A living body fingerprint detection method based on fusion of multispectral and micro-arterial pulse characteristics

Abstract

The present application relates to the technical field of live fingerprint detection, and specifically discloses a live fingerprint detection method based on multispectral and micro-arterial pulse feature fusion, which comprises signal synchronization generation, time domain alignment processing, feature matrix generation, weight dynamic adjustment, cross-validation analysis, abnormal area sampling, live model verification and live authentication output; the present application synchronously collects multispectral reflection intensity and skin displacement fluctuation sequence in the fingerprint area, dynamically adjusts the weight of spectral absorption and pulsation trajectory features, cross-verify outputs local abnormal areas and obtains supplementary feature data, uses a preset live judgment model to perform physiological coupling rule matching, generates a verification result, and activates a forgery attack alarm protocol when the result of three consecutive times does not conform to the live rule; the method effectively improves the accuracy of live detection and enhances the environmental adaptability of the detection algorithm by fusing multispectral and micro-arterial pulse features and combining with an environment adaptive weight mechanism.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of live fingerprint detection, in particular to a live fingerprint detection method based on fusion of multispectral and micro-arterial pulse characteristics. BACKGROUND

[0002] Fingerprint recognition, as a mature and widely used biometric identification technology, has been deeply involved in many fields such as mobile payment, access control and information security. In order to protect the security of the fingerprint recognition system and prevent fraud attacks by counterfeit products such as fake fingerprints, live fingerprint detection technology has emerged as the times require. The core of this technology is to analyze whether the collected fingerprint image or signal contains live biological characteristics to distinguish real fingers from artificial replicas, which is a key link in building a secure authentication system.

[0003] Currently, existing live fingerprint detection technologies mainly rely on detecting a single physiological characteristic. For example, some methods analyze the blood flow changes in the fingerprint area, i.e. photoplethysmography signals, to determine the live body; other methods use the elasticity deformation or micro-pulsation of the skin when the finger is pressed as the basis for live body. In addition, some technologies also attempt to fuse multiple characteristics, but usually require the configuration of multiple independent sensors, such as using optical sensors and pressure sensors to collect different signals.

[0004] However, the above existing technical solutions have obvious limitations. Detection methods based on a single characteristic are easily cracked, for example, a silicone finger mold with a micro light source can simulate blood flow signals, and a fake fingerprint made of high-elasticity material can simulate skin deformation. For solutions using multiple sensors, it is difficult to achieve accurate spatio-temporal synchronization between different sensors, and the system structure is complex and the cost is high. Even if multiple characteristics are fused, if a fixed fusion strategy is used, it is difficult to adapt to complex and variable use environments, for example, strong environmental light can seriously interfere with the quality of optical signals, and the dryness or wetness of the user's skin can affect the collection effect of multiple physiological signals, resulting in unstable detection performance. SUMMARY

[0005] In view of this, in order to solve the problems raised in the background art, a live fingerprint detection method based on fusion of multispectral and micro-arterial pulse characteristics is proposed.

[0006] The purpose of the present application can be achieved by the following technical solutions: The present application provides a live fingerprint detection method based on fusion of multispectral and micro-arterial pulse characteristics, comprising the following steps: S1, signal synchronization generation: acquiring a multispectral reflectance intensity sequence and a skin sub-pixel level displacement fluctuation sequence of the fingerprint area collected synchronously, and generating a spatio-temporal synchronous original signal.

[0007] S2, time domain alignment processing: the spatiotemporal synchronization original signal is subjected to time domain alignment processing to form a set of spatiotemporal correlation data pairs.

[0008] S3, feature matrix generation: the spectral absorption feature vector of the multispectral reflection intensity sequence in the set of spatiotemporal correlation data pairs is extracted, and the pulsation trajectory feature vector of the skin sub-pixel level displacement fluctuation sequence at the corresponding time point is synchronously extracted, the spectral absorption feature vector and the pulsation trajectory feature vector are fused and coded according to the time stamp to generate a dynamic coupling feature matrix.

[0009] S4, weight dynamic adjustment: the weight ratio of the spectral absorption feature and the pulsation trajectory feature in the dynamic coupling feature matrix is dynamically adjusted according to the environmental interference parameter to generate an adaptive weight decision rule.

[0010] S5, cross-validation analysis: the dynamic coupling feature matrix is subjected to cross-validation analysis based on the adaptive weight decision rule to output a local feature abnormal area identifier.

[0011] S6, abnormal area sampling: an enhanced sampling instruction is triggered for the local feature abnormal area identifier to obtain supplementary feature data of an extended collection period.

[0012] S7, living body model verification: the supplementary feature data is input into a preset living body determination model for physiological coupling rule matching to generate a living body verification result.

[0013] S8, living body authentication output: based on the living body verification result, a living body fingerprint authentication pass signal is output.

[0014] Compared with the prior art, the embodiments of the present application have at least the following advantages or beneficial effects: (1) the present application synchronously collects physiological signals of two different modalities of multispectral reflection intensity and skin sub-pixel level displacement fluctuation, and performs deep fusion analysis thereon. Since the signals originate from the same physical region and are perfectly synchronized in time, the registration and synchronization errors of the multi-sensor scheme are fundamentally eliminated, ensuring the high fidelity and intrinsic correlation of the original data, thereby providing a reliable data basis for accurate discrimination, and significantly improving the accuracy of living body detection.

[0015] (2) the present application introduces an environment adaptive weight decision mechanism, which can monitor environmental light intensity and skin dryness and other interference factors in real time, and dynamically adjusts the weight ratio of the spectral absorption feature and the pulsation trajectory feature in the final decision. This intelligent adaptive capability enables the system to selectively rely on the signal source with better quality, effectively compensating for signal distortion in complex and variable application scenarios, thereby enhancing the robustness and environmental adaptability of the detection algorithm.

[0016] (3) The application establishes a hierarchical intelligent verification process, when the preliminary cross-validation finds that there is a contradiction in the local characteristics, instead of being directly rejected, the enhanced sampling and deep verification for the abnormal area are triggered. The process can distinguish the real physiological abnormalities from the malicious fake attacks with higher confidence by analyzing the stability and physiological coupling law of the extended period signal. This fine processing strategy effectively reduces the false rejection rate caused by individual differences or poor contact under the premise of ensuring high security, and improves the user experience. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the application, the following will briefly introduce the drawings needed to be used in the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0018] Figure 1 It is a schematic diagram of the method steps of the application.

[0019] Figure 2 It is a conflict area identification diagram of the application.

[0020] Figure 3 It is a phase shift and fluctuation amplitude relationship diagram of the application. DETAILED DESCRIPTION

[0021] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only some of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the application.

[0022] Please refer to Figure 1 The application provides a living fingerprint detection method based on multispectral and micro-artery pulse feature fusion, which comprises the following steps: S1, signal synchronization generation: acquiring a multispectral reflection intensity sequence and a skin sub-pixel level displacement fluctuation sequence of a fingerprint area collected synchronously, and generating a time-space synchronous original signal.

[0023] In specific embodiments of the application, the specific process of generating the time-space synchronous original signal is: controlling the multi-band light source to alternately irradiate the fingerprint area at a preset frequency.

[0024] It is noted that the control system drives a multi-band light source integrated with multiple different central wavelength light units, such as blue light with wavelength of 450nm, green light with wavelength of 530nm, red light with wavelength of 660nm and near-infrared light with wavelength of 940nm. The control system cyclically triggers these light units at a fixed and accurate preset frequency, so that they alternately irradiate the fingerprint area placed on the sensor surface.

[0025] In one embodiment of the present application, the preset frequency is specifically 100Hz, which is based on the fact that the frequency needs to be accurately matched with the acquisition capability of the high-frame-rate macro optical sensor. The sensor synchronously captures the skin reflection images under each band of light with millisecond-level time accuracy, such as capturing the first frame of image immediately when the blue light is on, capturing the second frame of image when the green light is on, and cyclically triggering the four band light sources of 450nm blue light, 530nm green light, 660nm red light and 940nm near-infrared light at a fixed frequency of 100Hz, so as to ensure that the timing of each band of light and the exposure of the sensor strictly correspond, thereby generating a raw multispectral video stream without synchronization error, and providing a high-fidelity data basis for subsequent extraction of skin sub-pixel level displacement fluctuation sequence.

[0026] The skin surface reflection image sequence under each band of light is synchronously captured by the macro optical sensor.

[0027] It is noted that while the light sources are alternately irradiated, a high-frame-rate macro optical sensor synchronously captures the skin surface reflection image sequence of the fingerprint area under each single band of light. The synchronization here is a key technical feature, which means that the exposure time of each frame of image of the sensor strictly corresponds to the lighting time of a certain specific band of light source. For example, at the millisecond, the blue light is on, and the sensor captures the first frame of image; at the millisecond, the green light is on, and the sensor captures the second frame of image, and so on. Such continuous acquisition forms a raw video stream containing multispectral information, i.e. the skin surface reflection image sequence.

[0028] The multispectral reflection intensity sequence is separated from the skin surface reflection image sequence.

[0029] It is noted that the skin surface reflection image sequence is processed in parallel to separate two target signals. The first one is the multispectral reflection intensity sequence. The processing method is to demarcate multiple regions of interest in the image sequence, which are usually selected at positions with good quality of fingerprint image. For each frame of image, the average pixel gray value in the corresponding region of interest is automatically obtained. Since each frame corresponds to a known band of light source, organizing these average gray values in time sequence and band information forms the multispectral reflection intensity sequence . Represents the intensity of reflection. Represents spectral bands, Represents time.

[0030] The displacement fluctuation of the skin surface between adjacent frames is calculated based on the image subpixel displacement algorithm, and a skin subpixel displacement fluctuation sequence is generated.

[0031] It should be noted that the specific process of generating the skin subpixel displacement fluctuation sequence is as follows: 1) Selecting frame pairs: Selecting two adjacent frames from a continuous image sequence; 2) Calculating correlation: Calculating the correlation between the two frames. and Performing two-dimensional Fourier transforms on each yields the frequency domain representation: ,in, The cross-power spectrum is obtained by multiplying the frequency domain representations of the two frames conjugately and normalizing the results. , ;in, yes The complex conjugate, with the denominator being the modulus normalization term, for the cross power spectrum Perform a two-dimensional inverse Fourier transform: ; It is a matrix where the value of each point represents the similarity between two frames at that offset; 3) Subpixel localization: Locating peak points: Finding the maximum point in the correlation matrix, and its coordinates This refers to the integer pixel-level displacement between two frames of images. Within a small neighborhood around the peak point, a quadratic function is used to fit the data to obtain more accurate coordinates of the extreme points, assuming that the area around the peak point conforms to a quadratic surface model. Parameters fitted using the least squares method extreme point coordinates This is subpixel-level displacement; 4) Generate sequence: Repeat the above calculation for all consecutive frame pairs in the entire image sequence to obtain a time-varying two-dimensional displacement vector sequence. The magnitude of this sequence or its projection in a specific direction constitutes the skin subpixel-level displacement fluctuation sequence. ,in t represents the displacement fluctuation, and t represents time.

[0032] The embodiment of the present application can extract two different modal physiological signals of living body from the same part of the reflection image sequence collected by a single micro-distance optical sensor. The single-source synchronous acquisition mechanism fundamentally ensures the perfect alignment of the two signals in time, avoiding the synchronization error and calibration difficulty that may be introduced by using a multi-sensor solution. It provides a high-quality and high-fidelity original data basis for subsequent feature fusion analysis, significantly improving the reliability and robustness of the living fingerprint detection algorithm in distinguishing real fingers from fake fingerprints, because the fake material is difficult to simulate the accurate physiological coupling relationship between the two signals.

[0033] S2, time domain alignment processing: performing time domain alignment processing on the space-time synchronous original signal to form a set of space-time correlation data pairs.

[0034] In a specific embodiment of the present application, the specific process of forming a set of space-time correlation data pairs is to identify the starting synchronization timestamp of the multispectral reflection intensity sequence and the skin sub-pixel level displacement fluctuation sequence.

[0035] It should be noted that the starting point of the processing is to identify the starting synchronization timestamp of the multispectral reflection intensity sequence and the skin sub-pixel level displacement fluctuation sequence. Although the two signals are synchronized during acquisition, the analysis usually starts from a certain time point after the fingerprint and the sensor are stably contacted to avoid noise interference in the initial stage. This time point is the starting synchronization timestamp, denoted as .

[0036] The double sequence is slidingly intercepted with a preset time window to generate a time-aligned data segment.

[0037] In a specific embodiment of the present application, the length of the preset time window is It is usually set according to the general rule of human heart cycle, and can be 1.5 seconds, to ensure that each window can completely capture the physiological information of at least one heartbeat event, and the step length needs to ensure that each sliding can cover at least half of the heart cycle, so as to stably capture the physiological coupling features of skin displacement and spectral reflection. Therefore, the step length can be 0.75 seconds, and the window overlaps 50% after each sliding, which can ensure the continuity of the features and control the amount of calculation.

[0038] The spectral reflection intensity value and the displacement fluctuation amount at the same time point in each time window are bound as a correlation data pair.

[0039] It should be noted that for the data intercepted in any time window, there is a vector composed of different spectral band illumination intensities and a corresponding skin displacement amount , wherein indicates different spectral bands, denotes the total number of spectral bands, denotes the discrete time sampling point number, . The method binds the spectral reflectance intensity vector of the same time point in each time window and the skin displacement amount as an associated data pair. For example, at time point , the generated data pair can be represented as . The associated data pairs generated by all sampling points in a time window are collected, and the associated data pairs generated by all sliding windows are combined, and finally the spatiotemporal associated data pair set is formed.

[0040] The embodiment of the present application successfully converts two continuous and dynamically changing physiological signal streams into a series of discrete, structured and internal time sequence relationship preserved spatiotemporal associated data pair sets. The core technical effect of this processing method is that it not only ensures the alignment of multi-modal data in a macro time period, but also realizes the accurate binding in a micro time point. This lays a solid foundation for subsequent feature extraction, ensures that the features extracted from the spectral signal can be directly compared and fused with the physical displacement features caused by the arterial pulsation at the same instant, so as to deeply mine the internal physiological coupling rules between the two signals, greatly improving the analysis accuracy and reliability of the living body detection model.

[0041] S3, feature matrix generation: extract the spectral absorption feature vector of the multispectral reflectance intensity sequence in the spatiotemporal associated data pair set, and synchronously extract the pulsation trajectory feature vector of the skin sub-pixel level displacement fluctuation sequence at the corresponding time point, fuse and encode the spectral absorption feature vector and the pulsation trajectory feature vector according to the time stamp, and generate a dynamic coupling feature matrix.

[0042] In specific embodiments of the present application, the specific process of generating a dynamic coupling feature matrix is: calculating the gradient value of the spectral reflectance intensity with respect to the wavelength band in each associated data pair to form a spectral absorption feature vector.

[0043] It should be noted that the gradient value of the spectral reflectance intensity with respect to the wavelength band in each associated data pair is calculated to form a spectral absorption feature vector. For the associated data pair at any time in a time window , the multispectral part is extracted. The calculation of the spectral absorption feature vector is completed by evaluating the change rate of the spectral intensity between adjacent wavelength bands, i.e. the spectral gradient. For example, the first element of the vector can be calculated as follows: , wherein, is the value of the first element of the vector, and They are time points Below, the center wavelength is and The reflected intensity value corresponding to the light source.

[0044] Extract the frequency domain energy distribution of displacement fluctuations within the corresponding time window to form the pulsation trajectory feature vector.

[0045] It should be noted that the frequency domain energy distribution of displacement fluctuations within the corresponding time window is extracted to construct the pulsation trajectory feature vector. This step focuses on the entire time window. Subpixel displacement fluctuation sequence of skin within The specific process for constructing the pulsation trajectory feature vector is as follows: 1) Fast Fourier Transform: using a programming language to analyze the sub-pixel-level displacement fluctuation sequence of the skin. Perform a Fast Fourier Transform to obtain the frequency domain sequence. 2) Determine the frequency resolution and frequency range: Frequency resolution ,in It is the sampling frequency. It is the length of the time-domain sequence, representing the interval between two adjacent frequency points in the frequency domain. According to the sampling theorem, the frequency range in the frequency domain is from 0 to... 3) Locating the physiological heart rate frequency range: Based on actual needs and research subjects, determine the specific physiological heart rate frequency range. Based on the frequency resolution and frequency range, find the frequency domain sequence. The frequency point index range corresponding to the physiological heart rate frequency range, assuming frequency points Then find the satisfying 4) Calculate energy: frequency domain sequence It is a complex number, and its magnitude is... It can be done ;in and They are The real and imaginary parts. For each frequency point within the physiological heart rate frequency range. Calculate its amplitude For each frequency point Its energy Within the physiological heart rate frequency range, the energy at each frequency point is calculated; 5) The energy or amplitude of each frequency component within the calculated physiological heart rate frequency range is arranged in ascending order of frequency, thus forming the pulsation trajectory feature vector of this time window. .

[0046] The spectral absorption feature vectors and pulsation trajectory feature vectors of the same time window are tensor-concatenated to form a three-dimensional feature matrix.

[0047] It should be noted that the spectral absorption feature vector and the pulsation trajectory feature vector of the same time window are tensor spliced to form a three-dimensional feature matrix. In this step, the two features calculated in the same time window are fused. Since the spectral absorption feature vector varies with the time point, and the pulsation trajectory feature vector is a summary of the entire window, therefore, the fusion needs to align the latter with the former in the time dimension. The specific operation is that for each time point in the time window, the global pulsation trajectory feature vector is appended to the spectral absorption feature vector of the moment. The new combined vector of all time points in the window is stacked in time order, which forms a two-dimensional matrix. When we stack such two-dimensional matrices generated by multiple time windows in succession and possibly overlapping along the new dimension, we form the three-dimensional feature matrix, i.e. the dynamic coupling feature matrix, described in the present application. The three dimensions represent the time window index, the time step within the window, and the fused feature, respectively.

[0048] S4, dynamic adjustment of weight: dynamically adjusting the weight ratio of the spectral absorption feature and the pulsation trajectory feature in the dynamic coupling feature matrix according to the environmental interference parameter, to generate an adaptive weight decision rule.

[0049] In specific embodiments of the present application, the specific process of generating the adaptive weight decision rule is: detecting the ambient light intensity parameter and the skin dryness parameter as the environmental interference parameters.

[0050] It should be noted that the ambient light intensity parameter can be directly measured by a special ambient light sensor integrated in the fingerprint sensor module. The skin dryness parameter is obtained by a professional skin impedance meter.

[0051] When the ambient light intensity exceeds the preset threshold, the weight coefficient of the spectral absorption feature vector is reduced.

[0052] It should be noted that strong ambient light will interfere with the signal emitted by the active light source, pollute the purity of the spectral absorption feature vector, and reduce the signal-to-noise ratio. Therefore, an ambient light threshold is set. When , the weight coefficient of the spectral feature will be adjusted, , is the weight coefficient of the adjusted spectral feature, is the initial weight coefficient of the spectral feature, represents a natural constant, denotes the attenuation factor, and the specific value is obtained by a large number of experimental tests, and the signal-to-noise ratio of the spectral absorption feature vector is measured under different ambient light intensities, .

[0053] In one specific embodiment of the present application, the ambient light threshold can be set to 700 lux. If the threshold is set too low, the weight coefficient of the spectral absorption feature vector may be reduced under normal lighting conditions, affecting normal signal analysis. If the threshold is set too high, the strong ambient light interference may not be effectively processed in time. Therefore, setting the threshold to 700 lux can ensure normal signal processing while effectively dealing with strong ambient light interference.

[0054] When the skin dryness exceeds the preset threshold, the weight coefficient of the pulsation trajectory feature vector is increased.

[0055] It should be noted that excessive dryness of the skin can reduce the efficiency of light penetration to the dermis, resulting in a decrease in the spectral signal reflecting blood flow changes. However, the impact on the skin mechanical displacement directly driven by arterial pulsation, i.e., the pulsation trajectory feature, is relatively small. Therefore, a skin dryness threshold is set When , the weight coefficient of the pulsation trajectory feature is increased, , denotes the weight coefficient of the adjusted pulsation trajectory feature, is the initial weight coefficient of the pulsation trajectory feature, denotes the adjustment coefficient, which determines the rate at which the function increases with the skin dryness, and the specific value is obtained by a large number of experimental tests, .

[0056] In one specific embodiment of the present application, through a large number of experiments and clinical data statistics, when the skin is measured by a skin impedance meter, if the skin impedance value measured by the skin impedance meter is greater than 70 kΩ, it is considered that the skin is in a relatively dry state. Therefore, the skin dryness threshold is set to 70 kΩ.

[0057] The decision boundary condition of the dynamic coupling feature matrix is reconstructed based on the weight coefficient.

[0058] It should be noted that the decision boundary condition of the dynamic coupling feature matrix is reconstructed based on the weight coefficient. This step does not directly modify the numerical value of the dynamic coupling feature matrix, but applies these weights in subsequent classification or discriminant analysis. When the living body determination model analyzes the feature matrix, it will multiply the extracted spectral absorption feature by the adjusted weight coefficient , and multiply the pulsation trajectory feature by This essentially changes the contribution of different features in the final decision, thereby dynamically adjusting the decision hyperplane, i.e. the decision boundary condition, in the high-dimensional feature space that distinguishes live from fake fingerprints.

[0059] S5, cross-validation analysis: cross-validation analysis is performed on the dynamic coupling feature matrix based on the adaptive weight decision rule, and a local feature abnormal area identifier is output.

[0060] Referring to Figure 2 In the specific process of outputting the local feature abnormal area identifier in the embodiment of the application, under the reconstructed decision boundary condition, a conflict area is identified in which the spectral absorption feature conforms to the live body characteristics but the pulsation trajectory feature is lower than the intensity threshold.

[0061] It should be noted that the cross-validation analysis method of the application is a key link for checking the consistency of different modal features in the dynamic coupling feature matrix under the guidance of the adaptive weight decision rule generated in the previous step. This analysis is not an independent judgment of a single feature, but a check of the inherent correlation of two features, and the process is carried out in the spatial dimension of the fingerprint image. First, under the reconstructed decision boundary condition, a conflict area is identified in which the spectral absorption feature conforms to the live body characteristics but the pulsation trajectory feature is lower than the intensity threshold. The specific operation is to divide the effective area of the fingerprint into multiple sub-areas for analysis. For each sub-area, the two features are evaluated in parallel. On the one hand, a preset live spectrum model is used to determine whether the spectral absorption feature of the area conforms to the typical mode of the live blood absorption spectrum. On the other hand, it is detected whether the energy of the pulsation trajectory feature corresponding to the area exceeds a preset pulsation intensity threshold. When the analysis result of a certain sub-area shows that the spectral absorption feature is determined to conform to the typical mode of the live blood absorption spectrum, but the energy of the pulsation trajectory feature fails to reach the threshold, this feature contradiction is identified, and the sub-area is defined as a conflict area.

[0062] It should also be noted that the specific way of using the preset live spectrum model to determine whether the spectral absorption feature of the area conforms to the typical mode of the live blood absorption spectrum is as follows: a threshold interval of the spectral absorption feature determined based on a large amount of live blood sample spectral data is extracted from the preset live spectrum model, the spectral absorption feature value extracted from the spectrum of the sub-area of the fingerprint is compared with the threshold interval, if the spectral absorption feature value is located within the threshold interval, it is determined that the spectral absorption feature value of the area conforms to the typical mode of the live blood absorption spectrum, and if the spectral absorption feature value is not located within the threshold interval, it is determined that the spectral absorption feature value of the area does not conform to the typical mode of the live blood absorption spectrum.

[0063] In one specific embodiment of the present application, the preset pulsation intensity threshold value can be set as 0.2, which is based on the analysis of the pulsation trajectory feature energy of a large number of healthy living body finger samples. Statistics show that when the normal living body blood pulsation corresponds to a certain range of regional energy, the value in the range that can accurately distinguish the living body and non-living body and has a certain fault tolerance is taken as the typical value to ensure that the pulsation conforming to the living body characteristics can be effectively identified, and false judgments caused by individual differences or environmental interference and other factors can be avoided.

[0064] Marking the conflict region as a local feature abnormal region identifier.

[0065] It should be noted that the conflict region is marked as a local feature abnormal region identifier. This step aggregates all sub-regions identified as conflict regions to generate a marker, which can be a binary mask image, where the conflict region corresponds to a pixel value of 1 and other regions are 0, to accurately indicate the spatial position and range of feature inconsistency on the fingerprint image.

[0066] Calculate the proportion of the total area of the conflict region to the total effective area of the fingerprint, i.e. the conflict region area ratio. When the conflict region area ratio exceeds the preset tolerance, send an enhanced sampling instruction to the acquisition module.

[0067] It should be noted that the total effective area of the fingerprint is usually obtained by directly acquiring the original fingerprint image through an optical fingerprint sensor, and then accurately circumscribing and identifying the total effective area of the fingerprint through image preprocessing. The preset tolerance value can be set as 15%. The value is based on a large number of experimental tests, which comprehensively considers the reasonable conflict region fluctuation caused by individual differences, equipment errors and other factors during normal fingerprint acquisition, and sets a threshold ratio that can tolerate certain reasonable errors and effectively identify abnormal situations to ensure the accuracy and reliability of fingerprint living body detection and avoid false judgments caused by small conflict regions.

[0068] S6, Abnormal region sampling: Trigger enhanced sampling instruction for local feature abnormal region identifier to obtain supplementary feature data of extended acquisition period.

[0069] In specific embodiments of the present application, the specific process of obtaining supplementary feature data of an extended acquisition period is to extend the data acquisition time length of the position corresponding to the local feature abnormal region identifier to a set multiple of the reference time length.

[0070] It should be noted that the data acquisition time length of the position corresponding to the local feature abnormal region identifier is extended to a set multiple of the reference time length, where the reference time length is the preset time window length used in the initial analysis. .

[0071] In one embodiment of the present application, the set multiple can be 2, aiming to capture at least two complete physiological cardiac cycles, and the lengthening of this time is crucial for confirming the periodicity of the signal, because fluctuations in a single cycle can be caused by random noise or artifacts, while stable repetition of multiple consecutive cycles is strong evidence of a living physiological signal.

[0072] Extract the periodic fluctuation rule of the pulsation trajectory feature in the extended acquisition cycle.

[0073] It should be noted that extracting the periodic fluctuation rule of the pulsation trajectory feature in the extended acquisition cycle refers to analyzing the skin sub-pixel level displacement fluctuation sequence after obtaining more complete physiological cycle data by extending the data acquisition time of the local feature abnormal area, and extracting the pulsation frequency and pulsation amplitude, i.e. the feature parameters such as the number of pulsations per unit time and the displacement fluctuation amount, which exhibit periodic changes over time, from the sequence. These parameters maintain a relatively stable repetition pattern in different physiological cycles.

[0074] Referring to Figure 3 As shown in S7, living body model verification: input the supplementary feature data into the preset living body determination model for physiological coupling rule matching to generate a living body verification result.

[0075] In an embodiment of the present application, the specific process of generating the living body verification result is to establish a fluctuation amplitude reference range of the spectral absorption feature in the pulse cycle.

[0076] It should be noted that a judgment reference is needed, i.e. to establish a fluctuation amplitude reference range of the spectral absorption feature in the pulse cycle. This reference range is not a temporary calculation, but a preset parameter obtained by prior calibration on a large number of real living body fingerprint samples. It defines a physiologically reasonable interval representing the normal fluctuation amplitude of spectral absorption intensity caused by blood volume change in a single cardiac cycle.

[0077] Detect the phase shift amount of the spectral absorption fluctuation and the pulsation trajectory wave peak in the supplementary feature data.

[0078] It should be noted that the actual detection stage of the currently acquired supplementary feature data is entered. The first detection is to detect the phase shift amount of the spectral absorption fluctuation and the pulsation trajectory wave peak in the supplementary feature data. From the supplementary feature data, the start point of each pulse cycle and the time point of the wave peak are accurately identified by analyzing the periodicity of the skin sub-pixel level displacement fluctuation sequence. At the same time, the change of the multi-spectral reflection intensity sequence is analyzed in the same time period to find the time of the absorption feature extreme point caused by the maximum blood filling.Due to the characteristics of hemodynamics, there is a small but stable time delay between the mechanical pressure wave peak and the peak of blood volume. The phase shift is quantified by calculating this time difference .

[0079] When the fluctuation amplitude is within the reference range and the phase shift is less than the preset error threshold, it is determined that the living body physiological coupling rule is met; when the fluctuation amplitude is not within the reference range or the phase shift is greater than or equal to the preset error threshold, it is determined that the living body physiological coupling rule is not met.

[0080] It should be noted that the final determination is made according to the preset coupling rule. The determination condition includes two aspects. On the one hand, the average fluctuation amplitude of the spectral absorption feature in the supplementary feature data in multiple complete pulse cycles is obtained , and it is verified whether it falls within the preset reference range, i.e. . On the other hand, the calculated phase shift is compared with a preset error threshold. Only when the fluctuation amplitude is within the reference range and the phase shift is less than the preset error threshold, it is finally determined that the living body physiological coupling rule is met.

[0081] In a specific embodiment of the present application, the typical value of the preset error threshold can be set to 0.1 seconds. The value is based on the physiological characteristics of a large number of real living body fingerprint samples, considering that there is a certain time delay from the heart beat to the mechanical displacement of the finger skin, but this delay is relatively stable and within a small range. Through experimental statistical analysis, it is found that when the phase shift is less than 0.1 seconds, the living body and non-living body fingerprints can be accurately distinguished, which not only avoids misjudgment due to individual differences, but also effectively eliminates abnormal situations that may occur due to the inability of fake fingerprints to simulate this accurate physiological coupling relationship, thereby ensuring the accuracy and reliability of living body detection.

[0082] S8, living body authentication output: based on the living body verification result, output living body fingerprint authentication pass signal.

[0083] In a specific embodiment of the present application, the specific process of outputting the living body fingerprint authentication pass signal is: when the living body verification result is not in accordance with the living body physiological coupling rule for three times in a row, activate the anti-fake attack alarm protocol.

[0084] Freeze the current fingerprint authentication process and record the attack feature mode.

[0085] Upload the attack feature mode to the security authentication center database.

[0086] It should be noted that the additional security measures performed after outputting the living body verification result is an intelligent response protocol designed to deal with persistent attack attempts. This process is not triggered immediately after a single verification failure, but is based on a cumulative count mechanism of consecutive failures. The system maintains a failure counter for each authentication session. After completing each complete living body detection process and generating a living body verification result, if the verification result is not in line with the living body physiological coupling rule, it means that this attempt is judged as non-living body, and the failure counter is incremented. Otherwise, if any verification is successful, the counter is immediately cleared.

[0087] It should also be noted that the trigger condition of the intelligent response protocol is when the verification result is not in line with the living body physiological coupling rule for three consecutive times. That is, when the failure counter accumulates to 3, the system determines that it is currently being subjected to persistent, and likely attack attempts, and automatically activates the anti-fraud attack alarm protocol. This protocol is a set of predefined, automatically executed response procedures. Once the protocol is activated, the current fingerprint authentication process will be immediately frozen, the system will stop accepting any new fingerprint input, and the authentication lock or failure information will be displayed to the user interface, effectively preventing attackers from finding system vulnerabilities through repeated trial and error. At the same time, the system will perform data solidification operations, that is, record attack feature patterns. This pattern is a data packet containing all key information that leads to consecutive failures, which may include multispectral reflectance intensity sequences, skin sub-pixel level displacement fluctuation sequences, extracted spectral absorption features and pulsation trajectory features, finally generated dynamic coupling feature matrix, and all intermediate analysis steps. Abnormal identification, etc. Finally, the system will package the attack feature pattern and upload it to the secure authentication center database through an encrypted channel.

[0088] The above is only an example and description of the concept of the present application. Those skilled in the art can make various modifications or supplements to the described specific embodiments or use similar ways to replace them, as long as they do not deviate from the concept of the present application or exceed the scope defined by the present application. It should belong to the protection scope of the present application.

Claims

1. A living body fingerprint detection method based on fusion of multispectral and micro-arterial pulse features, characterized in that, The method comprises the following steps: S1, signal synchronization generation: acquiring a multispectral reflectance intensity sequence and a skin sub-pixel level displacement fluctuation sequence of a fingerprint area collected synchronously, and generating a time-space synchronous original signal; S2, time domain alignment processing: performing time domain alignment processing on the time-space synchronous original signal to form a time-space correlation data pair set; The specific process of forming the time-space correlation data pair set is: identifying the starting synchronization time stamp of the multispectral reflectance intensity sequence and the skin sub-pixel level displacement fluctuation sequence; slidingly intercepting the two sequences with a preset time window to generate a time-aligned data segment; binding the spectral reflectance intensity value and the displacement fluctuation value at the same time point in each time window as a correlation data pair; S3, feature matrix generation: extracting a spectral absorption feature vector of the multispectral reflectance intensity sequence in the time-space correlation data pair set, and synchronously extracting a pulsation trajectory feature vector of the skin sub-pixel level displacement fluctuation sequence at the corresponding time point, fusing and encoding the spectral absorption feature vector and the pulsation trajectory feature vector according to the time stamp to generate a dynamic coupling feature matrix; The specific process of generating the dynamic coupling feature matrix is: calculating the gradient value of the spectral reflectance intensity varying with the wave band in each correlation data pair to form a spectral absorption feature vector; extracting the frequency energy distribution of the displacement fluctuation value in the corresponding time window to form a pulsation trajectory feature vector; tensor splicing the spectral absorption feature vector and the pulsation trajectory feature vector in the same time window to form a three-dimensional feature matrix; S4, weight dynamic adjustment: dynamically adjusting the weight proportion of the spectral absorption feature and the pulsation trajectory feature in the dynamic coupling feature matrix according to the environmental interference parameter to generate an adaptive weight decision rule; S5, cross-validation analysis: performing cross-validation analysis on the dynamic coupling feature matrix based on the adaptive weight decision rule to output a local feature abnormal region identification; S6, abnormal region sampling: triggering an enhanced sampling instruction for the local feature abnormal region identification to acquire supplementary feature data of an extended collection period; S7, living body model verification: inputting the supplementary feature data into a preset living body judgment model to perform physiological coupling rule matching to generate a living body verification result; S8, living body authentication output: outputting a living body fingerprint authentication pass signal based on the living body verification result.

2. The method according to claim 1, characterized in that: The specific process of generating the time-space synchronous original signal is: controlling a multi-waveband light source to alternately irradiate the fingerprint area at a preset frequency; synchronously capturing a skin surface reflection image sequence under each waveband light irradiation through a macro optical sensor; separating a multispectral reflectance intensity sequence from the skin surface reflection image sequence; calculating the displacement fluctuation of the skin surface between adjacent frames based on an image sub-pixel displacement algorithm to generate a skin sub-pixel level displacement fluctuation sequence.

3. The method according to claim 1, characterized in that: The specific process of generating the adaptive weight decision rule is: detecting environmental light intensity parameters and skin dryness parameters as environmental interference parameters; when the environmental light intensity exceeds a preset threshold, reducing the weight coefficient of the spectral absorption feature vector; when the skin dryness exceeds a preset threshold, increasing the weight coefficient of the pulsation trajectory feature vector; reconstructing the decision boundary condition of the dynamic coupling feature matrix based on the weight coefficient.

4. The method according to claim 3, characterized in that: The specific process of outputting the local feature abnormal region identification is: Under the reconstructed decision boundary condition, a conflict region is identified, in which the spectral absorption feature meets the living body characteristics, but the pulsation trajectory feature is lower than the intensity threshold; The conflict region is marked as a local feature anomaly region identifier; A ratio of a total area of the conflict region to an effective total area of the fingerprint is calculated, i.e., a conflict region area ratio, and when the conflict region area ratio exceeds a preset tolerance, an enhanced sampling instruction is sent to the acquisition module.

5. The method according to claim 4, characterized in that: The specific process of acquiring the supplemental feature data with the extended acquisition period is as follows: The data acquisition time length of the local feature anomaly region identifier corresponding position is extended to a set multiple of the reference time length; The periodic fluctuation rule of the pulsation trajectory feature in the extended acquisition period is extracted.

6. The method according to claim 5, characterized in that: The specific process of generating the living body verification result is as follows: A fluctuation amplitude reference range of the spectral absorption feature in a pulse cycle is established; The phase offset amount of the spectral absorption fluctuation and the pulsation trajectory wave peak in the supplemental feature data is detected; When the fluctuation amplitude is within the reference range and the phase offset amount is less than a preset error threshold, it is determined that the living body physiological coupling rule is met, and when the fluctuation amplitude is not within the reference range or the phase offset amount is greater than or equal to the preset error threshold, it is determined that the living body physiological coupling rule is not met.

7. The method according to claim 6, characterized in that: The specific process of outputting the living body fingerprint authentication pass signal is as follows: When the living body verification result is not in line with the living body physiological coupling rule for three consecutive times, an anti-fraud attack alarm protocol is activated; The current fingerprint authentication process is frozen and an attack feature mode is recorded; The attack feature mode is uploaded to a secure authentication center database.

Images