Living body detection method based on palm pulse wave and related device

Through the live detection method based on palm pulse waves, the pulse wave signal is extracted using graph neural network and attention mechanism, and combined with time domain and frequency domain analysis, the identification accuracy and safety of existing biometric verification methods in high-risk scenarios are solved. The problem of insufficient sexuality has achieved higher identification accuracy and security, providing reliable technical guarantees for related applications.

CN118975782APending Publication Date: 2024-11-19XIAN EASY PALM TECH CO LTD
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411034112.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-30
Publication Date
2024-11-19

AI Technical Summary

Technical Problem

Existing biometric verification methods are prone to counterfeiting and impersonation in high-risk scenarios, resulting in insufficient identification accuracy and safety, and need to improve the accuracy and safety performance of live detection.

Method used

Using the live detection method based on palm pulse waves, RGB and near-infrared videos are obtained through dual-channel video acquisition, pulse wave signals are extracted using graph neural network and attention mechanism, and combined with time domain and frequency domain analysis, a real palm pulse wave database is established for Discrimination and improve identification accuracy.

Benefits of technology

It greatly improves the accuracy of the identification of the palm under test, enhances the system's attack resistance and security, provides higher security guarantees and user experience, and is suitable for application scenarios such as palm swiping payment and smart access control.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118975782A_ABST
    Figure CN118975782A_ABST
Patent Text Reader

Abstract

The invention relates to a biomedical signal processing technology, and discloses a palm pulse wave-based living body detection method and a related device, and the method comprises the steps: firstly obtaining an original RGB video and a near-infrared palm video through a two-channel video collection mode, and carrying out the palm detection to guarantee the effective collection; then, pulse wave signals are extracted by means of deep learning and multi-modal fusion, and the pulse wave signals are preprocessed and fused; then, an independent component analysis algorithm is used for enhancement, and key features are highlighted; and the dynamic characteristics of pulse waves, such as the change of frequency spectrum and amplitude along with time, are researched in a mode of combining time domain analysis and frequency domain analysis. And finally, the detection result is compared with the pre-stored real user data, and living body discrimination is performed by using the trained neural network model and the pre-training model, so that the accuracy of identifying the detected palm is greatly improved, and reliable technical guarantee is provided for related applications.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to biomedical signal processing technology, in particular to a living body detection method based on palm pulse waves and related devices. Background Art

[0002] With the rapid development of the Internet and deep learning technology, people have begun to explore a new type of biometric verification method. For example, palm payment, access control systems and other applications widely use biometric recognition technology to enhance security and convenience. However, these emerging technologies also bring some security risks.

[0003] Although traditional biometric authentication methods such as fingerprint recognition or facial recognition have been widely used, they are prone to duplication or imitation. Especially in high-risk scenarios, such as financial payment and security access control systems, traditional biometric recognition technology may face the challenge of counterfeiting and impersonation, which seriously affects system security and user trust.

[0004] However, there are still drawbacks in the recognition accuracy of existing biometric verification methods. There are often recognition errors or the impersonated living bodies are easily recognized, which will bring great security risks. How to improve the accuracy of biometric verification methods and the detection of security performance still needs further exploration. Summary of the invention

[0005] In order to overcome the defects of the above-mentioned prior art, the purpose of the present invention is to provide a liveness detection method based on palm pulse wave and related devices to solve the technical problem of how to improve the accuracy of liveness detection in the biometric verification method in the prior art.

[0006] The present invention is achieved through the following technical solutions:

[0007] In a first aspect, the present invention provides a method for detecting a living body based on a palm pulse wave, comprising:

[0008] Obtain a video sequence of the detected palm;

[0009] Extracting pulse wave signals from video sequences;

[0010] analyzing dynamic pulse wave characteristics of the extracted pulse wave signal;

[0011] Liveness detection is performed based on dynamic pulse wave features to determine the authenticity of the detected palm and complete liveness detection based on palm pulse waves.

[0012] Preferably, when acquiring the video sequence of the detected palm, a dual-channel video acquisition method is adopted to respectively acquire the original RGB video and the near-infrared palm video of the detected palm.

[0013] Preferably, the pulse wave signal is extracted from the video sequence using graph neural network means and attention mechanism means, and the pulse wave signals extracted by the graph neural network means and the attention mechanism means are weighted fused in turn.

[0014] Furthermore, when extracting pulse wave signals from video sequences using graph neural networks, the original video frames of the video sequence are divided into short sequences and long sequences for extracting different feature information, where the short sequence is used to extract local feature information, represented by L s , the long sequence is used to extract the global long-term periodic features, denoted as L l .

[0015] Furthermore, the attention mechanism dynamically adjusts the importance weights of different features extracted by the graph neural network, and performs weighted fusion on the adjusted features.

[0016] Preferably, in analyzing the dynamic pulse wave characteristics of the extracted pulse wave signal, the pulse wave signal related to human physiology is restored through signal processing and signal enhancement algorithms, and the periodicity of the real pulse wave signal and the forged prosthetic pulse wave and the pattern of amplitude change over time in the spectrum diagram are analyzed by combining time domain and frequency domain analysis.

[0017] Preferably, the specific process of liveness detection is as follows:

[0018] Establish a real palm pulse wave database to store the amplitude, period, and dynamic characteristic data of the real palm pulse wave;

[0019] Compare the detected palm pulse wave characteristics with the real palm pulse wave data;

[0020] The neural network model is used to perform discrimination, obtain discrimination results, and save the results of network training as a pre-trained model.

[0021] In a second aspect, the present invention further provides a liveness detection system based on palm pulse waves, comprising a video sequence acquisition module for acquiring a video sequence of a detected palm;

[0022] A signal extraction module, used for extracting a pulse wave signal from a video sequence;

[0023] A feature analysis module, used for analyzing the dynamic pulse wave features of the extracted pulse wave signal;

[0024] The liveness detection module is used to perform liveness detection based on dynamic pulse wave characteristics, determine the authenticity of the detected palm, and complete liveness detection based on palm pulse waves.

[0025] In a third aspect, the present invention also provides a mobile terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the liveness detection method based on palm pulse wave as described above when executing the computer program.

[0026] In a fourth aspect, the present invention further provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the steps of the liveness detection method based on palm pulse wave as described above.

[0027] Compared with the prior art, the present invention has the following beneficial technical effects:

[0028] The present invention provides a method for liveness detection based on palm pulse wave. First, the original RGB video and near-infrared palm video are acquired through a dual-channel video acquisition method, and palm detection is performed to ensure effective acquisition. Then, the pulse wave signal is extracted by deep learning and multimodal fusion means, and preprocessed and fused. Then, the independent component analysis algorithm is used for enhancement to highlight the key features. Then, the dynamic characteristics of the pulse wave, such as the change of spectrum and amplitude over time, are studied by combining time domain and frequency domain analysis. Finally, the detection results are compared with the real user data stored in advance, and the trained neural network model and the pre-trained model are used for liveness discrimination, which greatly improves the accuracy of identifying the palm under test and provides reliable technical guarantee for related applications.

[0029] Furthermore, liveness detection through palm pulse waves can reflect the unique physiological activity characteristics of an individual, which is not easy to be forged or imitated, greatly improving the system's anti-attack ability and security.

[0030] Furthermore, the palm pulse wave contains the dynamic physiological characteristics of an individual, such as changes in vascular networks and blood flow. Compared with static biometric features, the palm pulse wave is not easily forged or imitated, and has higher security and reliability. In addition, the palm area is relatively stable, easy to obtain, and does not require specific user posture or lighting conditions, making the technology widely applicable and operable in practical applications.

[0031] Furthermore, liveness detection technology based on palm pulse waves can not only effectively cope with the security challenges faced by current biometric technology, but also provide a higher level of security and user experience for application scenarios such as palm payment and smart access control. It is one of the important technological innovation directions in the field of biometrics. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 is a flow chart of a living body detection method based on palm pulse wave in the present invention;

[0033] Figure 2 It is a schematic diagram of the structure of the living body detection system based on the palm pulse wave in the present invention;

[0034] Figure 3 It is a schematic diagram of liveness detection based on pulse wave in the present invention;

[0035] Figure 4 Schematic diagram of pulse wave signals corresponding to real samples and forged samples in the present invention;

[0036] In the figure: 1- video sequence acquisition module; 2- signal extraction module; 3- feature analysis module; 4- living body detection module. DETAILED DESCRIPTION

[0037] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0038] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0039] The present invention is further described in detail below in conjunction with the accompanying drawings:

[0040] The purpose of the present invention is to provide a liveness detection method based on palm pulse wave and related devices to solve the technical problem of how to improve the accuracy of liveness detection in biometric verification methods in the prior art.

[0041] See also Figure 1 In one embodiment of the present invention, a method for detecting a living body based on a palm pulse wave is provided, comprising:

[0042] Step 1, obtaining a video sequence of the detected palm;

[0043] Specifically, high-precision video acquisition equipment is selected, and appropriate frame rates and resolutions are set to ensure that clear and continuous palm video sequences are obtained. A dual-channel video acquisition method is used during the acquisition process, one channel obtains the original RGB video, and the other channel obtains the near-infrared palm video. During the acquisition process, the palm detection technology is used to monitor the position and posture of the palm in the acquisition area in real time to ensure the effective capture and extraction of video data in the palm area. For acquisition results that do not meet the requirements, prompt re-acquisition in a timely manner. In order to improve the accuracy and stability of the acquisition, appropriate lighting conditions and background environment can be set to reduce the impact of external interference factors on the acquisition results. During the video acquisition process, ensure that the palm area is fully exposed and clearly visible to ensure the accuracy and stability of subsequent signal processing.

[0044] Step 2, extracting the pulse wave signal from the video sequence;

[0045] Specifically, graph neural networks (GNNs) can process irregular graph structure data, flexibly capture complex relationships and dependencies between nodes, and are suitable for more complex data representation. GNNs can process irregular and dynamically changing graph structure data, capture complex relationships and dependencies between nodes, and better capture spatiotemporal features through a unique message passing mechanism. They are particularly suitable for capturing dynamic changes in pulse wave signals. Through multi-level feature aggregation and updating, they can learn richer and higher-dimensional feature representations, improving recognition accuracy and robustness.

[0046] Among them, the specific process of extracting the pulse wave signal is as follows:

[0047] First, construct a graph structure: model the palm area as a graph structure G. The graph structure G consists of a node set V and an edge set E, G = (V, E); the node v∈V of the palm graph represents a pixel or superpixel in the palm image, and the edge (u, v)∈E represents the spatial relationship between node u and node v. Use a graph neural network (GNN) to learn the feature representation of each node. The initial feature representation of each node v is the pixel value or superpixel value in the palm image. It is expressed as the following formula:

[0048]

[0049] in is the initial feature representation of node v, x v is the input feature of node v. The feature is then updated through the message passing mechanism:

[0050]

[0051] in is the feature representation of node v at the kth layer, N(v) is the set of adjacent nodes of node v, Message and Update are message passing and update functions respectively. Considering that the pulse wave has periodicity on the time axis, but the local waveform contains rich physiological features, the original video frame is first divided into two sequences: short sequence and long sequence for extracting different feature information when using graph neural network for feature extraction. The short sequence is used to extract local feature information, represented by L s , the long sequence is used to extract the global long-term periodic features, denoted as L l By extracting features from these two different sequence blocks using graph neural networks, the key information of the pulse wave at the local position is retained while the periodic characteristics of the entire signal on the time axis are restored.

[0052] The two features are then fused. During the fusion process, the attention mechanism is introduced to enhance the effect of feature fusion. The attention mechanism can dynamically adjust the importance weights of different features, thereby improving the accuracy of feature representation. The weight calculation is achieved through the following formula:

[0053]

[0054] Among them, score(h i ,q) is the characteristic h i The similarity score between the query vector q is calculated by dot product. Finally, the features are weighted and fused, and the final feature representation is:

[0055]

[0056] Through the above steps, a complete pulse wave signal representation can be obtained. Then, the loss function is established by the pulse wave extracted from the original video frame and the real pulse wave signal. Since the pulse wave signal reflects the physiological characteristics of the human body and is related to the heart rate, the traditional mean square error (MSE) is used as the loss function, which can well reflect the prediction performance of the model. It is obtained by calculating the difference between the predicted value and the real value. The specific calculation method is as follows:

[0057]

[0058] where y i is the true value, is the predicted value, and P is the number of samples. The error can effectively determine the difference between the predicted value and the true value. To perform liveness detection, it is necessary to distinguish between real samples and fake prostheses. This can be achieved by setting an error threshold. By adjusting the threshold, the difference between the pulse waves of real and fake individuals can be distinguished. Generally, the palm video of the real sample has rich feature information, so the error with the real sample is also small, while the pulse wave information in the fake sample video is generally very weak, so the error with the real sample is relatively large. This mechanism can achieve liveness detection through pulse waves.

[0059] Step 3, analyzing the dynamic pulse wave characteristics of the extracted pulse wave signal;

[0060] When analyzing dynamic features, fast Fourier transform and inverse transform are used to process the pulse wave signal to obtain feature expressions in the frequency domain and time domain, and analyze its spectral features. Before Fourier transform, the extracted original signal is first detrended and smoothed to reduce the interference of noise and false peaks, so as to obtain physiological signals and make the heartbeat peak more obvious.

[0061] The signal enhancement algorithm, adaptive filtering algorithm and fast Fourier transform are used to restore the pulse wave signal related to human physiology and highlight its key features. The signal range is limited to 0.75-2.5 Hz by filtering. The signal within this range retains the physiological characteristics related to human heart rate and can highlight the key features. The implementation method of Fourier transform is:

[0062]

[0063] Among them, F(f) reflects the amplitude and phase of frequency f in the signal, x[b] represents the nth sample value of the time domain signal, f represents the current frequency component, n represents the index of the current sample, and N represents the total number of samples of the signal. By analyzing the frequency component, we can find the value of the corresponding frequency reflecting the instantaneous heartbeat.

[0064] In the time domain analysis, the signal is restored to the time domain dimension by performing an inverse Fourier transform on the signal, and then the peaks and peak intervals in the periodic signal are detected by the peak detection algorithm to accurately restore the periodic characteristics of the pulse wave. The peak detection algorithm is implemented using the following formula:

[0065] P = {x i |x i-1 <x i <x i+1 ∧x i >T}

[0066] Where P is the set of detected peaks, x i is the current signal point, x i-1 and xi+1 are the adjacent preceding and following signal points, and T is the threshold, which is used to exclude low-amplitude signal points that do not meet the peak standard.

[0067] Setting an appropriate threshold T is the key, which is obtained by dynamically calculating the mean and standard deviation:

[0068] T=μ+k·σ

[0069] Where μ is the mean value of the signal, σ is the standard deviation, and k is the adjustment coefficient used to control the strictness of the threshold. The appropriate k value is selected according to the signal characteristics.

[0070] After identifying the peaks, the characteristics of these peaks are analyzed, such as the time interval between peaks (i.e., heartbeat cycle), the height and width of the peaks. The pulse wave signal of the real sample has obvious periodic characteristics and a more obvious peak, while the forged sample individuals do not have this characteristic information. In this way, the difference between the real signal and the forged sample signal can be effectively identified, thereby achieving the purpose of liveness detection.

[0071] By combining time domain and frequency domain analysis, the periodic differences between the real pulse wave signal and the forged prosthetic pulse wave are compared in detail. At the same time, the change of amplitude over time in the spectrum diagram is observed to extract characteristic parameters that can effectively distinguish between real and forged pulse waves. Figure 4 As shown in: It is the result display of pulse wave extracted from real palm and forged palm, among which Figure (a), Figure (b) and Figure (c) are pulse waves extracted from real palm, Figure (d), Figure (e) and Figure (f) are from forged palm. It can be seen that the pulse wave of real palm has obvious periodicity and stable amplitude, while the pulse wave signal extracted from forged palm is very weak and does not have periodicity. Therefore, it can be clearly seen that the prosthesis does not contain dynamic physiological characteristics of human body.

[0072] Step 4: Perform liveness detection based on the dynamic pulse wave feature to determine the authenticity of the detected palm and complete the liveness detection based on the palm pulse wave.

[0073] Specifically, the specific process of liveness detection is as follows:

[0074] Establish a real palm pulse wave database to store the amplitude, period, and dynamic characteristic data of the real palm pulse wave;

[0075] Compare the detected palm pulse wave characteristics with the real palm pulse wave data;

[0076] In the comparison process, the differences in multiple characteristics of the time series, such as amplitude, period and dynamics, are comprehensively considered.

[0077] The neural network model is used to perform discrimination, obtain discrimination results, and save the results of network training as a pre-trained model.

[0078] Specifically, the trained neural network model is used for discrimination, and the network training result is saved as a pre-trained model. In actual detection, the network training parameters and pre-trained model can be directly called to perform fast and accurate liveness detection on the device.

[0079] According to the test results, give a clear judgment conclusion, such as "true" or "false", such as Figure 3 After being judged by the discriminator, the real or fake palm individuals can be identified, and the detection results can be recorded and stored for subsequent query and analysis.

[0080] In summary, this embodiment provides a method for liveness detection based on palm pulse waves. First, the original RGB video and near-infrared palm video are acquired through a dual-channel video acquisition method, and palm detection is performed to ensure effective acquisition. Then, the pulse wave signal is extracted by deep learning and multimodal fusion, and preprocessed and fused. Then, the independent component analysis algorithm is used for enhancement to highlight the key features. Then, the dynamic characteristics of the pulse wave, such as the change of spectrum and amplitude over time, are studied by combining time domain and frequency domain analysis. Finally, the detection results are compared with the real user data stored in advance, and the trained neural network model and pre-trained model are used for liveness discrimination, which greatly improves the accuracy of identifying the palm being tested and provides reliable technical support for related applications.

[0081] In this embodiment, no user-specific wearable equipment or complicated operation steps are required. Contactless biometric verification can be achieved by simply collecting video through an ordinary camera. The palm area is relatively stable and contains rich biometric information. It is suitable for a variety of application scenarios, including financial payment, access control systems, health monitoring, etc., providing new technical support for the development of an intelligent society. Palm video collection does not involve facial or other personal identity information, which better protects the user's privacy and reduces the risk of personal information leakage caused by biometric technology.

[0082] The liveness detection technology based on palm pulse waves in this embodiment can not only effectively cope with the security challenges faced by current biometrics technology, but also provide a higher level of security and user experience, promoting the advancement and popularization of biometrics technology. With the continuous evolution of technology and the expansion of application scenarios, this technology is expected to play an important role in various fields and promote the further development of an intelligent society.

[0083] Example 2

[0084] according to Figure 2 As shown, this embodiment provides a living body detection system based on palm pulse wave, including:

[0085] The video sequence acquisition module 1 is used to acquire the video sequence of the detected palm;

[0086] A signal extraction module 2 is used to extract a pulse wave signal from a video sequence;

[0087] A feature analysis module 3, used for analyzing the dynamic pulse wave features of the extracted pulse wave signal;

[0088] The liveness detection module 4 is used to perform liveness detection based on dynamic pulse wave characteristics, determine the authenticity of the detected palm, and complete the liveness detection work based on the palm pulse wave.

[0089] Example 3

[0090] The present invention also provides a mobile terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor, such as a liveness detection program based on palm pulse waves.

[0091] When the processor executes the computer program, the steps of the above-mentioned method for detecting a living body based on a palm pulse wave are implemented, for example:

[0092] Obtain a video sequence of the detected palm;

[0093] Extracting pulse wave signals from video sequences;

[0094] analyzing dynamic pulse wave characteristics of the extracted pulse wave signal;

[0095] Liveness detection is performed based on dynamic pulse wave features to determine the authenticity of the detected palm and complete liveness detection based on palm pulse waves.

[0096] Alternatively, when the processor executes the computer program, the functions of each module in the above system are realized, for example:

[0097] The video sequence acquisition module 1 is used to acquire the video sequence of the detected palm;

[0098] A signal extraction module 2 is used to extract a pulse wave signal from a video sequence;

[0099] A feature analysis module 3, used for analyzing the dynamic pulse wave features of the extracted pulse wave signal;

[0100] The liveness detection module 4 is used to perform liveness detection based on dynamic pulse wave characteristics, determine the authenticity of the detected palm, and complete the liveness detection work based on the palm pulse wave.

[0101] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to implement the present invention. The one or more modules / units may be a series of computer program instruction segments capable of implementing specific functions, which are used to describe the execution process of the computer program in the mobile terminal.

[0102] For example, the computer program may be divided into a video sequence acquisition module 1, a signal extraction module 2, a feature analysis module 3 and a living body detection module 4;

[0103] The specific functions of each module are as follows:

[0104] The video sequence acquisition module 1 is used to acquire the video sequence of the detected palm;

[0105] A signal extraction module 2 is used to extract a pulse wave signal from a video sequence;

[0106] A feature analysis module 3, used for analyzing the dynamic pulse wave features of the extracted pulse wave signal;

[0107] The liveness detection module 4 is used to perform liveness detection based on dynamic pulse wave characteristics, determine the authenticity of the detected palm, and complete the liveness detection work based on the palm pulse wave.

[0108] The mobile terminal may be a computing device such as a desktop computer, a notebook, a palm computer, a cloud server, etc. The mobile terminal may include, but is not limited to, a processor and a memory.

[0109] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the mobile terminal, and uses various interfaces and lines to connect various parts of the entire mobile terminal.

[0110] The memory may be used to store the computer programs and / or modules, and the processor implements various functions of the mobile terminal by running or executing the computer programs and / or modules stored in the memory and calling the data stored in the memory.

[0111] The memory may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (SmartMediaCard, SMC), a secure digital (SecureDigital, SD) card, a flash card (FlashCard), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0112] Example 4

[0113] The present invention also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of the living body detection method based on palm pulse wave are implemented.

[0114] If the module / unit integrated in the mobile terminal is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium.

[0115] Based on such understanding, the present invention implements all or part of the processes in the above method, and can also be completed by instructing related hardware through a computer program, and the computer program can be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of the above-mentioned liveness detection method based on palm pulse wave can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc.

[0116] The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium, etc.

[0117] It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media does not include electrical carrier signals and telecommunication signals.

[0118] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for detecting a living body based on palm pulse wave, characterized in that: include: Obtain a video sequence of the detected palm; Extracting pulse wave signals from video sequences; analyzing dynamic pulse wave characteristics of the extracted pulse wave signal; Liveness detection is performed based on dynamic pulse wave features to determine the authenticity of the detected palm and complete liveness detection based on palm pulse waves.

2. A method for detecting a living body based on palm pulse waves according to claim 1, characterized in that: When acquiring the video sequence of the detected palm, a dual-channel video acquisition method is used to respectively acquire the original RGB video and the near-infrared palm video of the detected palm.

3. The method for detecting a living body based on palm pulse wave according to claim 1, characterized in that: Graph neural network and attention mechanism are used to extract pulse wave signals from video sequences, and the pulse wave signals extracted by graph neural network and attention mechanism are weighted fused in turn.

4. A method for detecting living body based on palm pulse wave according to claim 3, characterized in that: When extracting pulse wave signals from video sequences using graph neural networks, the original video frames of the video sequence are divided into short sequences and long sequences for extracting different feature information, where the short sequence is used to extract local feature information, represented by L s , the long sequence is used to extract the global long-term periodic features, denoted as L l .

5. The method for detecting living body based on palm pulse wave according to claim 3, characterized in that: The attention mechanism dynamically adjusts the importance weights of different features extracted by the graph neural network method, and performs weighted fusion on the adjusted features.

6. The method for detecting living body based on palm pulse wave according to claim 1, characterized in that: In analyzing the dynamic pulse wave characteristics of the extracted pulse wave signal, the pulse wave signal related to human physiology is restored through signal processing and signal enhancement algorithms, and the periodicity of the real pulse wave signal and the forged prosthetic pulse wave as well as the amplitude change pattern over time in the spectrum diagram are analyzed by combining time domain and frequency domain analysis.

7. The method for detecting a living body based on palm pulse wave according to claim 1, characterized in that: The specific process of liveness detection is as follows: Establish a real palm pulse wave database to store the amplitude, period, and dynamic characteristic data of the real palm pulse wave; Compare the detected palm pulse wave characteristics with the real palm pulse wave data; The neural network model is used to perform discrimination, obtain discrimination results, and save the results of network training as a pre-trained model.

8. A liveness detection system based on palm pulse wave, characterized in that: include A video sequence acquisition module (1), used for acquiring a video sequence of a detected palm; A signal extraction module (2), used for extracting a pulse wave signal from a video sequence; A feature analysis module (3), used for analyzing the dynamic pulse wave features of the extracted pulse wave signal; The liveness detection module (4) is used to perform liveness detection based on dynamic pulse wave characteristics, determine the authenticity of the detected palm, and complete the liveness detection work based on the palm pulse wave.

9. A mobile terminal comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the living body detection method based on palm pulse wave as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the liveness detection method based on palm pulse wave as described in any one of claims 1 to 7 are implemented.

Citation Information

Cited By

  • Multi-modal identity authentication method and system based on PPG characteristic spectrum construction

    CN121350560A