Identity authentication method, device and equipment based on head tremor biological characteristics
By collecting and analyzing user head tremor signals, and utilizing denoising and event segmentation techniques combined with a pre-trained model, the inconvenience and vulnerability to eavesdropping in virtual reality device authentication are solved, providing a secure and efficient authentication solution.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-05
- Publication Date
- 2026-03-27
AI Technical Summary
Existing authentication methods for virtual reality devices are cumbersome, time-consuming, and easily spied on. Furthermore, biometric authentication methods require expensive external sensors, making them difficult to popularize.
Wearable devices are used to collect head tremor signals, and biometric features are extracted through denoising and event segmentation. A pre-trained authentication model is then used to calculate similarity and determine the user's identity.
It achieves convenient and secure identity authentication, reduces the risk of being spied on, and is suitable for widespread deployment.
Smart Images
Figure CN121744288A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the fields of biometrics and cybersecurity, and in particular to an identity authentication method, device, and equipment based on head tremor biometrics. Background Technology
[0002] Virtual reality technology is rapidly integrating into various fields such as entertainment, education, and healthcare, leading to the migration of a large amount of sensitive user data and functions to virtual reality devices. Therefore, providing a secure and convenient identity authentication system for virtual reality devices is crucial. Virtual reality device authentication schemes in related technologies, such as password or pattern unlocking, typically rely on handheld controllers or air gestures for input; some virtual reality device authentication schemes attempt to introduce biometrics for authentication.
[0003] However, password or pattern unlocking methods in related technologies are not only cumbersome and time-consuming, but also pose serious security risks as user input actions can be easily captured by bystanders or cameras. In addition, biometric authentication methods in related technologies generally require users to perform time-consuming specific tasks or rely on expensive, non-standard external sensors, making them difficult to popularize in commercial virtual reality devices and urgently need to be addressed. Summary of the Invention
[0004] This application provides a method, device, and equipment for identity authentication based on head tremor biometrics to solve the problems of inconvenient identity authentication and susceptibility to eavesdropping on head-mounted devices.
[0005] The first aspect of this application provides an identity authentication method based on head tremor biometrics, including the following steps: The device collects head tremor signals from the current user using wearable devices. The head tremor signal of the current user is denoised and segmented to obtain a first target tremor signal. A first biometric feature is extracted from the first target tremor signal and input into a pre-trained authentication model. The similarity between the first biometric feature and the pre-stored user template feature is calculated based on the pre-trained authentication model to obtain the authentication result. The authentication result is used to determine whether the current user is a legitimate user.
[0006] Optionally, the head tremor signal is denoised, including: Power spectral density analysis was performed on the head tremor signal, and motion artifact noise in the head tremor signal was filtered out based on the analysis results; Residual noise in the head tremor signal is filtered out using the maximum overlap discrete wavelet transform technique.
[0007] Optionally, the head tremor signal is subjected to event segmentation processing, including: Calculate the aggregated signal variance of the multi-axis sensor signals in the wearable device; The start and end points of the head tremor signal are determined based on the aggregated signal variance and a preset dual threshold method, and segments of the head tremor signal that meet preset valid conditions are located and segmented from the head tremor signal.
[0008] Optionally, the step of collecting the current user's head tremor signal using a wearable device includes: The wearable device is used to collect the initial tremor signal generated by the current user based on the rhythm of a preset challenge signal; Determine whether the initial tremor signal and the preset challenge signal meet the preset synchronization conditions; If the initial tremor signal and the preset challenge signal meet the preset synchronization condition, then the initial tremor signal is used as the head tremor signal of the current user.
[0009] Optionally, before calculating the similarity between the first biometric feature and the pre-stored user template features based on the pre-trained authentication model, the method further includes: Acquire head tremor signals from multiple target users; The head tremor signals of the multiple target users are denoised and segmented to obtain a second target tremor signal, and a second biometric feature is extracted from the second target tremor signal. The Siamese network training model is trained by receiving positive and negative sample pairs of the second biometric feature, thereby minimizing the distance between the positive sample pairs and maximizing the distance between the negative sample pairs, thus obtaining the trained Siamese network training model; wherein, the positive sample pair consists of two sample features from the same legitimate user, and the negative sample pair consists of one sample from a legitimate user and one sample from an illegitimate user. The trained Siamese network training model is optimized using a preset contrastive loss function to obtain the pre-trained authentication model.
[0010] A second aspect of this application provides an identity authentication device based on head tremor biometrics, comprising: The acquisition module is used to acquire the current user's head tremor signals using wearable devices; The determination module is used to perform noise reduction and event segmentation processing on the head tremor signal of the current user to obtain a first target tremor signal, extract a first biometric feature from the first target tremor signal, and input the first biometric feature into a pre-trained authentication model to calculate the similarity between the first biometric feature and the pre-stored user template feature based on the pre-trained authentication model to obtain the authentication result. The judgment module is used to determine whether the current user is a legitimate user based on the authentication result.
[0011] Optionally, the determining module is specifically used for: Power spectral density analysis was performed on the head tremor signal, and motion artifact noise in the head tremor signal was filtered out based on the analysis results; Residual noise in the head tremor signal is filtered out using the maximum overlap discrete wavelet transform technique.
[0012] Optionally, the determining module is specifically used for: Calculate the aggregated signal variance of the multi-axis sensor signals in the wearable device; The start and end points of the head tremor signal are determined based on the aggregated signal variance and a preset dual threshold method, and segments of the head tremor signal that meet preset valid conditions are located and segmented from the head tremor signal.
[0013] Optionally, the acquisition module is specifically used for: The wearable device is used to collect the initial tremor signal generated by the current user based on the rhythm of a preset challenge signal; Determine whether the initial tremor signal and the preset challenge signal meet the preset synchronization conditions; If the initial tremor signal and the preset challenge signal meet the preset synchronization condition, then the initial tremor signal is used as the head tremor signal of the current user.
[0014] Optionally, before calculating the similarity between the first biometric feature and the pre-stored user template features based on the pre-trained authentication model, the determining module is further configured to: Acquire head tremor signals from multiple target users; The head tremor signals of the multiple target users are denoised and segmented to obtain a second target tremor signal, and a second biometric feature is extracted from the second target tremor signal. The Siamese network training model is trained by receiving positive and negative sample pairs of the second biometric feature, thereby minimizing the distance between the positive sample pairs and maximizing the distance between the negative sample pairs, thus obtaining the trained Siamese network training model; wherein, the positive sample pair consists of two sample features from the same legitimate user, and the negative sample pair consists of one sample from a legitimate user and one sample from an illegitimate user. The trained Siamese network training model is optimized using a preset contrastive loss function to obtain the pre-trained authentication model.
[0015] A third aspect of this application provides an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being configured to perform an identity authentication method based on head tremor biometrics as described in the above embodiments.
[0016] A fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement the head tremor-based biometric authentication method as described in the above embodiments.
[0017] A fifth aspect of this application provides a computer program product storing a computer program that, when executed by a processor, implements the identity authentication method based on head tremor biometrics as described in the above embodiments.
[0018] Therefore, this embodiment utilizes a wearable device to collect the head tremor signal of the current user, and performs denoising and event segmentation processing on the head tremor signal to obtain a first target tremor signal. A first biometric feature is extracted from the first target tremor signal and input into a pre-trained authentication model. The similarity between the first biometric feature and pre-stored user template features is calculated based on the pre-trained authentication model to obtain the authentication result. The authentication result is then used to determine whether the current user is a legitimate user. Thus, by collecting and analyzing the muscle physiological characteristics generated by the user's active head tremors, the problems of inconvenient authentication and susceptibility to eavesdropping associated with head-mounted devices are solved.
[0019] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description
[0020] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1This is a flowchart of an identity authentication method based on head tremor biometrics provided in an embodiment of this application; Figure 2 This is a schematic diagram of the authentication model structure of an identity authentication method based on head tremor biometrics according to an embodiment of this application; Figure 3 This is a flowchart illustrating an identity authentication method based on head tremor biometrics according to an embodiment of this application; Figure 4 This is a schematic diagram of an identity authentication device based on head tremor biometrics provided in an embodiment of this application; Figure 5 This is a schematic diagram of the structure of an electronic device provided according to an embodiment of this application. Detailed Implementation
[0021] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.
[0022] The following describes an identity authentication method, apparatus, and device based on head tremor biometrics according to embodiments of this application, with reference to the accompanying drawings. Addressing the problems of inconvenient and easily spied-on authentication using head-mounted devices mentioned in the background art, this application provides an identity authentication method based on head tremor biometrics. In this method, embodiments of this application utilize a wearable device to collect the head tremor signal of the current user, and perform denoising and event segmentation processing on the current user's head tremor signal to obtain a first target tremor signal. A first biometric feature is extracted from the first target tremor signal and input into a pre-trained authentication model. The similarity between the first biometric feature and pre-stored user template features is calculated based on the pre-trained authentication model to obtain an authentication result. The authentication result is used to determine whether the current user is a legitimate user. Thus, by collecting and analyzing the muscle physiological characteristics generated by the user's active head tremors, the problems of inconvenient and easily spied-on authentication using head-mounted devices are solved.
[0023] Specifically, Figure 1 This is a flowchart illustrating an identity authentication method based on head tremor biometrics provided in an embodiment of this application.
[0024] like Figure 1 As shown, this biometric authentication method based on head tremor includes the following steps: In step S101, the wearable device is used to collect the head tremor signal of the current user. Specifically, when user authentication is required, this embodiment prompts the user to actively tense their neck skeletal muscles, and the wearable device's built-in inertial measurement unit collects tremor data during this process. The inertial measurement unit simultaneously acquires multidimensional time-series data generated by a three-axis accelerometer and a three-axis gyroscope at a sampling rate of 500Hz, forming the raw signal, namely the head tremor signal.
[0025] Optionally, in some embodiments, the use of a wearable device to collect the head tremor signal of the current user includes: using the wearable device to collect the initial tremor signal generated by the current user based on the rhythm of a preset challenge signal; determining whether the initial tremor signal and the preset challenge signal meet a preset synchronization condition; if the initial tremor signal and the preset challenge signal meet the preset synchronization condition, then the initial tremor signal is used as the head tremor signal of the current user.
[0026] It is understood that the embodiments of this application design a challenge-response liveness detection. At the start of authentication, the system displays an audio that flashes at a random rhythm, requiring the user to follow the rhythm and correspondingly tense their neck muscles. The system uses the initial tremor signal collected by the wearable device to calculate the cross-correlation between the energy envelope of the collected tremor signal (initial tremor signal) and the time series of the audio rhythm (preset challenge signal). Only when the correlation coefficient is higher than the preset liveness detection threshold (preset synchronization condition) is the current operation considered to be a real, real-time, and legitimate user behavior, and the initial tremor signal is taken as the current user's head tremor signal.
[0027] In step S102, the head tremor signal of the current user is denoised and segmented to obtain the first target tremor signal. The first biometric feature is extracted from the first target tremor signal and input into the pre-trained authentication model. The similarity between the first biometric feature and the pre-stored user template feature is calculated based on the pre-trained authentication model to obtain the authentication result.
[0028] Among them, the pre-stored user template feature is a unique biometric ID card or standard fingerprint that is stored in the device after a legitimate user has successfully registered in the system, representing their identity.
[0029] Specifically, effective head tremor signal segments are separated and located from the noisy raw signal (head tremor signal) to obtain the first target tremor signal. Biometric information (first biometric feature) capable of stably characterizing the physiological properties of the user's neck muscles is extracted from the first target tremor signal. The first biometric feature includes a muscle contraction descriptor and a muscle endurance descriptor. The muscle contraction descriptor characterizes the force and velocity characteristics of the user's muscle contraction by calculating the Mel-frequency cepstral coefficients, spectral centroid, and spectral spread of the signal. The muscle endurance descriptor characterizes the stability and fatigue characteristics of the user's muscles by calculating the spectral flux, spectral entropy, and zero-crossing rate of the signal. The first biometric feature is input into a pre-trained authentication model (Siamese network module); where, for example... Figure 2 As shown, Figure 2 This is a schematic diagram of the authentication model structure of an identity authentication method based on head tremor biometrics according to an embodiment of this application. The pre-trained authentication model includes two sub-networks with identical structures and shared weights, used to reconstruct the input biometrics to mitigate the impact of inconsistent user behavior. Each sub-network contains 8 layers: the first layer is the input layer, which takes the extracted biometric data as input; the second, third, and fourth layers are all convolutional layers, and batch normalization and ReLU activation functions are added after the second, third, and fourth layers; the fifth layer is a one-dimensional max pooling layer; the sixth layer is a random deactivation layer; the seventh layer is a flattening layer, which flattens the data into one-dimensional data; the eighth layer is a fully connected layer, which maps the input into a fixed-length one-dimensional vector. In this embodiment, the input is the embedding vector output by the two sub-networks of the Siamese network, and the output is the Euclidean distance between the first biometric feature and the pre-stored user template feature to measure the similarity between the two. The smaller the Euclidean distance, the higher the similarity, and the authentication result is obtained.
[0030] Optionally, in some embodiments, the head tremor signal is denoised, including: performing power spectral density analysis on the head tremor signal and filtering out motion artifact noise in the head tremor signal based on the analysis results; and using maximum overlap discrete wavelet transform technology to filter out residual noise in the head tremor signal.
[0031] Understandably, the first noise reduction process uses power spectral density analysis to identify and filter out motion artifact noise generated by large user movements (such as head rotation and walking); the second noise reduction process uses maximum overlap discrete wavelet transform technology to filter out residual noise generated by small user movements (such as breathing and facial expression changes) that overlap with the frequency band of the head tremor signal.
[0032] Optionally, in some embodiments, the head tremor signal is subjected to event segmentation processing, including: calculating the aggregated signal variance of the multi-axis sensor signals in the wearable device; determining the start and end points of the head tremor signal based on the aggregated signal variance and a preset dual threshold method; and locating and segmenting the head tremor signal segments that meet preset valid conditions from the head tremor signal.
[0033] Among them, the preset dual-threshold method is a signal processing technique that robustly and accurately detects the start and end times of a specific event from a continuous signal stream. The preset validity conditions are a set of standards used to determine whether a signal segment is qualified after the start and end points of the event are detected. They generally include conditions such as minimum duration, maximum duration, minimum signal strength, minimum signal amplitude, and signal-to-noise ratio.
[0034] Understandably, tremor events are highlighted by calculating the aggregated signal variance of multi-axis sensor signals (multi-axis inertial measurement unit signals). When the aggregated signal variance is low, it indicates that the signal is stable during this period, and the user may be in a static or normal activity state. When the aggregated signal variance is high, it indicates that the signal fluctuates violently, and head tremor may have occurred. In some embodiments, this application sets a high threshold and a low threshold. When the aggregated signal variance value rises from low to high and first exceeds the high threshold, it is determined as the start point of a valid tremor event. When the aggregated signal variance value falls from high to low after reaching a peak and falls back below the low threshold, it is determined as the end point of the event. Based on the found start and end points, head tremor signal segments that meet the preset valid conditions are segmented from the original signal as target signals.
[0035] In step S103, the authentication result is used to determine whether the current user is a legitimate user.
[0036] Specifically, in this embodiment, the authentication result output by the pre-trained authentication model is received, and this value is compared with a preset decision threshold. When the authentication result meets the preset decision threshold, the current user is determined to be a legitimate user; when the authentication result does not meet the preset decision threshold, the current user is determined to be an illegitimate user, and security measures are triggered (such as: denying access, requiring a retry, locking the system).
[0037] For example, when performing facial recognition for access control, the authentication result is the matching degree between the face captured by the camera and the face registered in the database. The preset judgment threshold is the matching degree qualification line (90%) set by the access control system. When the authentication result is greater than or equal to the preset judgment threshold, it is determined to be a legitimate user; otherwise, the authentication fails. At this time, the matching degree between the face captured by the camera and the face registered in the database is calculated to be 95%. The authentication result (95%) is greater than the preset judgment threshold (90%), so it is determined to be a legitimate employee, and the door is opened for entry.
[0038] Optionally, in some embodiments, before calculating the similarity between the first biometric feature and the pre-stored user template features based on the pre-trained authentication model, the method further includes: acquiring head tremor signals from multiple target users; performing denoising and event segmentation processing on the head tremor signals from multiple target users to obtain a second target tremor signal; extracting a second biometric feature from the second target tremor signal; training a pre-set Siamese network training model by receiving positive and negative sample pairs of the second biometric feature, minimizing the distance between positive sample pairs and maximizing the distance between negative sample pairs, to obtain a trained Siamese network training model; wherein, a positive sample pair consists of two sample features from the same legitimate user, and a negative sample pair consists of one sample from a legitimate user and one sample from an illegitimate user; and optimizing the trained Siamese network training model using a pre-set contrastive loss function to obtain a pre-trained authentication model.
[0039] Understandably, the training process of the pre-trained authentication model first requires acquiring head tremor signals generated by multiple target users tensing their neck muscles, collecting tremor data through wearable devices; denoising and segmenting the tremor data to obtain a second target tremor signal; extracting the second biometric features contained in the second target tremor signal; and training the model using a pre-defined Siamese network training model based on the extracted second biometric features. The Siamese network (pre-defined Siamese network training model) receives positive and negative sample pairs for training. A positive sample pair consists of two legitimate user samples, and a negative sample pair consists of a legitimate user sample and an illegitimate user sample. The goal of the pre-defined Siamese network training model is to minimize the distance between positive sample pairs and maximize the distance between negative sample pairs. A pre-defined contrastive loss function is used to optimize the trained Siamese network training model to obtain the pre-trained authentication model.
[0040] It should be noted that, in order to achieve passive continuous authentication, after the user completes the initial identity authentication, the pre-trained authentication model switches to a low-power monitoring mode. In this mode, the embodiments of this application collect the passive head tremor signal naturally generated by the user in a static or slightly active state once per minute, and process it using a parameter set optimized for weak signals. The user identity discrimination module then compares the extracted features with the user template. If multiple comparisons fail, the embodiments of this application can determine that the user has been replaced and automatically lock the device or require active authentication to be performed again.
[0041] To facilitate those skilled in the art to further understand the identity authentication method based on head tremor biometrics in the embodiments of this application, the following is combined with... Figure 3 The embodiments shown will be described in detail.
[0042] Specifically, such as Figure 3 As shown, Figure 3This is a flowchart illustrating an identity authentication method based on head tremor biometrics according to an embodiment of this application. This application provides an identity authentication system based on head tremor biometrics, comprising the following modules: a data acquisition module for collecting head tremor data; a data processing module for denoising and segmenting the tremor data to obtain a target signal; a feature extraction module for extracting at least one biometric feature from the target signal; and a user identity determination module for determining whether a user is a legitimate user based on the extracted biometric features and using an authentication model.
[0043] Furthermore, in this embodiment of the application, 30 subjects aged 18-27 years participated in the experiment, including 8 women and 22 men. During the experiment, the subjects wore commercial virtual reality headsets and actively generated head tremors as prompted. In this embodiment of the application, inertial measurement unit data was collected at a sampling rate of 500Hz. During data analysis, 10 samples were used to register and fine-tune the authentication model for each user. The experimental results show that the present invention can effectively identify legitimate users with a balance accuracy of 97.22%.
[0044] Therefore, when a user wears a virtual reality device, the user first tenses their neck muscles, and the head-mounted device collects tremor data. Next, the tremor data is denoised and segmented to obtain the target signal. Then, the biometric features contained in the target signal are extracted. Finally, based on the extracted biometric features, an authentication model is used to determine whether the user is legitimate. This invention collects tremor data when a user wears a virtual reality device using an inertial measurement unit. It leverages the differences in the contraction and endurance characteristics of different users' neck muscles, while maintaining consistency within the same user, to perform feature extraction and modeling to identify the user's identity. Therefore, it has advantages such as high security, strong practicality, and ease of deployment.
[0045] According to the head tremor biometric authentication method proposed in this application, this application utilizes a wearable device to collect the head tremor signal of the current user; performs denoising and event segmentation processing on the current user's head tremor signal to obtain a first target tremor signal; extracts a first biometric feature from the first target tremor signal; and inputs the first biometric feature into a pre-trained authentication model. The similarity between the first biometric feature and pre-stored user template features is calculated based on the pre-trained authentication model to obtain the authentication result; and the authentication result determines whether the current user is a legitimate user. Therefore, by collecting and analyzing the muscle physiological characteristics generated by the user's active head tremors, the problems of inconvenient authentication and susceptibility to eavesdropping associated with head-mounted devices are solved.
[0046] Next, with reference to the accompanying drawings, an identity authentication device based on head tremor biometrics proposed according to an embodiment of this application is described.
[0047] Figure 4This is a block diagram of an identity authentication device based on head tremor biometrics according to an embodiment of this application.
[0048] like Figure 4 As shown, the identity authentication device 10 based on head tremor biometrics includes: a data acquisition module 100, a determination module 200, and a judgment module 300.
[0049] The acquisition module 100 is used to acquire the head tremor signal of the current user using a wearable device; The determination module 200 is used to perform noise reduction and event segmentation processing on the head tremor signal of the current user to obtain the first target tremor signal, extract the first biometric feature from the first target tremor signal, and input the first biometric feature into the pre-trained authentication model to calculate the similarity between the first biometric feature and the pre-stored user template feature based on the pre-trained authentication model to obtain the authentication result. The judgment module 300 is used to determine whether the current user is a legitimate user based on the authentication result.
[0050] Optionally, the determination module 200 is specifically used for: performing power spectral density analysis on the head tremor signal and filtering out motion artifact noise in the head tremor signal based on the analysis results; and using maximum overlap discrete wavelet transform technology to filter out residual noise in the head tremor signal.
[0051] Optionally, the determining module 200 is specifically used to: calculate the aggregated signal variance of the multi-axis sensor signals in the wearable device; determine the start and end points of the head tremor signal based on the aggregated signal variance and a preset dual threshold method, and locate and segment the head tremor signal segments that meet the preset valid conditions from the head tremor signal.
[0052] Optionally, the acquisition module 100 is specifically used to: acquire the initial tremor signal generated by the current user based on the rhythm of the preset challenge signal using a wearable device; determine whether the initial tremor signal and the preset challenge signal meet the preset synchronization condition; if the initial tremor signal and the preset challenge signal meet the preset synchronization condition, then the initial tremor signal is used as the head tremor signal of the current user.
[0053] Optionally, before calculating the similarity between the first biometric feature and the pre-stored user template features based on the pre-trained authentication model, the determination module 200 is further configured to: acquire head tremor signals from multiple target users; perform denoising and event segmentation processing on the head tremor signals from multiple target users to obtain a second target tremor signal, and extract a second biometric feature from the second target tremor signal; train a pre-set Siamese network training model by receiving positive and negative sample pairs of the second biometric feature, minimizing the distance between positive sample pairs and maximizing the distance between negative sample pairs, to obtain a trained Siamese network training model; wherein, a positive sample pair consists of two sample features from the same legitimate user, and a negative sample pair consists of one sample from a legitimate user and one sample from an illegitimate user; optimize the trained Siamese network training model using a pre-set contrastive loss function to obtain a pre-trained authentication model.
[0054] It should be noted that the foregoing explanation of the embodiment of the biometric authentication method based on head tremor also applies to the biometric authentication device based on head tremor in this embodiment, and will not be repeated here.
[0055] According to the head tremor biometric authentication device proposed in this application, this embodiment utilizes a wearable device to collect the head tremor signal of the current user; performs denoising and event segmentation processing on the current user's head tremor signal to obtain a first target tremor signal; extracts a first biometric feature from the first target tremor signal; and inputs the first biometric feature into a pre-trained authentication model. The similarity between the first biometric feature and pre-stored user template features is calculated based on the pre-trained authentication model to obtain the authentication result; and the authentication result determines whether the current user is a legitimate user. Thus, by collecting and analyzing the muscle physiological characteristics generated by the user's active head tremors, the problems of inconvenient authentication and susceptibility to eavesdropping associated with head-mounted devices are solved.
[0056] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include: The memory 501, the processor 502, and the computer program stored on the memory 501 and capable of running on the processor 502.
[0057] When the processor 502 executes the program, it implements the identity authentication method based on head tremor biometrics provided in the above embodiments.
[0058] Furthermore, electronic devices also include: Communication interface 503 is used for communication between memory 501 and processor 502.
[0059] The memory 501 is used to store computer programs that can run on the processor 502.
[0060] Memory 501 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.
[0061] If the memory 501, processor 502, and communication interface 503 are implemented independently, then the communication interface 503, memory 501, and processor 502 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 5 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.
[0062] Optionally, in a specific implementation, if the memory 501, processor 502, and communication interface 503 are integrated on a single chip, then the memory 501, processor 502, and communication interface 503 can communicate with each other through an internal interface.
[0063] Processor 502 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.
[0064] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the above-described biometric authentication method based on head tremor.
[0065] This application also provides a computer program product that stores a computer program that, when executed by a processor, implements the above-described biometric authentication method based on head tremor.
[0066] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.
[0067] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.
[0068] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.
[0069] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.
[0070] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.
Claims
1. A biometric authentication method based on head tremor, characterized in that, Includes the following steps: The device collects head tremor signals from the current user using wearable devices. The head tremor signal of the current user is denoised and segmented to obtain a first target tremor signal. A first biometric feature is extracted from the first target tremor signal and input into a pre-trained authentication model. The similarity between the first biometric feature and the pre-stored user template feature is calculated based on the pre-trained authentication model to obtain the authentication result. The authentication result is used to determine whether the current user is a legitimate user.
2. The method according to claim 1, characterized in that, The head tremor signal is denoised, including: Power spectral density analysis was performed on the head tremor signal, and motion artifact noise in the head tremor signal was filtered out based on the analysis results; Residual noise in the head tremor signal is filtered out using the maximum overlap discrete wavelet transform technique.
3. The method according to claim 1, characterized in that, The head tremor signal is subjected to event segmentation processing, including: Calculate the aggregated signal variance of the multi-axis sensor signals in the wearable device; The start and end points of the head tremor signal are determined based on the aggregated signal variance and a preset dual threshold method, and segments of the head tremor signal that meet preset valid conditions are located and segmented from the head tremor signal.
4. The method according to claim 1, characterized in that, The method of collecting the current user's head tremor signal using a wearable device includes: The wearable device is used to collect the initial tremor signal generated by the current user based on the rhythm of a preset challenge signal; Determine whether the initial tremor signal and the preset challenge signal meet the preset synchronization conditions; If the initial tremor signal and the preset challenge signal meet the preset synchronization condition, then the initial tremor signal is used as the head tremor signal of the current user.
5. The method according to claim 1, characterized in that, Before calculating the similarity between the first biometric feature and the pre-stored user template features based on the pre-trained authentication model, the method further includes: Acquire head tremor signals from multiple target users; The head tremor signals of the multiple target users are denoised and segmented to obtain a second target tremor signal, and a second biometric feature is extracted from the second target tremor signal. The Siamese network training model is trained by receiving positive and negative sample pairs of the second biometric feature, thereby minimizing the distance between the positive sample pairs and maximizing the distance between the negative sample pairs, thus obtaining the trained Siamese network training model; wherein, the positive sample pair consists of two sample features from the same legitimate user, and the negative sample pair consists of one sample from a legitimate user and one sample from an illegitimate user. The trained Siamese network training model is optimized using a preset contrastive loss function to obtain the pre-trained authentication model.
6. An identity authentication device based on head tremor biometrics, characterized in that, include: The acquisition module is used to acquire the current user's head tremor signals using wearable devices; The determination module is used to perform noise reduction and event segmentation processing on the head tremor signal of the current user to obtain a first target tremor signal, extract a first biometric feature from the first target tremor signal, and input the first biometric feature into a pre-trained authentication model to calculate the similarity between the first biometric feature and the pre-stored user template feature based on the pre-trained authentication model to obtain the authentication result. The judgment module is used to determine whether the current user is a legitimate user based on the authentication result.
7. The apparatus according to claim 6, characterized in that, The determining module is specifically used for: Power spectral density analysis was performed on the head tremor signal, and motion artifact noise in the head tremor signal was filtered out based on the analysis results; Residual noise in the head tremor signal is filtered out using the maximum overlap discrete wavelet transform technique.
8. An electronic device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, the processor executing the program to implement the head tremor-based biometric authentication method as described in any one of claims 1-5.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the biometric authentication method based on head tremor as described in any one of claims 1-5.
10. A computer program product, said computer program product storing a computer program, characterized in that, When executed by the processor, the program implements the biometric authentication method based on head tremor as described in any one of claims 1-5.