and released, thereby calculating the time interval sequence between adjacent keys, i.e., the typing rhythm. These millisecond-level time differences vary from person to person but are relatively stable. For the mouse or touchscreen, it can similarly record the coordinate sequence of the movement trajectory, acceleration changes, pressure sensitivity, and other parameters. Simultaneously, the terminal also has a local secure storage area 304, which can physically be a TPM chip, a secure enclave, or a virtualization-based isolated area. This storage area is used to securely store encrypted baseline behavior feature models, device private keys, and other sensitive credentials required for operation locally. This data is invisible or only accessible with limited access to external operating systems and applications, ensuring that even if the device is infected with malware, critical data is difficult to steal or tamper with.
[0071] The cloud management server is deployed on cloud infrastructure and can adopt a microservice architecture, internally divided into multiple collaborative software modules. The user management module 322 is responsible for the lifecycle management of user identities, including device binding during user registration and storing user public keys for subsequent signature verification. It also maintains basic user attributes and a list of authorized devices. The risk adaptive engine 324 is the decision-making hub, integrating a hybrid rule-based and machine learning model that can receive rich contextual parameters and output risk scores. Its main evaluation dimensions include: whether the device fingerprint initiating the request belongs to the user's frequently used device set; the deviation between the source IP geolocation of the request and the user's historical location trajectory; whether the current time is within the user's usual active period; the historical access frequency and sensitivity level of the target application; and whether there have been any recent abnormal login attempts. The token management module 326 is responsible for protocol interaction with the authentication endpoint of the third-party service provider 34, automatically completing the OAuth 2.0 authorization code process or client credential process, securely obtaining, storing, and refreshing access tokens (dynamic security tokens); for websites that do not support standard protocols, it can also have browser automation capabilities for simulating logins and capturing session cookies. Vault 328 is a high-security encrypted database. All dynamic tokens are encrypted and stored using a user-unique data encryption key, which is protected by the system's root key. This achieves user- and application-specific encryption isolation. Vault also supports key rotation and access auditing.
[0072] Third-party service providers (34) broadly refer to various websites or applications that users need to access, and can expose login pages and APIs. For applications that support the OAuth 2.0 protocol, they can provide authentication endpoints and issue bearer tokens conforming to the RFC 6749 standard; for legacy applications, the traditional Cookie-Session model is used. The system obtains and injects credentials by interacting with these external endpoints.
[0073] When a user accesses the login page of the online office application using a laptop's browser, the client browser extension is activated by the page load event. Upon detecting that the current domain belongs to a known target URL requiring management, the extension immediately displays an unlock screen, obscuring the original login form, and prompts the user to enter their master password (master PIN) to activate the cloud password manager. Since the user has already completed initial initialization and behavior registration on this device, a unique baseline behavioral characteristic model of that user is stored in the local secure storage area 304.
[0074] As the user begins entering their master password, the behavior feature acquisition module 302 is activated, recording the keyboard event flow. It captures the timestamps of each "press" and "release" event, calculating a series of time interval differences in real time to form a feature vector representing the input rhythm. Simultaneously, the background logic of the unlock interface extracts the user-input master password string, processes it using a one-way hash function, and compares it with the master password hash value previously stored in the local secure storage area 304 to verify password correctness.
[0075] Optionally, the local collaborative decision-making process runs in a secure environment or trusted execution environment on the terminal CPU. The client compares the feature vectors collected in real time with the baseline behavioral feature model decrypted and loaded from the local secure storage area. Since the length and speed of different inputs may vary slightly, the system uses a dynamic time warping algorithm to align the time axis and calculates a cosine similarity or normalized Euclidean distance score as the similarity score of the current input behavior. At the same time, the system verifies the hash comparison result of the master password. The final local authentication result is determined by two conditions: the master password matches, and the similarity score exceeds a preset threshold (this threshold is dynamically set during the registration phase by analyzing the standard deviation of multiple user samples and the system security policy, for example, 0.85). The local authentication status is updated to pass only when both conditions are met. If the master password is correct but the similarity score is below the threshold, the system will determine that the current input behavior is significantly inconsistent with the owner, and it is possible that an attacker has guessed the password and attempted to log in. At this point, the system can trigger a preset response strategy, such as displaying an error message on the interface and disabling retry for a few seconds, or automatically initiating biometric secondary authentication based on user pre-configuration (e.g., sending a request to the laptop's built-in fingerprint scanner to require the user to verify their fingerprint). Local authentication is only considered successful after the user passes biometric authentication, providing a disaster recovery channel in case of behavioral feature drift (e.g., user finger injury).
[0076] Once local authentication is successful, the client immediately initiates the construction of a token acquisition request. This can begin by collecting the input behavior context of the current session, specifically including: device identification information obtained through system API calls (a device fingerprint generated by integrating hardware serial number, OS installation ID, etc.); a precise timestamp synchronized from the network time protocol; the target service provider identifier of the target online office application read from the extended configuration; and city-level login location information obtained through IP location services. To further enhance security, the client can also choose to compute a compact feature digest from the real-time behavioral feature vector. This feature digest is typically generated using a lightweight hash function or feature reduction algorithm (such as random projection), and while it may not be able to reconstruct the complete behavioral model, it retains sufficient distinguishing information for rapid pre-screening in the cloud. The client packages the aforementioned context information along with the feature digest to form the initial token acquisition request. Subsequently, the client calls the hardware key module in the local secure storage area 304, using the device private key stored within this module to digitally sign the initial request. The signing operation is performed in an isolated area of the secure chip, and the private key, generated during registration, never leaves the chip. The signed data packet is the complete token acquisition request, which is sent to the API endpoint of the cloud management server via an HTTPS secure connection.
[0077] Upon receiving a request, the cloud management server's transport security layer first decrypts the HTTPS and then sends the request payload to the user management module 322. This module retrieves the public key bound to the device from the user database based on the device identifier in the request and performs signature verification. If verification fails, it indicates that the request may have been tampered with or originated from a spoofed device; the server will record an alert and return an error. If verification is successful, the user management module 322 parses the plaintext context and, along with a behavioral feature digest (if any), forwards it to the risk adaptive engine 324.
[0078] The risk adaptive engine 324 initiates a multi-factor risk assessment process. It can perform profile matching based on device fingerprints. If the current device is one of the user's most frequently used devices in the last 30 days, the device reputation score is high. Geographical analysis of the login location shows the IP address is in city A, and the user frequently operates in cities A and B. Furthermore, the current location's offset from the user's home network IP is less than 100 kilometers, so the location reputation score is also high. Time analysis reveals the request occurred at 10 AM on Tuesday, consistent with the user's weekday activity patterns. The target application is an office application, falling into the low-to-medium sensitivity category. Additionally, if the client sends a feature summary, the engine retrieves the historical summary sequence previously uploaded by the device from the database and calculates the statistical deviation between the current summary and the historical summary set. If the deviation is within the normal fluctuation range, an additional positive gain is generated. Based on the above, if the engine calculates a risk value below the preset low-risk threshold, it is directly marked as low-risk and the request is approved. If the risk value falls into the medium-risk range, the engine will temporarily deny approval and send a push notification to the user's linked auxiliary device (such as a mobile phone), requiring the user to confirm within a limited time; if the confirmation expires or is rejected, the request is rejected. In high-risk situations, requests are directly rejected and may trigger account freezing. The engine's decision-making logic is not static and can adaptively adjust based on the global attack landscape over a period of time.
[0079] After the risk adaptive engine 324 makes an approval decision, it requests the token management module 326. This module parses the target service provider identifier and queries the credential vault 328 for the user's credential record for this application. The record shows that an access token obtained through the OAuth 2.0 authorization code process already exists, and its validity period has more than 1 hour remaining. Therefore, the module retrieves the encrypted access token from the vault and decrypts it using the vault's master key to obtain the plaintext token. Next, the token management module 326 generates a temporary symmetric session key for this response, encrypts the access token using the AES-GCM algorithm, and generates a ciphertext token. Then, it uses the terminal device's public key obtained from the user management module 322 to encrypt the session key using RSA-OAEP. Finally, the two ciphertexts (token ciphertext and key ciphertext) are packaged together in the response body and returned to the client. At the same time, if the module checks that the refresh token of this access token will expire in 24 hours, it will asynchronously initiate a background refresh process to ensure that the token remains available.
[0080] Upon receiving the response, the client first decrypts the session key using its local device private key, and then uses the session key to decrypt the plaintext access token. At this point, the client holds a valid access token for the online office application. Because the application supports OAuth 2.0, the browser extension employs a final step for passwordless automatic login: the extension calls the browser's WebRequest API to intercept requests sent to the application resource server and adds "Authorization:Bearer {access token}" to the request header; alternatively, it works in conjunction with the application's OAuth callback endpoint to inject the token into the page's JavaScript environment, triggering immediate login activation via the page script. After refreshing the browser page, the user directly enters the application's working interface without needing to touch any password input fields.
[0081] If the target website is a traditional website that does not support tokenized authentication protocols such as OAuth, the interaction process differs, but the user terminal 30 will hardly notice it. During the account registration and addition phase, the user is required to enter their account and password for the website once. The terminal encrypts these and uploads them to the credential vault 328 on the cloud server 32 via a secure channel. The token management module 326 has a built-in network automation engine that uses the stored password to submit authentication data to the website's login interface in the form of a headless browser or HTTP client. After successful login, the Set-Cookie field is captured from the response header, and multiple session cookie values are extracted. These cookies are essentially session tokens issued by the website, typically with a validity period of several hours. The token management module 326 treats them as a special type of access token (i.e., pseudo-dynamic token), annotates their validity period, domain name, and scope, encrypts them, and stores them in the credential vault 328. When the user subsequently initiates a login request, the local behavior verification and cloud risk decision-making processes are consistent with the above. When the token management module 326 receives a request, it retrieves the corresponding website's session cookie from the vault, encrypts it, and sends it to the client. After receiving the data, the client browser extension calls the browser's cookie setting interface to implant cookies into the target domain one by one, and can selectively set attributes such as HttpOnly and Secure to comply with the website's security policy. After implantation, when the user opens the website again, the browser naturally carries the valid cookie and directly enters the login page. To maintain this seamless experience, the token management module 326 also runs a scheduled task that periodically traverses the vault for pseudo-dynamic token records that are about to expire. Once it detects that the validity window of a cookie is insufficient (e.g., the remaining time is less than 30%), the module retrieves the original password from the vault, performs a simulated login again, captures a new cookie, and replaces the stored one. This refresh operation is all completed silently in the cloud background, and the number of retries and failure alarms can be set. The user is never aware of the password's existence, but the system securely proxies the lifecycle of the credential.
[0082] This application provides a login verification device suitable for use in terminal devices. Figure 4 This is a schematic diagram of the device. Figure 4As can be seen from the diagram, the device includes: a first processing module 40, used to determine the input behavior characteristics of the target object when a password input operation of the target object is detected; a second processing module 42, used to determine the similarity between the input behavior characteristics and the baseline behavior characteristics stored locally on the terminal device; a third processing module 44, used to generate a token acquisition request when the password entered by the target object passes verification and the similarity is greater than a preset similarity threshold; and a fourth processing module 46, used to send a token acquisition request to a cloud server and receive an access token returned by the cloud server corresponding to the token acquisition request, wherein the access token is used to log in to the target website.
[0083] In some embodiments of this application, before determining the input behavior characteristics of the target object, the first processing module 40 is further configured to: guide the target object to repeatedly enter the target password when the target object registers the target password for the corresponding target website, and repeatedly collect the input behavior information of the target object when entering the target password during the process of the target object repeatedly entering the target password; and determine the baseline behavior characteristics based on the input behavior information.
[0084] In some embodiments of this application, the input behavior information includes at least one of the following: the key delay time sequence when the target object enters the target password, the movement trajectory of the mouse pointer, the acceleration characteristics of the mouse pointer, the sliding trajectory on the touch screen, and the pressure change information of the touch screen.
[0085] In some embodiments of this application, the step of the third processing module 44 generating a token acquisition request includes: determining an input behavior context corresponding to the input behavior features, wherein the input behavior context includes at least one of the following: device identification information of the terminal device, timestamp, target service provider identification corresponding to the target URL, and login location information; generating an initial token acquisition request based on the input behavior context; and encrypting the initial token acquisition request using the hardware key of the terminal device to obtain the token acquisition request.
[0086] In some embodiments of this application, the fourth processing module 46 is further configured to: determine feature summary information of the baseline behavioral features; and send the feature summary information to a cloud server.
[0087] It should be noted that each module in the above login verification device can be a program module (e.g., a set of program instructions to implement a specific function) or a hardware module. For the latter, it can take the following forms, but is not limited to them: each of the above modules is represented by a processor, or the functions of each of the above modules are implemented by a processor.
[0088] According to another aspect of the embodiments of this application, a non-volatile storage medium is also provided, which stores a program. When the program runs, it controls the device where the non-volatile storage medium is located to execute the following login verification method: when a password input operation of a target object is detected, the input behavior characteristics of the target object are determined; the similarity between the input behavior characteristics and the baseline behavior characteristics stored locally on the terminal device is determined; if the password entered by the target object passes verification and the similarity is greater than a preset similarity threshold, a token acquisition request is generated; the token acquisition request is sent to a cloud server, and an access token corresponding to the token acquisition request is returned by the cloud server, wherein the access token is used to log in to the target website.
[0089] According to another aspect of the embodiments of this application, an electronic device is also provided, including: a memory and a processor, wherein the processor is configured to run a program stored in the memory, wherein the program executes the following login verification method: when a password input operation of a target object is detected, determining the input behavior characteristics of the target object; determining the similarity between the input behavior characteristics and the baseline behavior characteristics stored locally on the terminal device; if the password entered by the target object passes verification and the similarity is greater than a preset similarity threshold, generating a token acquisition request; sending the token acquisition request to a cloud server, and receiving an access token returned by the cloud server corresponding to the token acquisition request, wherein the access token is used to log in to the target website.
[0090] According to another aspect of the embodiments of this application, a computer program product is also provided, including a computer program that, when executed by a processor, implements the following login verification method steps: when a password input operation of a target object is detected, determining the input behavior characteristics of the target object; determining the similarity between the input behavior characteristics and the baseline behavior characteristics stored locally on the terminal device; if the password entered by the target object passes verification and the similarity is greater than a preset similarity threshold, generating a token acquisition request; sending the token acquisition request to a cloud server and receiving an access token returned by the cloud server corresponding to the token acquisition request, wherein the access token is used to log in to the target website.
[0091] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.
[0092] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.
[0093] The units described as separate components may or may not be physically separate. 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 units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0094] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.
[0095] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to related technologies, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.
[0096] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.
Claims
1. A login verification method, characterized in that, include: When a password input operation is detected from a target object, the input behavior characteristics of the target object are determined; Determine the similarity between the input behavior features and the baseline behavior features stored locally on the terminal device; If the password entered by the target object is verified and the similarity is greater than a preset similarity threshold, a token acquisition request is generated. Send the token acquisition request to the cloud server and receive the access token corresponding to the token acquisition request returned by the cloud server, wherein the access token is used to log in to the target website.
2. The login verification method according to claim 1, characterized in that, Before determining the input behavior features of the target object, the method further includes: When the target object registers the target password corresponding to the target website, the target object is guided to repeatedly enter the target password, and during the process of the target object repeatedly entering the target password, the input behavior information of the target object when entering the target password is repeatedly collected; The baseline behavioral features are determined based on the input behavioral information.
3. The login verification method according to claim 2, characterized in that, The input behavior information includes at least one of the following: the key delay time sequence when the target object inputs the target password, the movement trajectory of the mouse pointer, the acceleration characteristics of the mouse pointer, the sliding trajectory on the touch screen, and the pressure change information of the touch screen.
4. The login verification method according to claim 1, characterized in that, The token generation request includes: Determine the input behavior context corresponding to the input behavior feature, wherein the input behavior context includes at least one of the following: device identification information of the terminal device, timestamp, target service provider identification corresponding to the target URL, and login location information; An initial token acquisition request is generated based on the input behavior context; The initial token acquisition request is encrypted using the hardware key of the terminal device to obtain the token acquisition request.
5. The login verification method according to claim 1, characterized in that, The method further includes: If the target URL does not support the tokenized authentication protocol, the cloud server uses the password registered by the target object for logging into the target URL to simulate logging into the target URL, and captures the session cookie or session token returned by the target URL after successful login; Upon receiving the token acquisition request for the target URL from the terminal device, the session cookie or session token of the target URL is sent to the terminal device as an access token.
6. The login verification method according to claim 1, characterized in that, The method further includes: Determine the feature summary information of the benchmark behavioral features; The feature summary information is sent to the cloud server.
7. The login verification method according to claim 1, characterized in that, After sending the token acquisition request to the cloud server, the method further includes: The cloud server verifies the token acquisition request and determines the risk level corresponding to the token acquisition request; If the risk level is low, the access token is sent to the terminal device; If the risk level is medium or high, an authentication instruction is sent to the terminal device, and if the target object passes the authentication, the access token is sent to the terminal device.
8. A login verification device, suitable for use in terminal equipment, characterized in that, include: The first processing module is used to determine the input behavior characteristics of the target object when a password input operation of the target object is detected; The second processing module is used to determine the similarity between the input behavior features and the baseline behavior features stored locally on the terminal device; The third processing module is used to generate a token acquisition request when the password entered by the target object passes verification and the similarity is greater than a preset similarity threshold. The fourth processing module is used to send the token acquisition request to the cloud server and receive the access token corresponding to the token acquisition request returned by the cloud server, wherein the access token is used to log in to the target website.
9. A non-volatile storage medium, characterized in that, The non-volatile storage medium stores a program, wherein when the program is executed, it controls the device where the non-volatile storage medium is located to execute the login verification method according to any one of claims 1 to 7.
10. An electronic device, characterized in that, include: A memory and a processor, the processor being configured to run a program stored in the memory, wherein the program, when running, executes the login verification method according to any one of claims 1 to 7.
11. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the login verification method according to any one of claims 1 to 7.
