Blood pressure measurement method and device based on multi-dimensional transformation neural network model

Abstract

The invention relates to a blood pressure measurement method and device based on a multi-dimensional transformation neural network model. The method comprises the steps that pulse wave signals of a person to be measured are obtained; the pulse wave signals are input into a multi-dimensional transformation neural network model, the blood pressure value is obtained, the multi-dimensional transformation neural network model comprises a dimension transformation block based on the extreme value, a dimension transformation block based on fast Fourier transformation and a one-dimensional deep residual shrinkage network, and the blood pressure value is obtained; the dimension transformation block based on the extreme value is used for converting the pulse wave signal into a first signal, and then the first signal is converted into a second signal through the dimension transformation block based on fast Fourier transform, so that a blood pressure value is obtained based on the second signal through the one-dimensional deep residual shrinkage network. The blood pressure can be quickly and accurately measured, and the continuous monitoring capability can be kept.

Description

Technical Field

[0001] The present disclosure relates to the technical field of medical equipment, and in particular to a blood pressure measurement method, device and electronic equipment based on a multi-dimensional transformation neural network model. Background Art

[0002] Hypertension is one of the most common chronic diseases worldwide. Long-term hypertension significantly increases the risk of heart disease, stroke and other cardiovascular diseases. Therefore, timely monitoring of blood pressure changes is crucial for the prevention and management of these diseases.

[0003] However, the data processing process in the related technology may cause certain errors in the final detection results. Summary of the invention

[0004] The blood pressure measurement method, device and electronic device based on the multi-dimensional transformation neural network model provided by the exemplary embodiments of the present disclosure can at least solve the above-mentioned technical problems and other technical problems not mentioned above.

[0005] According to one aspect of the present disclosure, a blood pressure measurement method based on a multidimensional transformation neural network model is provided, the method comprising: obtaining a pulse wave signal of a person to be measured; inputting the pulse wave signal into the multidimensional transformation neural network model to obtain a blood pressure value, wherein the multidimensional transformation neural network model comprises a dimensional transformation block based on an extreme value, a dimensional transformation block based on a fast Fourier transform, and a one-dimensional deep residual shrinkage network, the dimensional transformation block based on the extreme value is used to convert the pulse wave signal into a first signal, and then convert the first signal into a second signal through the dimensional transformation block based on the fast Fourier transform, so as to obtain the blood pressure value based on the second signal through the one-dimensional deep residual shrinkage network.

[0006] Optionally, the maximum value-based dimensional transformation block is used to perform the following operations: convert the pulse wave signal into a two-dimensional pulse wave signal based on the maximum value dimensional transformation; input the two-dimensional pulse wave signal into a residual network with the last layer of connection removed to obtain a first output; subject the first output to pooling processing to convert it into a one-dimensional pulse wave signal; and perform residual linking of the one-dimensional pulse wave signal and the pulse wave signal to obtain the first signal.

[0007] Optionally, the maximum value-based dimensional transformation converts the pulse wave signal into a two-dimensional pulse wave signal, including: identifying the maximum value and the minimum value in each cycle of the pulse wave signal; taking the maximum value or the minimum value in each cycle as the starting point, and intercepting a plurality of signal segments with a length of a first preset threshold from the pulse wave signal; and stacking each of the signal segments as a row of a two-dimensional matrix to obtain the two-dimensional pulse wave signal.

[0008] Optionally, the fast Fourier transform-based dimensionality transformation block is used to perform the following operations: normalize the first signal output to obtain a second output; perform multiple fast Fourier transform-based dimensionality transformations on the second output to obtain a third output; and residually link the third output with the pulse wave signal to obtain the second signal.

[0009] Optionally, obtaining the pulse wave signal of the subject to be tested includes: obtaining a photoelectric volumetric pulse wave signal of the subject to be tested; performing data cleaning on the photoelectric volumetric pulse wave signal to obtain a first pulse wave signal; extracting a second pulse wave signal that meets preset conditions by performing a quality check on the first pulse wave signal; wherein, inputting the pulse wave signal into the multidimensional transformation neural network model includes: inputting the second pulse wave signal into the multidimensional transformation neural network model.

[0010] Optionally, the data cleaning of the photoelectric volumetric pulse wave signal to obtain the first pulse wave signal includes: low-pass filtering the photoelectric volumetric pulse wave signal to obtain a third pulse wave signal; discrete wavelet transforming the third pulse wave signal and eliminating the low-frequency component coefficients of the first layer obtained by decomposition during the discrete wavelet transform process to obtain the first pulse wave signal.

[0011] Optionally, extracting the second pulse wave signal that meets preset conditions by performing a quality check on the first pulse wave signal includes: performing outlier detection on the pulse wave signal in each window in the first pulse wave signal based on kernel density estimation, wherein each window is obtained by pre-windowing the first pulse wave signal; and obtaining the second pulse wave signal that meets the preset conditions by extracting the pulse wave signal of the window in which no outliers exist in the first pulse wave signal.

[0012] Optionally, the outlier detection method based on kernel density estimation performs outlier detection on the pulse wave signal in each window of the first pulse wave signal, including: calculating the quality score of each data point in each window of the first pulse wave signal based on a kernel density estimation formula using a smooth and symmetric Gaussian kernel, wherein the quality score is the inverse of the kernel density estimation, which is used to determine whether the data point is an outlier, and the kernel density estimation formula is expressed as:

[0013]

[0014] Wherein, n is all data point samples in the first pulse wave signal, ω is the bandwidth parameter, and x is at least one of the following characteristics of the data points in each window in the first pulse wave signal: mean, variance, skewness, peak state, rising amplitude, dicrotic wave amplitude, descending isthmus amplitude, and area under the curve.

[0015] According to another aspect of the present disclosure, a blood pressure measurement device based on a multi-dimensional transformation neural network model is also provided, the device comprising: a signal acquisition unit, configured to: acquire a pulse wave signal of a person to be measured; a blood pressure calculation unit, configured to: input the pulse wave signal into the multi-dimensional transformation neural network model to obtain a blood pressure value, wherein the multi-dimensional transformation neural network model comprises a dimensional transformation block based on an extreme value, a dimensional transformation block based on a fast Fourier transform, and a one-dimensional deep residual shrinkage network, the dimensional transformation block based on the extreme value is used to convert the pulse wave signal into a first signal, and then convert the first signal into a second signal through the dimensional transformation block based on the fast Fourier transform, so as to obtain the blood pressure value based on the second signal through the one-dimensional deep residual shrinkage network.

[0016] Optionally, the blood pressure calculation unit is configured to: convert the pulse wave signal into a two-dimensional pulse wave signal based on the maximum dimensional transformation; input the two-dimensional pulse wave signal into a residual network with the last layer of connection removed to obtain a first output; subject the first output to pooling processing to convert it into a one-dimensional pulse wave signal; and perform residual linking of the one-dimensional pulse wave signal and the pulse wave signal to obtain the first signal.

[0017] Optionally, the blood pressure calculation unit is configured to: identify the maximum and minimum values ​​in each cycle of the pulse wave signal; take the maximum or minimum value in each cycle as a starting point, and extract multiple signal segments with a length of a first preset threshold from the pulse wave signal; and stack each of the signal segments as a row of a two-dimensional matrix to obtain the two-dimensional pulse wave signal.

[0018] Optionally, the blood pressure calculation unit is configured to: perform a normalization operation on the first signal output to obtain a second output; perform multiple fast Fourier transform-based dimensional transformations on the second output to obtain a third output; and perform a residual linking operation on the third output and the pulse wave signal to obtain the second signal.

[0019] Optionally, the signal acquisition unit is further configured to: acquire a photoelectric volumetric pulse wave signal of the subject to be tested; perform data cleaning on the photoelectric volumetric pulse wave signal to obtain a first pulse wave signal; extract a second pulse wave signal that meets preset conditions by performing a quality check on the first pulse wave signal; wherein the blood pressure calculation unit is further configured to: input the second pulse wave signal into the multidimensional transformation neural network model.

[0020] Optionally, the signal preprocessing unit is configured to: perform low-pass filtering on the photoelectric volumetric pulse wave signal to obtain a third pulse wave signal; perform discrete wavelet transform on the third pulse wave signal, and eliminate the low-frequency component coefficients of the first layer obtained by decomposition during the discrete wavelet transform process to obtain the first pulse wave signal.

[0021] Optionally, the signal preprocessing unit is configured to: perform outlier detection on the pulse wave signal in each window in the first pulse wave signal based on an outlier detection method based on kernel density estimation, wherein each window is obtained by pre-windowing the first pulse wave signal; and obtain the second pulse wave signal that meets the preset conditions by extracting the pulse wave signal of the window in which no outliers exist in the first pulse wave signal.

[0022] Optionally, the signal preprocessing unit is configured to: calculate the quality score of each data point in each window in the first pulse wave signal based on a kernel density estimation formula using a smooth and symmetric Gaussian kernel, wherein the quality score is the inverse of the kernel density estimation, which is used to determine whether the data point is an outlier, and the kernel density estimation formula is expressed as:

[0023]

[0024] Wherein, n is all data point samples in the first pulse wave signal, ω is the bandwidth parameter, and x is at least one of the following characteristics of the data points in each window in the first pulse wave signal: mean, variance, skewness, peak state, rising amplitude, dicrotic wave amplitude, descending isthmus amplitude, and area under the curve.

[0025] According to another aspect of an embodiment of the present disclosure, an electronic device is also provided, comprising: at least one processor; and at least one memory storing computer executable instructions, wherein when the computer executable instructions are executed by the at least one processor, the at least one processor is prompted to execute any of the blood pressure measurement methods based on the multi-dimensional transformation neural network model as described above.

[0026] According to another aspect of an embodiment of the present disclosure, a computer-readable storage medium storing instructions is also provided. When the instructions are executed by at least one processor, the at least one processor is prompted to execute any of the blood pressure measurement methods based on the multi-dimensional transformation neural network model as described above.

[0027] According to another aspect of an embodiment of the present disclosure, a system is also provided, comprising at least one computing device and at least one storage device storing instructions, wherein when the instructions are executed by the at least one computing device, the at least one computing device is prompted to execute any of the blood pressure measurement methods based on the multidimensional transformation neural network model as described above.

[0028] According to another aspect of an embodiment of the present disclosure, a computer program product is also provided, including a computer program / instruction, which, when executed by a processor, implements the blood pressure measurement method based on the multi-dimensional transformation neural network model as described in any one of the above.

[0029] The technical solution provided by the embodiments of the present disclosure brings at least the following beneficial effects:

[0030] According to the blood pressure measurement method, device and electronic device based on the multi-dimensional transformation neural network model disclosed in the present invention, by capturing the pulse wave signal of the subject and applying a series of dimensional conversion techniques and feature extraction techniques to process the pulse wave signal, not only the instantaneous fluctuations in the pulse wave signal can be captured, but also the overall trend of the pulse wave signal can be grasped, thereby achieving rapid and accurate measurement of blood pressure.

[0031] In addition, on the basis of obtaining the photoelectric volumetric pulse wave signal, a high-quality pulse wave signal is obtained through data processing methods such as cleaning and optimization, and then after signal quality screening, an effective pulse wave signal can be obtained. The effective pulse wave signal is subjected to a series of dimensional transformations, that is, the one-dimensional signal is converted into a variety of two-dimensional signals through multiple dimensional transformations, so that richer information can be extracted. Combined with feature extraction processing, it can further and more accurately capture the instantaneous fluctuations of the pulse wave signal and grasp the overall trend, thereby achieving faster and more accurate measurement of blood pressure. In addition, it can also maintain the ability of continuous monitoring, and can be applied to miniaturized, wearable blood pressure detection equipment, which is convenient for frequent use and improves the portability and comfort of the blood pressure measurement process. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The drawings herein are incorporated into and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the description are used to explain the principles of the present disclosure, and do not constitute improper limitations on the present disclosure.

[0033] Figure 1 A flow chart showing a blood pressure measurement method based on a multi-dimensional transformation neural network model in an exemplary embodiment of the present disclosure;

[0034] Figure 2 A schematic diagram showing the structure of a multi-dimensional transformation neural network model in an exemplary embodiment of the present disclosure;

[0035] Figure 3A schematic diagram showing a data processing flow of a dimension transformation block based on a maximum value in an exemplary embodiment of the present disclosure is shown;

[0036] Figure 4 A schematic diagram showing a data processing flow of a dimension transformation block based on fast Fourier transform in an exemplary embodiment of the present disclosure is shown;

[0037] Figure 5 A schematic diagram showing a data processing flow of a one-dimensional deep residual shrinkage network in an exemplary embodiment of the present disclosure is shown;

[0038] Figure 6 A schematic diagram showing a structure of a multi-dimensional transformation neural network model in an exemplary embodiment of the present disclosure is shown;

[0039] Figure 7 A block diagram of a blood pressure measurement device based on a multi-dimensional transformation neural network model in an exemplary embodiment of the present disclosure is shown;

[0040] Figure 8 A schematic diagram showing a structure of a blood pressure measurement device based on a multi-dimensional transformation neural network model in an exemplary embodiment of the present disclosure is shown;

[0041] Fig. 9 A block diagram illustrating an electronic device according to an exemplary embodiment of the present disclosure. DETAILED DESCRIPTION

[0042] In order to enable ordinary persons in the art to better understand the technical solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the accompanying drawings.

[0043] It should be noted that the terms "first", "second", etc. in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. The implementation methods described in the following examples do not represent all implementation methods consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the attached claims.

[0044] It should be noted that the phrase "at least one of the items" in the present disclosure includes three types of parallel situations: "any one of the items", "a combination of any number of the items", and "all of the items". For example, "including at least one of A and B" includes the following three parallel situations: (1) including A; (2) including B; (3) including A and B. Another example is "executing at least one of step 1 and step 2" which means the following three parallel situations: (1) executing step 1; (2) executing step 2; (3) executing step 1 and step 2.

[0045] In blood pressure monitoring technology, there are currently two main mainstream methods: pressure sensing-based method and photoelectric sensing-based method.

[0046] The pressure sensing method provides relatively accurate blood pressure readings by directly measuring the pressure changes in the arterial wall. However, this method has some obvious disadvantages, including discomfort when wearing, large size of the device making it difficult to carry, and difficulty in frequent use. These problems limit the widespread application of this method in daily life.

[0047] The photoelectric sensing method detects blood flow changes by optical means, thereby indirectly calculating blood pressure values. This method improves user comfort, portability, and the feasibility of continuous monitoring. However, this method also has some shortcomings, such as the data is easily affected by external interference (such as movement, ambient light, etc.), and the data processing process is relatively complicated, resulting in certain errors in the final detection results.

[0048] In order to solve one or more problems in the related art such as large blood pressure detection error, difficulty in carrying, discomfort, and difficulty in frequent use, the present disclosure provides a blood pressure measurement method, device and electronic device based on a multi-dimensional transformation neural network model. By capturing the pulse wave signal of the subject and applying a series of dimensional conversion techniques and feature extraction techniques to process the pulse wave signal, not only the instantaneous fluctuations in the pulse wave signal can be captured, but also the overall trend of the pulse wave signal can be grasped, thereby achieving rapid and accurate measurement of blood pressure.

[0049] Next, we will refer to Figures 1 to 9 The blood pressure measurement method, device and electronic device based on the multi-dimensional transformation neural network model disclosed in the present invention are specifically described.

[0050] Figure 1 A flow chart of a blood pressure measurement method based on a multi-dimensional transformation neural network model in an exemplary embodiment of the present disclosure is shown.

[0051] Reference Figure 1 In step 101, a pulse wave signal of a subject is obtained.

[0052] Specifically, the pulse wave signal of the subject to be measured may be obtained by ultrasonic Doppler method, impedance volume plethysmography method, or the like.

[0053] According to an exemplary embodiment of the present disclosure, a photoplethysmogram signal of a subject may be acquired.

[0054] Specifically, the photoelectric volumetric pulse wave signal can be collected by using a photoelectric module of photoplethysmography (PPG). The exemplary embodiments of the present disclosure are illustrated by taking the photoplethysmography signal as an example.

[0055] According to an exemplary embodiment of the present disclosure, the photoelectric volumetric pulse wave signal can be cleaned to obtain a first pulse wave signal (i.e., a high-quality pulse wave signal); by performing a quality check on the first pulse wave signal, a second pulse wave signal (i.e., a valid pulse wave signal) that meets preset conditions can be extracted.

[0056] Specifically, the photoplethysmographic signal can be cleaned by filtering, thresholding, statistics, machine learning, etc. The quality check of the first pulse wave signal can be achieved by time domain analysis, frequency domain analysis, statistical features, machine learning, etc.

[0057] According to an exemplary embodiment of the present disclosure, the process of data cleaning of the photoplethysmogram signal may include but is not limited to the following operations: the photoplethysmogram signal may be low-pass filtered to obtain a third pulse wave signal; the third pulse wave signal may be discretely transformed, and the low-frequency component coefficients of the first layer obtained by decomposition during the discrete wavelet transform process may be eliminated to obtain the first pulse wave signal.

[0058] Specifically, a 0-8 Hz low-pass filter based on a Butterworth template can be used to low-pass filter the photoplethysmogram signal, and the low-pass filter can strictly follow the design principle of the infinite impulse response (IIR) filter. The frequency domain formula of the Butterworth analog low-pass filter is as follows:

[0059]

[0060] Where N is the order, ω cis the 3dB cutoff frequency. After obtaining the S-domain expression H(s) of the analog low-pass filter, the impulse response invariance method can be used to realize the transformation from S-domain to Z-domain. The impulse response invariance method uses the time domain approximation principle to transform the S-domain expression H(s) inverse Laplace to obtain the continuous time domain expression h[t], sample the time domain to obtain the discrete time domain expression h[k], and finally use the Z transform to obtain the Butterworth digital low-pass filter Z-domain expression H(z). The Butterworth filter has the advantages of passband flatness, etc. It can filter out high-frequency interference to a large extent while maintaining the original waveform blood pressure information to obtain the third pulse wave signal.

[0061] Afterwards, the db8 wavelet function can be used to perform discrete wavelet transform on the third pulse wave signal, and the low-frequency component coefficients in the first layer obtained by decomposition during the discrete wavelet transform can be designed to eliminate the first pulse wave signal (i.e., high-quality pulse wave signal). Discrete wavelet transform can achieve high-frequency and low-frequency subdivision of the time series waveform by stretching and translating the wavelet function. In addition, discrete wavelet transform has the characteristic of concentrated energy and can effectively derive detail functions and approximate functions of the signal at different scales. Finally, the signal baseline drift can be effectively removed to obtain a high-quality pulse wave signal (i.e., the first pulse wave signal).

[0062] According to an exemplary embodiment of the present disclosure, the process of performing quality inspection on the first pulse wave signal may include but is not limited to the following operations: outlier detection may be performed on the pulse wave signal in each window in the first pulse wave signal based on the outlier detection method of kernel density estimation, wherein each window is obtained by pre-windowing the first pulse wave signal; and the second pulse wave signal that meets the preset condition is obtained by extracting the pulse wave signal of the window in which no outliers exist in the first pulse wave signal.

[0063] Specifically, the window size of the first pulse wave signal can be set to 3 to 5 seconds for windowing to perform related operations of signal quality inspection. In addition, the quality score can be obtained by the outlier detection method based on kernel density estimation to identify outliers. The preset condition can be that there are no outliers in the pulse wave signal of any window. Therefore, the pulse wave signal of each window without outliers can be extracted to obtain a valid pulse wave signal (i.e., the second pulse wave signal) to achieve quality inspection of the first pulse wave signal.

[0064] In addition, the pulse wave signal of each window with outliers can be extracted to obtain invalid pulse wave signals.

[0065] According to an exemplary embodiment of the present disclosure, the outlier detection method based on kernel density estimation is used to perform outlier detection on the pulse wave signal in each window of the first pulse wave signal, which may include but is not limited to the following operations: the quality score of each data point in each window of the first pulse wave signal may be calculated based on the kernel density estimation formula using a smooth and symmetric Gaussian kernel, wherein the quality score is the inverse of the kernel density estimation, which is used to determine whether the data point is an outlier. The kernel density estimation formula is expressed as:

[0066]

[0067] Among them, n is all data point samples in the first pulse wave signal, ω is the bandwidth parameter, and x is at least one of the following characteristics of the data points in each window in the first pulse wave signal: mean, variance, skewness, peak state, rising amplitude, dicrotic wave amplitude, descending isthmus amplitude, and area under the curve.

[0068] Specifically, in the process of signal quality inspection, the outlier detection method based on kernel density estimation can be used, and the PPG signal mean, variance, skewness, peak state, rising amplitude, dicrotic wave amplitude, descending isthmus amplitude, and area under the curve are used as the eight data dimensions of outlier detection to determine whether each cycle in the signal window is normalized. The kernel density estimation formula can be as follows:

[0069]

[0070] Where n is all data samples, ω is the bandwidth parameter, and the formula is essentially the relative distance calculation between a certain signal window and all other signal windows. The key to the kernel density estimation method lies in the selection of kernel function. In the present disclosure, a smooth and symmetrical Gaussian kernel can be selected, and the Gaussian kernel formula can be as follows:

[0071]

[0072] Substituting the Gaussian kernel into the kernel density estimation formula yields:

[0073]

[0074] The quality score can be set as the inverse of the kernel density estimate to determine whether the data point is in a low-density area, thereby identifying outliers.

[0075] In step 102, the pulse wave signal is input into a multi-dimensional transformation neural network model to obtain a blood pressure value, wherein the multi-dimensional transformation neural network model includes a dimensional transformation block based on an extreme value, a dimensional transformation block based on a fast Fourier transform, and a one-dimensional deep residual shrinkage network. The dimensional transformation block based on the extreme value is used to convert the pulse wave signal into a first signal, and then the first signal is converted into a second signal through the dimensional transformation block based on the fast Fourier transform, so as to obtain the blood pressure value based on the second signal through the one-dimensional deep residual shrinkage network.

[0076] Specifically, the blood pressure estimation process can be implemented by a blood pressure estimation model (ie, a multi-dimensional transformation neural network model).

[0077] Figure 2 A schematic diagram of the structure of a multi-dimensional transformation neural network model in an exemplary embodiment of the present disclosure is shown.

[0078] Reference Figure 2 The multi-dimensional transformation neural network model mainly consists of three main parts: a maximum value-based dimensional transformation block 201 (MMBlock), a fast Fourier transform-based dimensional transformation 202 (FFTBlock), and a one-dimensional deep residual shrinkage network 203 (DRSNs1D).

[0079] Here, the photoelectric volumetric pulse wave signal or the first pulse wave signal after data cleaning (high-quality pulse wave signal) or the second pulse wave signal after data cleaning and quality inspection (valid pulse wave signal) can be input into the multi-dimensional transformation neural network model, and finally the corresponding blood pressure value is output. According to an exemplary embodiment of the present disclosure, the dimension transformation block based on the maximum value can be used to perform operations including but not limited to the following: converting the pulse wave signal into a two-dimensional pulse wave signal based on the maximum value dimension transformation; inputting the two-dimensional pulse wave signal into the residual network with the last layer of connection removed to obtain a first output; subjecting the first output to pooling processing to convert it into a one-dimensional pulse wave signal; residually linking the one-dimensional pulse wave signal and the pulse wave signal to obtain a first signal.

[0080] Specifically, in the first stage, the input photoelectric volumetric pulse wave signal (or high-quality pulse wave signal, or effective pulse wave signal) can be first converted into two-dimensional data (i.e., two-dimensional pulse wave signal) through a maximum-value-based dimensional transformation, and then, these data can be sent to a modified ResNet model (residual network), which removes the connection of the last layer to obtain a first output while keeping the shape of the two-dimensional data (two-dimensional pulse wave signal) unchanged; after the first output is subjected to a layer of pooling processing, the two-dimensional data (two-dimensional pulse wave signal) can be flattened into one dimension (one-dimensional pulse wave signal), and a residual structure is formed with the original input (photoelectric volumetric pulse wave signal, or high-quality pulse wave signal, or effective pulse wave signal) to obtain a first signal to enhance feature expression.

[0081] It can be understood that the original input mentioned in the present disclosure refers to the input of the multi-dimensional transformation neural network model.

[0082] According to an exemplary embodiment of the present disclosure, converting a pulse wave signal into a two-dimensional pulse wave signal based on a dimensional transformation of the maximum value may include but is not limited to the following operations: identifying the maximum value and the minimum value in each cycle of the pulse wave signal; taking the maximum value or the minimum value in each cycle as the starting point, and intercepting a plurality of signal segments with a length of a first preset threshold from the pulse wave signal; and stacking each signal segment as a row of a two-dimensional matrix to obtain a two-dimensional pulse wave signal.

[0083] Specifically, Figure 3 A schematic diagram of a data processing flow of a dimension transformation block based on a maximum value in an exemplary embodiment of the present disclosure is shown.

[0084] Reference Figure 3 Based on the maximum value dimensional transformation block (MMBlock), the maximum value points of high-quality pulse wave signals (or photoelectric volume pulse wave signals, or effective pulse wave signals, the following takes high-quality pulse wave signals as an example) can be identified and arranged alternately, that is, X mm ={X max1 ,X min1 ,X max2 ,X min2 ,X max3 ,X min3 ,...}. Before the dimension transformation, the high-quality pulse wave signal X 1D Then, a sliding window length of L (first preset threshold) can be set to be selected from the X of the high-quality pulse wave signal. mm Start by intercepting a high-quality pulse wave signal X with a length of L 1dThe fragment is regarded as a row of two-dimensional data (two-dimensional pulse wave signal). After that, each row can be stacked up to form a two-dimensional matrix to obtain a two-dimensional pulse wave signal. The dimension transformation process can be described by the following formula:

[0085] X 2D =append{slide(X 1D ),size=L,head=X mm}

[0086] Since the number of extreme points of the data is different, the matrix as a whole can be expanded by n rows by zero padding. Subsequently, this two-dimensional matrix can be fed into a two-dimensional deep neural network for training. For example, the two-dimensional deep neural network can be a ResNet model. Finally, the two-dimensional model (two-dimensional pulse wave signal) can be converted back to one dimension (one-dimensional pulse wave signal), and the one-dimensional pulse wave signal and the original input (photoelectric volume accumulation pulse wave signal, or high-quality pulse wave signal, or effective pulse wave signal) can be used to establish a residual structure, and finally form a first signal, which is also the first output.

[0087] According to an exemplary embodiment of the present disclosure, a fast Fourier transform-based dimensionality transformation block can be used to perform operations including but not limited to the following: normalizing the first signal output to obtain a second output; performing multiple fast Fourier transform-based dimensionality transformations on the second output to obtain a third output; and residually linking the third output with the pulse wave signal to obtain a second signal.

[0088] Specifically, after the first stage one-dimensional data normalization operation is performed to obtain the second output, multiple fast Fourier transform-based dimensional transformations (FFTBlock) may be performed on the second output to obtain the third output.

[0089] Unlike MMBlock, FFTBlock does not rely on the obvious period of the signal and can be reused to deepen the network structure and learn richer features. Therefore, the signal is processed by FFTblock and combined with the original input using residual connection. In order to fully extract features, the multi-dimensional transform neural network model can reuse FFTBlock multiple times, for example, three times.

[0090] Specifically, Figure 4 A schematic diagram showing a data processing flow of a dimension transformation block based on fast Fourier transform in an exemplary embodiment of the present disclosure is shown.

[0091] Reference Figure 4, the fast Fourier transform (FFT) can be introduced as the basis for dividing the data, and the dimension transformation block based on the fast Fourier transform (FFTBlock) can be responsible for accessing the first signal. At this time, the step of normalizing the first signal output can be integrated into the FFTBlock. By performing FFT analysis on the accessed signal, the amplitude A={A f1,..., A fk} of the first k prominent frequencies f={f1,...,f k}, based on these frequencies, the one-dimensional data can be decomposed k times, each decomposition length is f, and the insufficient length is filled with zeros, and these fragments can be recombined to form k groups of two-dimensional data. The process can be shown in the following formula:

[0092] A=Amp(FFT(X 1D )),{f1,f2,...,f k} = argTopk(A)

[0093]

[0094] Next, the signal features can be extracted using a two-dimensional convolution kernel, which can capture both the dependencies within a time period and the dependencies between adjacent time periods in the signal. In this process, the model can use four convolution kernels of different sizes (1x1, 3x3, 5x5, and 7x7) (Inception Block) for convolution operations, and then the two-dimensional data can be converted back to one-dimensional form through the Flatten layer. In order to show the differences between different time periods, a set of programmable coefficients can be introduced in the adaptive aggregation layer, and combined with the softmax function to determine the weighted sum of k one-dimensional data fragments. The length of this weighted sum can be equal to k.

[0095] Finally, the k weighted one-dimensional data can be combined with the original input to form a residual structure, and finally form a second signal. Here, k can be 3 to 5.

[0096] According to an exemplary embodiment of the present disclosure, in the third stage, a one-dimensional deep residual shrinkage network (DRSNs1D) can be used to convert the one-dimensional data second signal into a blood pressure value. Specifically, the DRSNs (deep residual shrinkage network) modified to adapt to one-dimensional data, that is, the one-dimensional deep residual shrinkage network (DRSNs1D), can be better than the two-dimensional version in effect. After operations such as convolution, soft thresholding, and pooling, the model can finally compress the extracted information to obtain the label value of the blood pressure. For high pressure and low pressure, the same model structure can be used to establish a personalized model respectively.

[0097] Specifically, Figure 5 A schematic diagram showing the data processing flow of a one-dimensional deep residual shrinkage network in an exemplary embodiment of the present disclosure is shown.

[0098] Reference Figure 5 The one-dimensional deep residual contraction network includes key components such as batch normalization (BN), rectified linear unit (ReLU), convolution operation (Conv), global average pooling (GAP) and fully connected layer (FC), as well as absolute value operation (Absolute), average operation (Average), activation function (ReLU, Sigmoid), soft threshold (Softthresholding), back propagation (BP) and other components. The one-dimensional deep residual contraction network can be responsible for accessing the second signal and outputting the final blood pressure value. In addition, DRSNs1D can introduce soft thresholding as a nonlinear transformation layer. The threshold used can be automatically determined in the deep architecture. The function of the soft threshold can be expressed as:

[0099]

[0100] Among them, x is the input feature, y is the output feature, and τ is the threshold, which is a positive parameter. The soft threshold method does not set the negative features in the ReLU activation function to zero, but sets the features close to zero to zero, thereby retaining useful negative features. This can effectively prevent the gradient vanishing and exploding problems.

[0101] Figure 6 A schematic diagram showing one of the structures of a multi-dimensional transformation neural network model in an exemplary embodiment of the present disclosure is shown.

[0102] Reference Figure 6 The multi-dimensional transformation neural network model can include a maximum value-based dimensional transformation block (MMblock), three fast Fourier transform-based dimensional transformation blocks (FFTBlock), and a one-dimensional deep residual shrinkage network (DRSNs). The input can be a valid pulse wave signal that has been cleaned and quality checked. After the input is processed by the maximum value-based dimensional transformation block, three fast Fourier transform-based dimensional transformation blocks, and a one-dimensional deep residual shrinkage network in sequence, the output, i.e., the blood pressure value, can be obtained.

[0103] According to an exemplary embodiment of the present disclosure, a blood pressure monitoring method integrating a multi-dimensional conversion model is proposed. By capturing the optical volume pulse signal of the subject, optimizing it into a high-quality pulse wave signal and filtering the signal quality, a series of dimensional conversion technologies are used to process the pulse wave signal, which can not only capture instantaneous fluctuations but also grasp the overall trend, thereby achieving rapid and accurate measurement of blood pressure and maintaining the ability of continuous monitoring. In addition, the blood pressure monitoring method can be applied to portable devices.

[0104] Figure 7 A block diagram of a blood pressure measurement device based on a multi-dimensional transformation neural network model in an exemplary embodiment of the present disclosure is shown.

[0105] Reference Figure 7 An exemplary embodiment of the present disclosure further provides a blood pressure measurement device 700 based on a multi-dimensional transformation neural network model, which may include but is not limited to a signal acquisition unit 701 and a blood pressure calculation unit 702.

[0106] The signal acquisition unit 701 can acquire the pulse wave signal of the subject.

[0107] The blood pressure calculation unit 702 can input the pulse wave signal into a multi-dimensional transformation neural network model to obtain a blood pressure value, wherein the multi-dimensional transformation neural network model includes a dimensional transformation block based on the maximum value, a dimensional transformation block based on the fast Fourier transform, and a one-dimensional deep residual shrinkage network. The dimensional transformation block based on the maximum value is used to convert the pulse wave signal into a first signal, and then convert the first signal into a second signal through the dimensional transformation block based on the fast Fourier transform, so as to obtain the blood pressure value based on the second signal through the one-dimensional deep residual shrinkage network.

[0108] According to an exemplary embodiment of the present disclosure, the blood pressure calculation unit 702 can convert the pulse wave signal into a two-dimensional pulse wave signal based on the maximum dimensional transformation; input the two-dimensional pulse wave signal into a residual network with the last layer of connection removed to obtain a first output; subject the first output to pooling processing to convert it into a one-dimensional pulse wave signal; and perform residual linking of the one-dimensional pulse wave signal and the pulse wave signal to obtain a first signal.

[0109] According to an exemplary embodiment of the present disclosure, the blood pressure calculation unit 702 can identify the maximum and minimum values ​​in each cycle of the pulse wave signal; take the maximum or minimum value in each cycle as the starting point, and extract multiple signal segments with a length of a first preset threshold from the pulse wave signal; stack each signal segment as a row of a two-dimensional matrix to obtain a two-dimensional pulse wave signal.

[0110] According to an exemplary embodiment of the present disclosure, the blood pressure calculation unit 702 can perform a normalization operation on the first signal output to obtain a second output; perform multiple fast Fourier transform-based dimensional transformations on the second output to obtain a third output; and perform a residual linking operation on the third output and the pulse wave signal to obtain a second signal.

[0111] According to an exemplary embodiment of the present disclosure, the signal acquisition unit 701 can also acquire the photoelectric volumetric pulse wave signal of the subject to be tested; perform data cleaning on the photoelectric volumetric pulse wave signal to obtain a first pulse wave signal; and extract a second pulse wave signal that meets preset conditions by performing a quality check on the first pulse wave signal; wherein the blood pressure calculation unit 702 can also input the second pulse wave signal into the multidimensional transformation neural network model.

[0112] According to an exemplary embodiment of the present disclosure, the signal acquisition unit 701 can also perform low-pass filtering on the photoelectric volumetric pulse wave signal to obtain a third pulse wave signal; perform discrete wavelet transform on the third pulse wave signal, and eliminate the low-frequency component coefficients of the first layer obtained by decomposition during the discrete wavelet transform process to obtain the first pulse wave signal.

[0113] According to an exemplary embodiment of the present disclosure, the signal acquisition unit 701 can also perform outlier detection on the pulse wave signal in each window in the first pulse wave signal based on the outlier detection method of kernel density estimation, wherein each window is obtained by pre-windowing the first pulse wave signal; and a second pulse wave signal that meets preset conditions is obtained by extracting the pulse wave signal of the window in which no outliers exist in the first pulse wave signal.

[0114] According to an exemplary embodiment of the present disclosure, the signal preprocessing unit can calculate the quality score of each data point in each window in the first pulse wave signal based on the kernel density estimation formula using a smooth and symmetric Gaussian kernel, wherein the quality score is the inverse of the kernel density estimation, which is used to determine whether the data point is an outlier. The kernel density estimation formula is expressed as:

[0115]

[0116] Wherein, n is all data point samples in the first pulse wave signal, ω is the bandwidth parameter, and x is at least one of the following characteristics of the data points in each window in the first pulse wave signal: mean, variance, skewness, peak state, rising amplitude, dicrotic wave amplitude, descending isthmus amplitude, and area under the curve.

[0117] It can be understood that in the exemplary embodiment of the blood pressure measurement device 700 based on the multi-dimensional transformation neural network model, the specific implementation process is roughly the same as the exemplary embodiment of the blood pressure measurement method based on the multi-dimensional transformation neural network model, which will not be repeated here. The blood pressure measurement device 700 based on the multi-dimensional transformation neural network model can be configured as software, hardware, firmware or any combination of the above items to perform specific functions. For example, these devices may correspond to dedicated integrated circuits, pure software codes, or modules that combine software and hardware. In addition, one or more functions implemented by these devices can also be uniformly executed by components in physical entity devices (for example, processors, clients or servers, etc.).

[0118] According to an exemplary embodiment of the present disclosure, Figure 8 A schematic diagram of a blood pressure measurement device based on a multi-dimensional transformation neural network model in an exemplary embodiment of the present disclosure is shown. Figure 8 The device may include a data collection unit 801, a signal preprocessing unit 802, a signal quality inspection unit 803, a blood pressure calculation unit 804 and an information storage unit 805 connected in sequence.

[0119] The data collection unit 801 may be used to obtain the photoplethysmogram signal of the person to be tested, and transmit the photoplethysmogram signal to the signal preprocessing unit 802 .

[0120] The signal preprocessing unit 802 can be used to perform data cleaning on the photoplethysmography signal to obtain a high-quality pulse wave signal, and transmit the high-quality pulse wave signal to the signal quality inspection unit 803 .

[0121] The signal quality inspection unit 803 may be configured to perform a signal quality inspection on the high-quality pulse wave signal, and transmit the high-quality pulse wave signal that meets the inspection conditions to the blood pressure calculation unit 804 .

[0122] The blood pressure calculation unit 804 can be used to analyze the features and data correlation in the high-quality pulse wave signal, perform blood pressure detection on the person to be tested, and transmit the high-quality pulse wave signal and the blood pressure calculation result to the information storage unit 805.

[0123] The information storage unit 805 can be used to store high-quality pulse wave signals and blood pressure calculation results in a storage device.

[0124] The data collection unit 801 can be constructed based on a photoelectric module for acquiring a photoplethysmogram signal. The photoelectric module can include a photoelectric sensing device and a light source device, the former can be a photoelectric sensor, and the latter can be a device capable of emitting green light.

[0125] After capturing these PPG signals, the data collection unit 801 may pass them to the connected signal preprocessing unit 802. The signal preprocessing unit 802 may receive the PPG signals from the data collection unit 801 and preprocess them to ensure that the signals meet the requirements of the subsequent blood pressure estimation algorithm (i.e., the multi-dimensional transformation neural network model) for input data, thereby generating a high-quality pulse wave signal.

[0126] Fig. 9 A block diagram illustrating an electronic device according to an exemplary embodiment of the present disclosure.

[0127] Reference Fig. 9 The electronic device 900 includes at least one memory 901 and at least one processor 902, wherein the at least one memory 901 stores a set of computer executable instructions. When the computer executable instruction set is executed by the at least one processor 902, a blood pressure measurement method based on a multi-dimensional transformation neural network model according to an exemplary embodiment of the present disclosure is executed.

[0128] As an example, the electronic device 900 may be a PC, a tablet device, a personal digital assistant, a smart phone, or other device capable of executing the above-mentioned instruction set. Here, the electronic device 900 is not necessarily a single electronic device, but may also be any device or circuit capable of executing the above-mentioned instructions (or instruction sets) individually or in combination. The electronic device 900 may also be part of an integrated control system or system manager, or may be configured as a portable electronic device interconnected with a local or remote (e.g., via wireless transmission) interface.

[0129] In the electronic device 900, the processor 902 may include a central processing unit (CPU), a graphics processing unit (GPU), a programmable logic device, a dedicated processor system, a microcontroller or a microprocessor. As an example and not limitation, the processor may also include an analog processor, a digital processor, a microprocessor, a multi-core processor, a processor array, a network processor, etc.

[0130] The processor 902 may execute instructions or codes stored in the memory 901, wherein the memory 901 may also store data. Instructions and data may also be sent and received over a network via a network interface device, wherein the network interface device may employ any known transmission protocol.

[0131] The memory 901 may be integrated with the processor 902, for example, by placing RAM or flash memory within an integrated circuit microprocessor or the like. In addition, the memory 901 may include a separate device, such as an external disk drive, a storage array, or any other storage device that can be used by a database system. The memory 901 and the processor 902 may be operatively coupled, or may communicate with each other, such as through an I / O port, a network connection, etc., so that the processor 902 can read files stored in the memory.

[0132] In addition, the electronic device 900 may further include a video display (such as a liquid crystal display) and a user interaction interface (such as a keyboard, a mouse, a touch input device, etc.) All components of the electronic device 900 may be connected to each other via a bus and / or a network.

[0133] According to an exemplary embodiment of the present disclosure, a computer-readable storage medium storing instructions may also be provided, wherein, when the instructions are executed by at least one computing device, the at least one computing device is prompted to execute the above-mentioned blood pressure measurement method based on the multi-dimensional transformation neural network model.

[0134] Examples of computer-readable storage media include read-only memory (ROM), random-access programmable read-only memory (PROM), electrically erasable programmable read-only memory (EEPROM), random-access memory (RAM), dynamic random-access memory (DRAM), static random-access memory (SRAM), flash memory, nonvolatile memory, CD-ROM, CD-R, CD+R, CD-RW, CD+RW, DVD-ROM, DVD-R, DVD+R, DVD-RW, DVD+RW, DVD-RAM, BD-ROM, BD-R, BD-R LTH, BD-RE, Blu-ray or optical disk storage, hard disk drive (HDD), solid state drive (SSD), card storage (such as, multimedia card, secure digital (SD) card or extreme digital (XD) card), magnetic tape, floppy disk, magneto-optical data storage device, optical data storage device, hard disk, solid state disk and any other device, any other device is configured to store computer programs and any associated data, data files and data structures in a non-transitory manner and provide the computer programs and any associated data, data files and data structures to a processor or computer so that the processor or computer can execute the computer program. The computer program in the above-mentioned computer-readable storage medium can be run in an environment deployed in a computer device such as a client, a host, an agent device, a server, etc. In addition, in one example, the computer program and any associated data, data files and data structures are distributed on a networked computer system, so that the computer program and any associated data, data files and data structures are stored, accessed and executed in a distributed manner by one or more processors or computers. It should be noted that the instructions can also be used to execute additional steps in addition to the above steps or to perform more specific processing when executing the above steps. These additional steps and further processing contents have been mentioned in the description of the relevant methods, so they will not be repeated here to avoid repetition.

[0135] Another embodiment of the present disclosure relates to a system comprising at least one computing device and at least one storage device storing instructions, wherein the instructions, when executed by the at least one computing device, prompt the at least one computing device to execute the above-mentioned blood pressure measurement method based on the multi-dimensional transformation neural network model.

[0136] It should be noted that the system according to the exemplary embodiments of the present disclosure can completely rely on the execution of computer programs or instructions to realize corresponding functions, that is, each unit corresponds to each step in the functional architecture of the computer program, so that the entire system is called through a special software package (e.g., lib library) to realize the corresponding functions.

[0137] On the other hand, when the above-mentioned system is implemented in software, firmware, middleware or microcode, the program code or code segment for performing the corresponding operation can be stored in a computer-readable medium such as a storage medium, so that at least one processor or at least one computing device can perform the corresponding operation by reading and running the corresponding program code or code segment.

[0138] According to an exemplary embodiment of the present disclosure, the storage device may be integrated with the computing device, for example, RAM or flash memory is arranged within an integrated circuit microprocessor, etc. In addition, the storage device may include an independent device, such as an external disk drive, a storage array, or any other storage device that can be used by a database system. The storage device and the computing device may be operationally coupled, or may communicate with each other, such as through an I / O port, a network connection, etc., so that the computing device can read instructions stored in the storage device.

[0139] Another embodiment of the present disclosure relates to a computer program product, including a computer program / instruction, which, when executed by a processor, implements the blood pressure measurement method based on the multi-dimensional transformation neural network model as described in any one of the above.

[0140] According to the blood pressure measurement method, device and electronic device based on the multi-dimensional transformation neural network model provided by the present invention, by capturing the pulse wave signal of the subject and applying a series of dimensional conversion technologies and feature extraction technologies to process the pulse wave signal, not only the instantaneous fluctuations in the pulse wave signal can be captured, but also the overall trend of the pulse wave signal can be grasped, thereby achieving rapid and accurate measurement of blood pressure.

[0141] In addition, on the basis of the photoelectric volumetric pulse wave signal, after obtaining a high-quality pulse wave signal through data processing methods such as cleaning and optimization, and then after signal quality screening, an effective pulse wave signal can be obtained, and the effective pulse wave signal is subjected to a series of dimensional transformations, that is, the one-dimensional signal is converted into a variety of two-dimensional signals through multiple dimensional transformations, so that richer information can be extracted. Combined with feature extraction processing, it can further and more accurately capture the instantaneous fluctuations of the pulse wave signal and grasp the overall trend, thereby achieving faster and more accurate measurement of blood pressure. In addition, it can also maintain the ability of continuous monitoring, and can be applied to miniaturized, wearable blood pressure detection equipment, which is convenient for frequent use and improves the portability and comfort of the blood pressure measurement process.

[0142] The above describes various exemplary embodiments of the present disclosure, and it should be understood that the above description is only exemplary and not exhaustive, and the present disclosure is not limited to the disclosed exemplary embodiments. Without departing from the scope and spirit of the present disclosure, many modifications and changes are obvious to those of ordinary skill in the art. Therefore, the scope of protection of the present disclosure should be based on the scope of the claims.

Claims

1. A blood pressure measurement method based on a multi-dimensional transformation neural network model, characterized in that: The method comprises: Acquiring a pulse wave signal of a subject to be tested; The pulse wave signal is input into the multi-dimensional transformation neural network model to obtain a blood pressure value, wherein the multi-dimensional transformation neural network model includes a maximum value-based dimensional transformation block, a fast Fourier transform-based dimensional transformation block, and a one-dimensional deep residual shrinkage network, the maximum value-based dimensional transformation block is used to convert the pulse wave signal into a first signal, and then convert the first signal into a second signal through the fast Fourier transform-based dimensional transformation block, so as to obtain the blood pressure value based on the second signal through the one-dimensional deep residual shrinkage network.

2. The blood pressure measurement method based on the multi-dimensional transformation neural network model according to claim 1, characterized in that: The maximum value-based dimensionality transformation block is used to perform the following operations: Converting the pulse wave signal into a two-dimensional pulse wave signal based on the dimensional transformation of the maximum value; Inputting the two-dimensional pulse wave signal into a residual network with the last layer of connections removed to obtain a first output; The first output is subjected to pooling processing to be converted into a one-dimensional pulse wave signal; The one-dimensional pulse wave signal and the pulse wave signal are residually linked to obtain the first signal.

3. The blood pressure measurement method based on the multi-dimensional transformation neural network model as claimed in claim 2, characterized in that: The maximum value-based dimensional transformation converts the pulse wave signal into a two-dimensional pulse wave signal, including: Identifying the maximum and minimum values ​​in each cycle of the pulse wave signal; Taking the maximum value or the minimum value in each cycle as a starting point, extracting a plurality of signal segments having a length of a first preset threshold from the pulse wave signal; Each of the signal segments is stacked as a row of a two-dimensional matrix to obtain the two-dimensional pulse wave signal.

4. The blood pressure measurement method based on the multi-dimensional transformation neural network model according to claim 1, characterized in that: The fast Fourier transform based dimensionality transformation block is used to perform the following operations: Performing a normalization operation on the first signal output to obtain a second output; Performing multiple fast Fourier transform-based dimensional transformations on the second output to obtain a third output; The third output is residually linked with the pulse wave signal to obtain the second signal.

5. The blood pressure measurement method based on the multi-dimensional transformation neural network model according to claim 1, characterized in that: The step of obtaining the pulse wave signal of the subject to be measured comprises: obtaining a photoplethysmogram signal of the subject to be measured; Performing data cleaning on the photoplethysmographic signal to obtain a first pulse wave signal; By performing a quality check on the first pulse wave signal, a second pulse wave signal satisfying a preset condition is extracted; Wherein, the step of inputting the pulse wave signal into the multi-dimensional transformation neural network model comprises: The second pulse wave signal is input into the multi-dimensional transformation neural network model.

6. The blood pressure measurement method based on the multi-dimensional transformation neural network model according to claim 5, characterized in that: The step of performing data cleaning on the photoplethysmogram signal to obtain a first pulse wave signal includes: Low-pass filtering the photoplethysmogram signal to obtain a third pulse wave signal; The third pulse wave signal is subjected to discrete wavelet transform, and the low-frequency component coefficients of the first layer obtained by decomposition during the discrete wavelet transform are eliminated to obtain the first pulse wave signal.

7. The blood pressure measurement method based on the multi-dimensional transformation neural network model according to claim 5, characterized in that: The extracting a second pulse wave signal that meets a preset condition by performing a quality check on the first pulse wave signal comprises: Based on the outlier detection method of kernel density estimation, outlier detection is performed on the pulse wave signal in each window of the first pulse wave signal, wherein each window is obtained by pre-windowing the first pulse wave signal; The second pulse wave signal satisfying the preset condition is obtained by extracting the pulse wave signal of a window without outliers in the first pulse wave signal.

8. The blood pressure measurement method based on the multi-dimensional transformation neural network model according to claim 7, characterized in that: The outlier detection method based on kernel density estimation performs outlier detection on the pulse wave signal in each window of the first pulse wave signal, including: Based on the kernel density estimation formula using a smooth and symmetrical Gaussian kernel, the quality score of each data point in each window in the first pulse wave signal is calculated, wherein the quality score is the inverse of the kernel density estimation, which is used to determine whether the data point is an outlier. The kernel density estimation formula is expressed as: Wherein, n is all data point samples in the first pulse wave signal, ω is the bandwidth parameter, and x is at least one of the following characteristics of the data points in each window in the first pulse wave signal: mean, variance, skewness, peak state, rising amplitude, dicrotic wave amplitude, descending isthmus amplitude, and area under the curve.

9. A blood pressure measurement device based on a multi-dimensional transformation neural network model, characterized in that: The device comprises: The signal acquisition unit is configured to: acquire a pulse wave signal of a subject to be tested; The blood pressure calculation unit is configured to: input the pulse wave signal into the multi-dimensional transformation neural network model to obtain a blood pressure value, wherein the multi-dimensional transformation neural network model includes a maximum value-based dimensional transformation block, a fast Fourier transform-based dimensional transformation block, and a one-dimensional deep residual shrinkage network, the maximum value-based dimensional transformation block is used to convert the pulse wave signal into a first signal, and then convert the first signal into a second signal through the fast Fourier transform-based dimensional transformation block, so as to obtain the blood pressure value based on the second signal through the one-dimensional deep residual shrinkage network.

10. An electronic device, characterized in that: include: at least one processor; at least one memory storing computer executable instructions, Wherein, when the computer executable instructions are executed by the at least one processor, the at least one processor is prompted to execute the blood pressure measurement method based on the multi-dimensional transformation neural network model as described in any one of claims 1-8.

Images