Biometric detection method, apparatus, electronic device, and storage medium

Abstract

This disclosure relates to a biometric detection method, apparatus, electronic device, and storage medium. The method acquires skin feature signals corresponding to the skin of a target area within a preset time interval to characterize changes in skin image feature values. The skin feature signals are then processed to obtain pulse wave signals, which are used to calculate the biometric information of the target object. This disclosure acquires image features of the skin in the target area of ​​the target object to determine the pulse wave signal, and further calculates biometric information based on the pulse wave signal, automatically and accurately detecting biometric information without contact. Furthermore, the entire detection process requires no manual intervention, has a high degree of automation, reduces the detection difficulty for users, and improves the user experience.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, and in particular to a biometric detection method, apparatus, electronic device, and storage medium. Background Technology

[0002] With the increasing proportion of an aging population and the rapid rise in the incidence of chronic diseases, the demand for continuous monitoring of vital signs such as heart rate and blood pressure, as well as risk warnings for specific populations, has increased significantly. However, healthcare organizations are constrained by limited human and material resources, making it challenging to accommodate large numbers of patients over a long period. Society urgently needs a system that can provide daily health monitoring and emergency early warning for the elderly or those living alone, thereby improving national well-being. Currently available biometric detection methods are all contact-based, requiring users to operate them themselves. The process is relatively complex, and it is difficult for some patients to perform the tests independently. Summary of the Invention

[0003] In view of this, this disclosure proposes a biometric detection method, device, electronic device and storage medium, aiming to provide a simple and accurate contactless biometric detection method.

[0004] According to a first aspect of this disclosure, a biometric detection method is provided, the method comprising:

[0005] Within a preset time interval, skin feature signals corresponding to the target area skin of the target object are acquired, and the skin feature signals characterize the changes in skin image feature values.

[0006] The skin feature signals are processed to obtain pulse wave signals;

[0007] The biometric information of the target object is calculated based on the pulse wave signal, and the biometric information is used to characterize at least one vital sign of the target object.

[0008] In one possible implementation, acquiring the skin feature signal corresponding to the target area skin of the target object within a preset time interval includes:

[0009] Within the preset time interval, skin images of the target area of ​​the target object are collected multiple times at a preset frequency;

[0010] By calculating the image feature value corresponding to each skin image, the corresponding skin feature signal is obtained.

[0011] In one possible implementation, obtaining the corresponding skin feature signal by calculating the image feature value corresponding to each skin image includes:

[0012] Calculate the mean pixel value of the green channel in each skin image to obtain the corresponding feature pixels;

[0013] The original signal trajectory is determined based on the feature pixels corresponding to each skin image;

[0014] The original signal trajectory is subjected to baseline removal processing to obtain skin feature signals.

[0015] In one possible implementation, the signal processing of the skin feature signal to obtain the pulse wave signal includes:

[0016] By separating the interference sources of the skin feature signals, candidate signals are obtained;

[0017] The candidate signal is processed by a trained neural network model to output a pulse wave signal.

[0018] In one possible implementation, the method further includes:

[0019] The neural network model is trained according to a preset training set, which includes multiple sample feature signals and corresponding labeled pulse signals.

[0020] The training process includes:

[0021] Each of the sample feature signals is input into the neural network model, and the corresponding predicted pulse signal is output.

[0022] The corresponding model loss is obtained by calculating the predicted pulse signal and the Pearson correlation coefficient of the labeled pulse signal for each of the sample feature signals.

[0023] Adjust the neural network model based on the model loss.

[0024] In one possible implementation, the biometric information includes at least one of heart rate, blood pressure, and heartbeat duration.

[0025] In one possible implementation, when the biometric information includes heart rate, calculating the biometric information of the target object based on the pulse wave signal includes:

[0026] The pulse wave components were determined by performing independent component analysis on the pulse wave signal.

[0027] The power of the pulse wave component is determined as the corresponding heart rate.

[0028] In one possible implementation, when the biometric information includes the heartbeat period, calculating the biometric information of the target object based on the pulse wave signal includes:

[0029] The peak value of the pulse wave signal was statistically analyzed;

[0030] Calculate the time interval between adjacent peaks to obtain the corresponding interval sequence as the corresponding heartbeat period.

[0031] In one possible implementation, when the biometric information includes blood pressure, calculating the biometric information of the target object based on the pulse wave signal includes:

[0032] The pulse wave signal corresponding to the skin of the target area is identified as the target pulse wave signal;

[0033] Within a preset time interval, acquire the reference pulse wave signal corresponding to the reference area skin of the target object;

[0034] The pulse wave transmission time is determined based on the target pulse wave signal and the reference pulse wave signal;

[0035] The blood pressure corresponding to the target object is calculated based on the pulse wave transmission time.

[0036] According to a second aspect of this disclosure, a biometric detection device is provided, the device comprising:

[0037] The signal acquisition module is used to acquire skin feature signals corresponding to the skin of the target area of ​​the target object within a preset time interval. The skin feature signals represent the changes in skin image feature values.

[0038] The signal processing module is used to process the skin feature signals to obtain pulse wave signals;

[0039] The feature determination module is used to calculate the biometric information of the target object based on the pulse wave signal, wherein the biometric information is used to characterize at least one vital sign of the target object.

[0040] In one possible implementation, the signal acquisition module is further configured to:

[0041] Within the preset time interval, skin images of the target area of ​​the target object are collected multiple times at a preset frequency;

[0042] By calculating the image feature value corresponding to each skin image, the corresponding skin feature signal is obtained.

[0043] In one possible implementation, the signal acquisition module is further configured to:

[0044] Calculate the mean pixel value of the green channel in each skin image to obtain the corresponding feature pixels;

[0045] The original signal trajectory is determined based on the feature pixels corresponding to each skin image;

[0046] The original signal trajectory is subjected to baseline removal processing to obtain skin feature signals.

[0047] In one possible implementation, the signal processing module is further configured to:

[0048] By separating the interference sources of the skin feature signals, candidate signals are obtained;

[0049] The candidate signal is processed by a trained neural network model to output a pulse wave signal.

[0050] In one possible implementation, the device further includes:

[0051] The model training module is used to train the neural network model according to a preset training set, wherein the training set includes multiple sample feature signals and corresponding labeled pulse signals;

[0052] The training process includes:

[0053] Each of the sample feature signals is input into the neural network model, and the corresponding predicted pulse signal is output.

[0054] The corresponding model loss is obtained by calculating the predicted pulse signal and the Pearson correlation coefficient of the labeled pulse signal for each of the sample feature signals.

[0055] Adjust the neural network model based on the model loss.

[0056] In one possible implementation, the biometric information includes at least one of heart rate, blood pressure, and heartbeat duration.

[0057] In one possible implementation, where the biometric information includes heart rate, the feature determination module is further configured to:

[0058] The pulse wave components were determined by performing independent component analysis on the pulse wave signal.

[0059] The power of the pulse wave component is determined as the corresponding heart rate.

[0060] In one possible implementation, when the biometric information includes the period of heartbeat, the feature determination module is further configured to:

[0061] The peak value of the pulse wave signal was statistically analyzed;

[0062] Calculate the time interval between adjacent peaks to obtain the corresponding interval sequence as the corresponding heartbeat period.

[0063] In one possible implementation, where the biometric information includes blood pressure, the feature determination module is further configured to:

[0064] The pulse wave signal corresponding to the skin of the target area is identified as the target pulse wave signal;

[0065] Within a preset time interval, acquire the reference pulse wave signal corresponding to the reference area skin of the target object;

[0066] The pulse wave transmission time is determined based on the target pulse wave signal and the reference pulse wave signal;

[0067] The blood pressure corresponding to the target object is calculated based on the pulse wave transmission time.

[0068] According to a third aspect of this disclosure, an electronic device is provided, comprising: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured to implement the above-described method when executing instructions stored in the memory.

[0069] According to a fourth aspect of this disclosure, a non-volatile computer-readable storage medium is provided that stores computer program instructions thereon, wherein the computer program instructions, when executed by a processor, implement the above-described method.

[0070] According to a fifth aspect of this disclosure, a computer program product is provided, including computer-readable code or a non-volatile computer-readable storage medium carrying the computer-readable code, wherein when the computer-readable code is run in a processor of an electronic device, the processor in the electronic device performs the above-described method.

[0071] In this embodiment, the method acquires skin feature signals corresponding to the skin of a target area within a preset time interval to characterize changes in skin image feature values. The skin feature signals are then processed to obtain pulse wave signals, which are used to calculate the biometric information of the target object. This disclosure acquires image features of the skin in the target area of ​​the target object to determine pulse wave signals, and further calculates biometric information based on these signals, automatically and accurately detecting biometric information without contact. Furthermore, the entire detection process requires no manual intervention, has a high degree of automation, reduces the detection difficulty for users, and improves the user experience.

[0072] Other features and aspects of this disclosure will become clear from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description

[0073] The accompanying drawings, which are included in and form part of this specification, illustrate exemplary embodiments, features, and aspects of this disclosure together with the specification and serve to explain the principles of this disclosure.

[0074] Figure 1 A flowchart of a biometric detection method according to an embodiment of the present disclosure is shown;

[0075] Figure 2 This diagram illustrates a process for determining a training set during a neural network model training process according to an embodiment of the present disclosure.

[0076] Figure 3 A schematic diagram illustrating a process for determining pulse wave propagation time according to an embodiment of the present disclosure is shown;

[0077] Figure 4 A comparative diagram showing the effects of different methods for determining pulse wave transmission time according to embodiments of the present disclosure is provided.

[0078] Figure 5 A schematic diagram of a biometric detection device according to an embodiment of the present disclosure is shown;

[0079] Figure 6 A schematic diagram of an electronic device according to an embodiment of the present disclosure is shown. Detailed Implementation

[0080] Various exemplary embodiments, features, and aspects of this disclosure will now be described in detail with reference to the accompanying drawings. The same reference numerals in the drawings denote elements that have the same or similar functions. Although various aspects of the embodiments are shown in the drawings, they are not necessarily drawn to scale unless specifically indicated otherwise.

[0081] The term “exemplary” as used herein means “serving as an example, embodiment, or illustration.” Any embodiment illustrated herein as “exemplary” is not necessarily to be construed as superior to or better than other embodiments.

[0082] Furthermore, to better illustrate this disclosure, numerous specific details are set forth in the following detailed description. Those skilled in the art will understand that this disclosure can be practiced without certain specific details. In some instances, methods, means, components, and circuits well known to those skilled in the art have not been described in detail in order to highlight the main points of this disclosure.

[0083] The biometric detection method of this disclosure can be executed by an electronic device such as a terminal device or a server. The terminal device can be any fixed or mobile terminal, such as a user equipment (UE), mobile device, user terminal, terminal, cellular phone, cordless phone, personal digital assistant (PDA), handheld device, computing device, vehicle-mounted device, or wearable device. The server can be a single server or a server cluster consisting of multiple servers. Any electronic device can implement the biometric detection method of this disclosure by having its processor call computer-readable instructions stored in its memory.

[0084] Figure 1 A flowchart illustrating a biometric detection method according to an embodiment of this disclosure is shown. Figure 1 As shown, the biometric detection method of this disclosure embodiment may include the following steps S10-S30.

[0085] Step S10: Within a preset time interval, acquire the skin feature signal corresponding to the target area skin of the target object.

[0086] In one possible implementation, when an electronic device needs to perform biometric detection on a target object, it can first acquire skin feature signals corresponding to the skin of the target area of ​​the target object. The target object can be an animal with skin, such as a human, monkey, or lizard. The target area skin can be the skin of a specific region on the target object, and in different application scenarios, the target area can be determined according to the user's needs. For example, when the target object is a human, the target area skin can be facial skin or hand skin. The skin feature signal characterizes the changes in skin image feature values. The skin image can be an image obtained by acquiring images of the target area skin of the target object; that is, the skin feature signal can be a curve showing the change in feature values ​​of multiple skin images acquired within a preset time interval.

[0087] Optionally, the method for determining the skin signal in this embodiment of the disclosure may be to first acquire skin images of the target area of ​​the target object multiple times within a preset time interval and at a preset frequency. Then, the corresponding skin feature signal is obtained by calculating the image feature value corresponding to each skin image. For example, when the target object is a person and the target area of ​​the target object is facial skin.

[0088] This embodiment of the disclosure can first acquire the face of a target object multiple times within a preset time interval and at a preset frequency to obtain an image frame sequence composed of multiple skin images. Optionally, the electronic device can first acquire images of the target object multiple times at a preset frequency, and then perform face recognition and face tracking on the acquired images to extract the facial region as a skin image, thus obtaining an image frame sequence. After obtaining the image frame sequence composed of multiple skin images, the feature value corresponding to each skin image can be extracted as the corresponding image feature value, and the skin feature signal can be further determined according to the acquisition order of the skin images. Specifically, the average pixel value of the green channel of each skin image can be calculated to obtain the corresponding feature pixel, and the feature pixel of each skin image can be used as the corresponding image feature value to determine the original signal trajectory based on the feature pixel corresponding to each skin image. The original signal trajectory is then subjected to baseline removal processing to obtain the skin feature signal.

[0089] Furthermore, target image regions within the skin image can be further identified to extract corresponding image feature values, thereby improving the accuracy of the obtained skin feature signals. For example, a preset number of marker points can be marked on the face of the first frame of the skin image, and a facial bounding box used to determine the size of the face patch can be obtained based on these marker points. The electronic device can directly determine the image region within the facial bounding box as the target image region, or divide the image region within the facial bounding box into multiple small blocks and then randomly select a region within one small block as the target image region. After obtaining the target image region in the first frame of the skin image, target image regions in other skin images can be determined by locating and tracking the target image region. Optionally, the method of extracting image feature values ​​from the target image region can also involve calculating the average pixel value of the green channel of each target image region to obtain the corresponding feature pixels as the image feature values ​​of the skin image.

[0090] Step S20: Perform signal processing on the skin feature signal to obtain the pulse wave signal.

[0091] In one possible implementation, after acquiring the skin feature signal corresponding to the skin of the target object in the target area, the electronic device can obtain a pulse wave signal by performing signal processing on the skin feature signal. This pulse wave signal is used to characterize the arterial pulsation of the target object within a preset time interval. That is, this signal processing is an end-to-end signal processing method, converting one signal into another without changing the format of the processed data. For example, this signal processing may include separating interference sources and noise reduction processing, that is, separating the component characterizing arterial pulsation from the skin feature signal as the pulse wave signal.

[0092] Optionally, the signal processing procedure in this embodiment may include first separating interference sources from the skin feature signal to obtain candidate signals. Then, a trained neural network model is used to perform time-series processing on the candidate signals to output a denoised pulse wave signal. The interference source separation can be achieved by using a Variational Mode Decomposition (VMD) algorithm to remove interference sources from the skin feature signal. The VMD algorithm determines the center frequency and bandwidth of each narrowband intrinsic mode component (BIMF) of the signal by iteratively searching for the optimal solution of the variational model. This essentially decomposes the signal into a combination of various frequency components. The original skin feature signal is decomposed into a sum of multiple sub-signals. Each sub-signal is a narrowband signal concentrated near its respective center frequency, effectively separating the signal's frequency band. This method is equivalent to constructing a Wiener filter based on the center frequency of each component. By considering the narrowband characteristics of these components, the filtered frequency band is ensured to be more concentrated, thereby generating a higher signal-to-noise ratio in the obtained signal components.

[0093] Furthermore, after signal processing to remove interference sources from the skin feature signals and obtain candidate signals, the trained neural network model can further process the candidate signals using time series analysis to denoise them and output the corresponding pulse wave signal. Optionally, this neural network model can be a Long Short-Term Memory (LSTM) network, such as a Deep Recurrent Neural Network (DRNN) as the model architecture. This model consists of four layers, including one LSTM layer followed by three fully connected layers, also known as dense layers. The LSTM layer includes an input gate, a forget gate, a storage unit, and an output gate. The memory cells play a crucial role in selective memory, enabling the model to choose what information to retain and what information to filter out as noise. During each time step, information from the current time step, information from the current storage unit, and the output from the previous time step are fed into the output gate to generate the output for the current time step. When applied to signal denoising, DRNN can effectively filter out non-physiological information in the candidate signals, such as false peaks caused by light, camera, or motion. Simultaneously, it handles noise that closely matches the heart rate frequency and is difficult to filter, retaining only information related to the pulse wave. Therefore, the neural network model can achieve end-to-end filtering of irrelevant noise included in candidate signals while preserving waveform features.

[0094] In one possible implementation, the neural network model of this disclosure embodiment can be pre-trained or trained in an electronic device performing the method of this disclosure embodiment. The neural network model can be trained according to a preset training set, which includes multiple sample feature signals and corresponding labeled pulse signals. The training process can involve inputting each sample feature signal into the neural network model and outputting a corresponding predicted pulse signal. By calculating the predicted pulse signal strength of each sample feature signal and the Pearson correlation coefficient of the labeled pulse signal, the corresponding model loss is obtained, and the neural network model is adjusted based on the model loss.

[0095] Figure 2 This diagram illustrates a process for determining a training set during the training of a neural network model according to an embodiment of the present disclosure. Figure 2 As shown, the method for determining the sample feature signals in the training set for training the neural network model can be the same as the method for determining the skin feature signals in step S10. For example, multiple consecutive face images within the target time interval can be acquired first, and then a preset number of facial markers can be marked on the acquired face images. Then, based on these facial markers in the first frame face image, a bounding box can be defined on the cheek areas of the acquired face image. The patch size is determined based on the size of the bounding box in the first frame. Within the bounding box, N small blocks can be randomly divided, each block serving as a small region of interest (ROI). The original change trajectory of the N facial pixel values ​​is then calculated using the average spatial pixels in the green channel. A detrending algorithm is used to remove the signal baseline, and the resulting pixel trajectory is used as the sample feature signal. Furthermore, the contact pulse signal acquired by a finger pulse oximeter within the same time interval can be used as the labeled pulse signal corresponding to these sample feature signals.

[0096] Furthermore, during the training of the neural network model, the Pearson correlation coefficient can be used as the loss function. This loss function deviates from the more commonly used maximum absolute distance (MAD) but achieves better results. This is because the MAD can better preserve the amplitude characteristics of the pulse wave when the input sample is a single-cycle pulse wave. However, the focus of signal processing in this embodiment is the timing of the peaks and troughs in the pulse wave. Meanwhile, the signal input to the neural network model contains multiple pulse wave cycles. Therefore, using the Pearson correlation coefficient to measure the correlation between the input and output signals helps the neural network model better capture the unique waveform characteristics of the pulse wave, thereby improving the performance of the trained model.

[0097] Step S30: Calculate the biometric information of the target object based on the pulse wave signal.

[0098] In one possible implementation, after determining the pulse wave signal through signal processing, the electronic device can further calculate the biometric information of the target object based on the pulse wave signal. The biometric information can be at least one type, used to characterize at least one vital sign of the target object. In other words, each type of biometric information can characterize a corresponding vital sign of the target object. Optionally, the biometric information may include at least one of heart rate, blood pressure, and cardiac duration. Further, it may also include other vital signs such as blood oxygen saturation.

[0099] The calculation methods of electronic devices differ depending on the type of biometric information being calculated. When the biometric information includes heart rate, calculating the biometric information of a target object based on a pulse wave signal can include: determining the pulse wave components by performing independent component analysis (ICA) on the pulse wave signal, and then determining the power of each pulse wave component as the corresponding heart rate. Specifically, in this embodiment, when calculating heart rate based on a pulse wave signal, the heart rate can be determined using the peak-frequency method, which combines the peak-frequency method and the autocorrelation method. This peak-frequency method can use signal separation algorithms such as ICA to process the pulse wave signal. This process first calculates the peak power of each component, takes the component with the largest peak power as the pulse wave, and then uses the power corresponding to that component as the heart rate. Furthermore, a continuous heart rate variation curve and heart rate variability can be obtained.

[0100] In the process of determining heart rate using a combination of peak-frequency and autocorrelation methods to calculate pulse wave signals, the autocorrelation method calculates the cyclic autocorrelation function of the signal. This involves multiplying the time series at different lag lengths by the un-lagped time series to calculate similarity. The autocorrelation function value is maximized when the lag value is periodic. First, the peak-frequency method is used to extract candidate period values, and then autocorrelation is used for verification. If the candidate period provided by the peak-frequency method is at the peak of the autocorrelation function, it is considered a valid period; otherwise, if it is at the trough, it is considered a false alarm and discarded. Therefore, when using the peak-frequency method to determine heart rate, if the peak with the largest power value corresponds to a point in the autocorrelation graph that is at a peak-trough, there is no obvious periodicity in the time domain. In this case, the autocorrelation method can be used as a supplement, selecting a larger approximate peak value as the period, making the calculated heart rate closer to the true value. Furthermore, combining autocorrelation and the peak-frequency method allows for more accurate heart rate calculation and reduces noise interference.

[0101] Furthermore, when biometric information includes heartbeat intervals, the method for calculating the biometric information of a target object based on pulse wave signals can be to statistically analyze the peak values ​​of the pulse wave signals and calculate the time intervals between adjacent peaks, obtaining the corresponding interval sequence as the corresponding heartbeat interval. In other words, the electronic device can detect peak values ​​in the obtained pulse wave information, calculate the time intervals between adjacent peaks, which represent heartbeat intervals, and thus obtain a heartbeat-to-heartbeat interval sequence as the heartbeat interval.

[0102] After calculating the heartbeat interval, further quantitative analysis of heart rate variability can be performed. Specifically, linear methods can be used to analyze the pulsatile changes exhibited by the sinoatrial node, and more complex nonlinear analysis methods derived from nonlinear dynamics can be used, considering the influence of various nonlinear mechanisms such as the sympathetic and vagus nerves on heart rate regulation. These include Poincaré plots and detrended fluctuation analysis (DFA). Poincaré plots represent the dynamic rules governing the relationship between continuous sampled values ​​of dynamic variables and can be used to predict severe arrhythmias in patients with ventricular fibrillation. Poincaré treats the next IBI (interval between heartbeats) value as a mapping to the current IBI value, which can be quantitatively described using SD1 and SD2. SD1 is the standard deviation of the scatter plot projected onto the line y = -x, and SD2 is the standard deviation of the scatter plot projected onto the line y = x. Poincaré plots reveal information about the heart's spontaneous activity, with SD1 related to parasympathetic activity and SD2 related to sympathetic activity.

[0103] IBI N+1 =f(IBI) N )

[0104] The process of quantitative analysis using the Poincaré plot is as follows: The length of the IBI sequence (cardiac interval sequence) is denoted as N, meaning there are a total of N RR intervals. The integral RR interval time series is calculated by subtracting the average value of the entire IBI sequence from each IBI value (RRi). The integral time series of total length N is divided into equal segments, each with n RR intervals. For each segment, a linear fit is performed using the least squares method, denoted as yn(k). This least squares line is considered the trend of that segment. Then, the root mean square method is used to calculate the least squares line yn(k) and the volatility F(n) of the integral time series for each segment, thus obtaining the relationship between volatility and segment length n:

[0105]

[0106] Here, n can be taken from a preset range, such as from 4 to 100, and the slope of the line is denoted as the scaling exponent α. The scaling exponent is different in different time ranges. Therefore, a breakpoint can also be taken within the preset range, for example, a breakpoint of 13 in the preset range of 4-100, to calculate the short-term scaling exponent α1 for n from 4 to 13 and the long-term scaling exponent α2 for n from 13 to 100, as well as the overall scaling exponent αall at all time scales. The short-term scaling exponent is a commonly used predictor of mortality in patients with cardiovascular diseases such as myocardial infarction and angina pectoris, and it quantifies the fractal changes in cardiac dynamics.

[0107] Optionally, when the biometric information includes blood pressure, the process of calculating the biometric information of the target object based on the pulse wave signal can be as follows: The pulse wave signal corresponding to the skin of the target area is determined as the target pulse wave signal. Within a preset time interval, a reference pulse wave signal corresponding to the skin of the reference area of ​​the target object is acquired. Then, the pulse wave transmission time is determined based on the target pulse wave signal and the reference pulse wave signal. Finally, the blood pressure corresponding to the target object is calculated based on the pulse wave transmission time. The method for determining the reference pulse wave signal can be the same as that for the target pulse wave signal, i.e., it can be determined by executing steps S10 and S20. Optionally, there can be one or more reference pulse wave signals. For example, when the target object is a person and the target area skin is collected from the person's facial skin, the reference area skin corresponding to the reference pulse wave can be the skin of the palm or fingertips, etc. After determining the target pulse wave signal and the reference pulse wave signal, the time delay of the pulse wave peak reaching the target area skin and the reference area skin can be determined by calculating the time delay of the peak value of the target pulse wave signal and the reference area skin, thus obtaining the pulse wave transmission time.

[0108] Figure 3 A schematic diagram illustrating a process for determining pulse wave propagation time according to an embodiment of the present disclosure is shown. Figure 3 As shown, when the target area skin is the skin of a person's face and the reference area skin is the skin of a person's palm, the difference between the target pulse wave signal (which is the facial pulse wave) and the reference pulse wave signal (which is the palm pulse wave) can be used as the pulse wave transmission time (PTT) at the proximal and distal ends.

[0109] Furthermore, after determining the pulse wave transit time, the target subject's blood pressure can be calculated based on the pulse wave transit time. Specifically, a mathematical model representing the mapping relationship between pulse wave transit time and blood pressure can be obtained through mathematical modeling. Then, the pulse wave transit time is substituted into this mathematical model to estimate the target subject's blood pressure value. In this process, electronic devices can acquire pre-measured pulse transit times and cuff blood pressure from multiple subjects, and then define a mathematical model that correlates the pulse transit time and cuff blood pressure using a small number of unknown parameters. The unknown parameters of the mathematical model are determined by fitting the model to the pre-measured values.

[0110] For example, in this embodiment of the disclosure, the mathematical model can be simplified to a linear model BP = K1ln(PTT) + K2, where diastolic (the diastolic BP = m·PTT + b) The slope and y-intercept of the linear regression are used to calibrate coefficients K1 and K2. For animals, healthy adults, and patients, the parameters in the diastolic blood pressure calculation formula are m = -1.4, -0.5, -1.6 mmHg / ms, and b = 197, 164, 125 mmHg, respectively. PWV (pulse wave velocity) can also be used as a quantitative indicator to assess arterial elasticity and stiffness; a higher PWV indicates a stiffer artery. D represents the average arterial length in adults, with a normal reference value between 5-10 ms. Values ​​exceeding this range will trigger a warning to the user.

[0111] Figure 4 A comparative diagram showing the effects of different methods for determining pulse wave transmission time according to embodiments of the present disclosure is presented. Figure 4 As shown, the disclosed embodiments can also compare the pulse wave transit time determined by other methods with the pulse wave transit time determined by this method. The comparison results show that the pulse wave transit times obtained by different methods exhibit consistent trends. In other words, we can use the pulse wave transit time determined by collecting pulse wave signals from different locations to replace contact-type fingertip pulse oximeters or continuous cuffless blood pressure monitors for non-contact measurement. Furthermore, the corresponding blood pressure can be obtained based on the accurate pulse wave transit time.

[0112] Based on the aforementioned technical features, the biometric detection method of this disclosure can acquire image features of the skin in the target area of ​​a target object to determine the pulse wave signal, and further calculate biometric information based on the pulse wave signal, automatically and accurately detecting biometric information under non-contact conditions. Simultaneously, the entire detection process requires no manual intervention, has a high degree of automation, reduces the detection difficulty for users, and improves user experience.

[0113] Figure 5 A schematic diagram of a biometric detection device according to an embodiment of the present disclosure is shown. Figure 5 As shown, the biometric detection device in this embodiment may include:

[0114] The signal acquisition module 50 is used to acquire skin feature signals corresponding to the skin of the target area of ​​the target object within a preset time interval, wherein the skin feature signals characterize the changes in skin image feature values;

[0115] Signal processing module 51 is used to process the skin feature signal to obtain a pulse wave signal;

[0116] The feature determination module 52 is used to calculate the biometric information of the target object based on the pulse wave signal, wherein the biometric information is used to characterize at least one vital sign of the target object.

[0117] In one possible implementation, the signal acquisition module 50 is further configured to:

[0118] Within the preset time interval, skin images of the target area of ​​the target object are collected multiple times at a preset frequency;

[0119] By calculating the image feature value corresponding to each skin image, the corresponding skin feature signal is obtained.

[0120] In one possible implementation, the signal acquisition module 50 is further configured to:

[0121] Calculate the mean pixel value of the green channel in each skin image to obtain the corresponding feature pixels;

[0122] The original signal trajectory is determined based on the feature pixels corresponding to each skin image;

[0123] The original signal trajectory is subjected to baseline removal processing to obtain skin feature signals.

[0124] In one possible implementation, the signal processing module 51 is further configured to:

[0125] By separating the interference sources of the skin feature signals, candidate signals are obtained;

[0126] The candidate signal is processed by a trained neural network model to output a pulse wave signal.

[0127] In one possible implementation, the device further includes:

[0128] The model training module is used to train the neural network model according to a preset training set, wherein the training set includes multiple sample feature signals and corresponding labeled pulse signals;

[0129] The training process includes:

[0130] Each of the sample feature signals is input into the neural network model, and the corresponding predicted pulse signal is output.

[0131] The corresponding model loss is obtained by calculating the predicted pulse signal and the Pearson correlation coefficient of the labeled pulse signal for each of the sample feature signals.

[0132] Adjust the neural network model based on the model loss.

[0133] In one possible implementation, the biometric information includes at least one of heart rate, blood pressure, and heartbeat duration.

[0134] In one possible implementation, where the biometric information includes heart rate, the feature determination module 52 is further configured to:

[0135] The pulse wave components were determined by performing independent component analysis on the pulse wave signal.

[0136] The power of the pulse wave component is determined as the corresponding heart rate.

[0137] In one possible implementation, when the biometric information includes the period of heartbeat, the feature determination module 52 is further configured to:

[0138] The peak value of the pulse wave signal was statistically analyzed;

[0139] Calculate the time interval between adjacent peaks to obtain the corresponding interval sequence as the corresponding heartbeat period.

[0140] In one possible implementation, when the biometric information includes blood pressure, the feature determination module 52 is further configured to:

[0141] The pulse wave signal corresponding to the skin of the target area is identified as the target pulse wave signal;

[0142] Within a preset time interval, acquire the reference pulse wave signal corresponding to the reference area skin of the target object;

[0143] The pulse wave transmission time is determined based on the target pulse wave signal and the reference pulse wave signal;

[0144] The blood pressure corresponding to the target object is calculated based on the pulse wave transmission time.

[0145] In some embodiments, the functions or modules of the apparatus provided in this disclosure can be used to perform the methods described in the above method embodiments. The specific implementation can be referred to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0146] This disclosure also proposes a computer-readable storage medium storing computer program instructions that, when executed by a processor, implement the above-described method. The computer-readable storage medium can be volatile or non-volatile.

[0147] This disclosure also proposes an electronic device, including: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured to implement the above method when executing the instructions stored in the memory.

[0148] This disclosure also provides a computer program product, including computer-readable code, or a non-volatile computer-readable storage medium carrying computer-readable code, wherein when the computer-readable code is run in a processor of an electronic device, the processor in the electronic device performs the above-described method.

[0149] Figure 6 A schematic diagram of an electronic device 1900 according to an embodiment of the present disclosure is shown. For example, the electronic device 1900 may be provided as a server or a terminal device. (Refer to...) Figure 6 The electronic device 1900 includes a processing component 1922, which further includes one or more processors, and a memory resource represented by a memory 1932 for storing instructions, such as application programs, that can be executed by the processing component 1922. The application programs stored in the memory 1932 may include one or more modules, each corresponding to a set of instructions. Furthermore, the processing component 1922 is configured to execute instructions to perform the methods described above.

[0150] Electronic device 1900 may also include a power supply component 1926 configured to perform power management of electronic device 1900, a wired or wireless network interface 1950 configured to connect electronic device 1900 to a network, and an input / output interface 1958 (I / O interface). Electronic device 1900 can operate on an operating system, such as Windows Server, stored in memory 1932. TM Mac OS X TM Unix TM Linux TM FreeBSD TM Or similar.

[0151] In an exemplary embodiment, a non-volatile computer-readable storage medium is also provided, such as a memory 1932 including computer program instructions that can be executed by a processing component 1922 of an electronic device 1900 to perform the above-described method.

[0152] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0153] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination thereof. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0154] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0155] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0156] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0157] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0158] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0159] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0160] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or technical improvements to the embodiments in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A biometric detection method, characterized in that, The method includes: Within a preset time interval, skin feature signals corresponding to the target area skin of the target object are acquired, and the skin feature signals characterize the changes in skin image feature values. The skin feature signals are processed to obtain pulse wave signals; The biometric information of the target object is calculated based on the pulse wave signal, and the biometric information is used to characterize at least one vital sign of the target object.

2. The method according to claim 1, characterized in that, The step of acquiring skin feature signals corresponding to the target area skin of the target object within a preset time interval includes: Within the preset time interval, skin images of the target area of ​​the target object are collected multiple times at a preset frequency; By calculating the image feature value corresponding to each skin image, the corresponding skin feature signal is obtained.

3. The method according to claim 2, characterized in that, The step of calculating the corresponding skin feature value for each skin image to obtain the corresponding skin feature signal includes: Calculate the mean pixel value of the green channel in each skin image to obtain the corresponding feature pixels; The original signal trajectory is determined based on the feature pixels corresponding to each skin image; The original signal trajectory is subjected to baseline removal processing to obtain skin feature signals.

4. The method according to claim 1, characterized in that, The step of processing the skin feature signals to obtain pulse wave signals includes: By separating the interference sources of the skin feature signals, candidate signals are obtained; The candidate signal is processed by a trained neural network model to output a pulse wave signal.

5. The method according to claim 4, characterized in that, The method further includes: The neural network model is trained according to a preset training set, which includes multiple sample feature signals and corresponding labeled pulse signals. The training process includes: Each of the sample feature signals is input into the neural network model, and the corresponding predicted pulse signal is output. The corresponding model loss is obtained by calculating the predicted pulse signal and the Pearson correlation coefficient of the labeled pulse signal for each of the sample feature signals. Adjust the neural network model based on the model loss.

6. The method according to claim 1, characterized in that, The biometric information includes at least one of heart rate, blood pressure, and heart rate.

7. The method according to claim 6, characterized in that, When the biometric information includes heart rate, calculating the biometric information of the target object based on the pulse wave signal includes: The pulse wave components were determined by performing independent component analysis on the pulse wave signal. The power of the pulse wave component is determined as the corresponding heart rate.

8. The method according to claim 6, characterized in that, When the biometric information includes the heartbeat period, calculating the biometric information of the target object based on the pulse wave signal includes: The peak value of the pulse wave signal was statistically analyzed; Calculate the time interval between adjacent peaks to obtain the corresponding interval sequence as the corresponding heartbeat period.

9. The method according to claim 6, characterized in that, When the biometric information includes blood pressure, calculating the biometric information of the target object based on the pulse wave signal includes: The pulse wave signal corresponding to the skin of the target area is identified as the target pulse wave signal; Within a preset time interval, acquire the reference pulse wave signal corresponding to the reference area skin of the target object; The pulse wave transmission time is determined based on the target pulse wave signal and the reference pulse wave signal; The blood pressure corresponding to the target object is calculated based on the pulse wave transmission time.

10. A biometric detection device, characterized in that, The device includes: The signal acquisition module is used to acquire skin feature signals corresponding to the skin of the target area of ​​the target object within a preset time interval. The skin feature signals represent the changes in skin image feature values. The signal processing module is used to process the skin feature signals to obtain pulse wave signals; The feature determination module is used to calculate the biometric information of the target object based on the pulse wave signal, wherein the biometric information is used to characterize at least one vital sign of the target object.

11. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to implement the method of any one of claims 1 to 9 when executing instructions stored in the memory.

12. A non-volatile computer-readable storage medium storing computer program instructions thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1 to 9.

Images