Continuous pulse signal processing and cardiac function evaluation system

Abstract

The application provides a continuous pulse signal processing and heart function evaluation system, and comprises a data acquisition module, a preprocessing module, a signal evaluation module, a feature extraction module and a heart function evaluation module. The sample of the continuous pulse signal is acquired, the heart function of the sample is evaluated, the sample dataset is established, the noise threshold of the continuous pulse signal in the sample dataset is set for filtering and noise reduction processing, and normalization processing is performed to form a preprocessing dataset. According to the amplitude and power spectral density of the continuous pulse signal, the pulse quality score of each continuous pulse signal is calculated, the continuous pulse signal is screened according to the pulse quality score, the time domain feature and the frequency domain feature of the screened continuous pulse signal are extracted, the heart function evaluation model is established according to the neural network and is trained, the continuous pulse signal collected by the acquisition module and the age and gender are taken as the input of the evaluation model, and the evaluation of the heart function is completed.

Description

Technical Field

[0001] This invention relates to the field of cardiac function assessment technology, specifically to a continuous pulse signal processing and cardiac function assessment system. Background Technology

[0002] Cardiac disorders have become a leading cause of death worldwide, posing a serious threat to human health and constituting a major public health issue. With socio-economic development and significant changes in people's lifestyles, coupled with the accelerating pace of population aging and urbanization, unhealthy lifestyles are becoming increasingly prominent, leading to a continuous increase in the prevalence and incidence of cardiac disorders. Therefore, strengthening the monitoring and assessment of cardiac health status is both necessary and urgent.

[0003] Continuous pulse signal monitoring is a crucial technology in the monitoring and assessment of cardiac function. Traditional methods of cardiac function assessment rely on intermittent electrocardiograms and echocardiograms. While these methods provide important information about cardiac health, they cannot achieve long-term and continuous cardiac monitoring. How to analyze pathological information in continuous pulse signals to provide a more accurate and real-time assessment of functional health status has become an urgent problem to be solved.

[0004] In the prior art, CN106539562A discloses a method for assessing cardiac function by pulse wave, which includes the following steps: (1) acquiring the fingertip photoplethysmography (PPG) pulse wave signal of the human body using a photoplethysmography (PPG) sensor; (2) amplifying and filtering the fingertip PPG pulse wave signal; (3) separating the AC quantity; (4) performing analog-to-digital conversion on the AC quantity, resampling the converted digital signal to obtain the pulse wave waveform; (5) determining the pulse rate (PR) using the feature points on the pulse wave waveform; (6) segmenting the signal using the starting point of the ascending limb of the pulse wave waveform, and finally obtaining the waveform feature parameters reflecting the cardiovascular resistance and microvascular microcirculation status of the human body, and then obtaining the parameters reflecting cardiac function: cardiac output, cardiac output index, stroke volume, and heart rate index. However, although this method amplifies and filters the pulse signal, its operation is too simple and cannot effectively eliminate signal quality problems caused by the environment and equipment during signal acquisition. At the same time, it uses the characteristics on the pulse waveform to analyze cardiac function, but it cannot capture the detailed maintenance details in the pulse signal, which may cause bias in the judgment of cardiac function. Furthermore, it does not have the ability to automatically learn and adapt to the data characteristics of different patients.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide a continuous pulse signal processing and cardiac function assessment system to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] A continuous pulse signal processing and cardiac function assessment system includes:

[0009] The data acquisition module acquires continuous pulse signal samples from samplers of different ages and genders, evaluates the cardiac function of each person, establishes a sample dataset, and maps the continuous pulse signal, age, gender and timestamp in the sample dataset one by one. At the same time, it acquires the continuous pulse signal and age and gender information of the person to be evaluated.

[0010] The preprocessing module analyzes the noise characteristics of the continuous pulse signal in the sample dataset, sets a noise threshold for filtering and noise reduction, filters out the noise caused by the device acquisition, and then normalizes the noise-reduced continuous pulse signal to form a preprocessed dataset.

[0011] The signal evaluation module is used to obtain the amplitude and power spectral density of the continuous pulse signal in the preprocessed dataset, calculate the average value and variance of the amplitude of the continuous pulse signal and the median of the low-frequency ratio, calculate the pulse quality score of each continuous pulse signal, filter the continuous pulse signals according to the pulse quality score, and use the filtered continuous pulse signals as the analysis dataset.

[0012] The feature extraction module is used to perform time-domain feature analysis on the continuous pulse signal in the analysis dataset, and then extract the frequency-domain features of the pulse signal through fast Fourier transform, and form a feature dataset by combining the time-domain features and the frequency-domain features.

[0013] The cardiac function assessment module is used to establish a cardiac function assessment model through a neural network. The model is trained by using a feature dataset of continuous pulse signals, along with age and gender, as inputs and cardiac function assessment results as outputs. The pulse signals to be assessed, along with age and gender, collected by the acquisition module, are input into the trained assessment model to complete the assessment of cardiac function.

[0014] Furthermore, the specific steps for filtering and noise reduction by setting a noise threshold are as follows:

[0015] First, wavelet transform is performed on the continuous pulse signal to decompose it into a series of approximation coefficients and detail coefficients of different frequencies;

[0016] ,

[0017] in, For continuous pulse signals in the sample dataset, These are approximate coefficients. For detail coefficients, For scaling function, For wavelet functions, These are the acquisition time points for continuous pulse signals. These are the scale parameter and the displacement parameter, respectively.

[0018] By setting a noise threshold, the wavelet coefficients are processed to remove the detail coefficients affected by noise.

[0019] ,

[0020] in, The detail coefficient after noise reduction. The noise threshold is set.

[0021] The denoised signal is reconstructed by performing inverse wavelet transform using the processed wavelet coefficients.

[0022] ,

[0023] in, This represents the continuous pulse signal of the denoised sample.

[0024] Furthermore, the calculation formula for normalizing the noise-reduced continuous pulse signal is as follows:

[0025] ,

[0026] in, The pulse signal after normalization. This is the denoised pulse signal. The minimum value in the original pulse signal data. This is the maximum value in the original pulse signal data.

[0027] Furthermore, the formulas for calculating the average value and variance of the amplitude of the continuous pulse signal are as follows;

[0028] ,

[0029] ,

[0030] in, For continuous pulse signals, the first The amplitude of the main peak wave, The number of main peak waves in a continuous pulse signal. The average amplitude of the main peak wave of the continuous pulse signal. The variance of the amplitude of the main peak wave of the continuous pulse signal;

[0031] The formula for calculating the median low-to-high frequency ratio is:

[0032] ,

[0033] in, The median of the low-to-high frequency ratio, For frequency The power spectral density at that point These represent the frequency range of the continuous pulse signal. These represent the low-frequency range of the continuous pulse signal;

[0034] The formula for calculating the continuous pulse quality score is as follows:

[0035] ,

[0036] in, For continuous pulse quality scoring, These represent the weights of the average amplitude, variance, and median low-to-high frequency ratio of the continuous pulse signal in relation to the quality score.

[0037] Furthermore, the process of time-domain feature analysis is as follows: by periodically segmenting the continuous pulse signal, the time-domain features of the pulse signal in each period are obtained. A window of constant width is set according to the length and characteristics of the pulse signal. The new sampling point is placed at the highest position of the window. The average value and standard deviation of the continuous pulse signal period are calculated. The segmentation period of the continuous pulse signal is obtained through the average value and standard deviation of the period. The specific calculation formula is as follows.

[0038] ,

[0039] ,

[0040] ,

[0041] in, The average value of the continuous pulse signal is the periodic value. The standard deviation of the period of the continuous pulse signal. To set a constant window width, To analyze continuous pulse signals in the sample, This represents the position of the last point of the window. For the first one inside the window The signal of the point, The segmentation period of a continuous pulse signal;

[0042] The specific formula for extracting the frequency domain features of the pulse signal using the Fast Fourier Transform is as follows;

[0043] ,

[0044] in, The frequency domain signal of a periodic pulse signal One portion, The first in a continuous pulse signal One sample, The total number of samples, Representing the first frequency domain One frequency component, , It is the imaginary unit.

[0045] Furthermore, the specific method for establishing a cardiac function assessment model through a neural network is as follows:

[0046] The cardiac function assessment model established by the neural network is based on the long short-term memory neural network. The neural network model includes an input gate, a forgetting gate, and an output gate.

[0047] The equation for the forgetting gate is:

[0048] ,

[0049] in, The current open / closed state of the forget gate. This represents the sigmoid function. For the weight of the forget gate, Forget gate bias parameters, This is the hidden state from the previous moment. Input switch for the current data;

[0050] Determine how many cell states from the previous time step need to be retained in the current time step;

[0051] The equation for the input gate is:

[0052] ,

[0053] in, To input the current open / closed state of the gate, This represents the sigmoid function. For the input gate weights, For the input gate bias parameters, This is the output hidden state from the previous time step. The features include pulse characteristics and age and gender information in the feature dataset;

[0054] ,

[0055] in, Candidate cell state To represent the hyperbolic tangent function, For candidate cell weights, These are the candidate cell state bias parameters. This is the hidden state from the previous moment. Input switch for the current data;

[0056] Determine how much of the network's input data at the current moment needs to be saved to the cell state;

[0057] The updated equation is:

[0058] ,

[0059] in, Current cell state This represents the cell state at the previous moment;

[0060] The equation for the output gate is:

[0061] ,

[0062] ,

[0063] in, To output the current open / closed state of the gate. This represents the sigmoid function. For the output gate weights, These are the output gate bias parameters. This is the hidden state from the previous moment. The current data input switch, The results of the cardiac function assessment;

[0064] Control how much of the current cell state needs to be output to the current output value;

[0065] The mathematical expressions for the sigmoid function and the hyperbolic tangent function are:

[0066] ,

[0067] ,

[0068] in, For the sigmoid function, For the input of the sigmoid function, It is the hyperbolic tangent function. For function input.

[0069] Compared with the prior art, the beneficial effects of the present invention are:

[0070] This invention acquires samples of continuous pulse signals and assesses the cardiac function of these samples to establish a sample dataset. The continuous pulse signals in the sample dataset are then filtered and normalized to form a preprocessed dataset. Based on the amplitude and power spectral density of the continuous pulse signals, a pulse quality score is calculated for each signal. Continuous pulse signals are then filtered according to their pulse quality scores. Time-domain and frequency-domain features are extracted from the filtered signals. A cardiac function assessment model is established and trained using a neural network. The continuous pulse signals, along with age and gender, are collected by the acquisition module as input to the assessment model to complete the assessment of cardiac function.

[0071] This invention filters and reduces noise in continuous pulse signals from a sample dataset by setting a noise threshold, and then normalizes the signals. Through wavelet transform, the signals can be observed at different frequency levels, effectively separating noise from useful signal components while preserving important signal features. Based on the amplitude and power spectral density of the continuous pulse signals, a pulse quality score is calculated for each signal. Continuous pulse signals are then filtered based on these scores, allowing for the selection of high-quality signals for analysis, significantly improving the accuracy of cardiac function assessment. Time-domain and frequency-domain feature extraction from the pulse signals serves as the basis for cardiac function assessment, enhancing data representation, accuracy, and adaptability in multiple ways. A cardiac function assessment model is established based on a neural network, effectively capturing the complex dynamics of cardiac health status changes over time, and possessing the ability to automatically learn and adapt to the data characteristics of different patients. Attached Figure Description

[0072] Figure 1 This is a schematic diagram of the overall system structure of the present invention. Detailed Implementation

[0073] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0074] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0075] Example:

[0076] Please see Figure 1 The present invention provides a technical solution:

[0077] A continuous pulse signal processing and cardiac function assessment system includes a data acquisition module, a preprocessing module, a signal assessment module, a feature extraction module, and a cardiac function assessment module, wherein;

[0078] The data acquisition module acquires continuous pulse signal samples from participants of different ages and genders, assesses the cardiac function of each participant, establishes a sample dataset, and maps the continuous pulse signals, age, gender, and timestamps in the sample dataset one by one. At the same time, it acquires the continuous pulse signals, age, and gender information of the participants to be evaluated.

[0079] The pulse signal is a vibration signal caused by the heart pumping blood and the impact of blood on blood vessels. It captures the amount of blood pumped by each heartbeat by sensing the expansion of the blood vessel walls. It is a linear, non-stationary, weak physiological signal. There is a close relationship between the pulse signal and cardiac function. Therefore, the pulse signal is an important physiological parameter reflecting cardiac function and the state of the circulatory system.

[0080] Pulse wave signals can usually be acquired using photoplethysmography (PPG), a non-invasive pulse wave detection technique that uses photoelectric elements to detect changes in blood volume caused by heartbeats on the surface of human skin, thereby reflecting the pulse wave morphology and obtaining the pulse signal.

[0081] As we age, blood vessel walls may lose elasticity, causing pulse waves to travel faster and resulting in higher pulse pressures in the pulse waveform. Furthermore, women typically have smaller hearts than men, leading to slightly higher resting heart rates. This means that gender differences can cause variations in the normal range of heart rate and cardiac output when assessing cardiac function. Therefore, when using pulse signals to assess cardiac function, it is generally necessary to consider individual age and gender. This helps in the more accurate interpretation of pulse signal changes and requires adjusting the interpretation of pulse waveform analysis based on factors such as the potential influence of physiological differences on pulse characteristics.

[0082] The preprocessing module analyzes the noise characteristics of the continuous pulse signals in the sample dataset, sets a noise threshold for filtering and noise reduction, filters out noise caused by device acquisition, and then normalizes the noise-reduced continuous pulse signals to form a preprocessed dataset.

[0083] Continuous pulse signals are easily interfered with during acquisition, generating noise. This noise includes baseline drift due to skin temperature changes, environmental noise, operator movements, and significant circuit noise. This noise makes further diagnostic and monitoring using direct pulse signals difficult. Therefore, processing noisy pulse signals is crucial and has practical value. Wavelet transform can analyze signals at different scales, making it particularly suitable for processing biosignals with transient characteristics, such as pulse signals. Through wavelet transform, signals can be observed at different frequency levels, effectively separating noise from useful signal components. By setting an appropriate threshold, wavelet coefficients below that threshold, typically considered noise, can be eliminated. This method is particularly suitable for removing random noise while preserving important signal characteristics, such as the basic rhythm and transient beating features of the heart.

[0084] In this embodiment, the specific steps for filtering and noise reduction by setting a noise threshold are as follows:

[0085] First, wavelet transform is performed on the continuous pulse signal to decompose it into a series of approximation coefficients and detail coefficients of different frequencies;

[0086] ,

[0087] in, For continuous pulse signals in the sample dataset, These are approximate coefficients. For detail coefficients, For scaling function, For wavelet functions, These are the acquisition time points for continuous pulse signals. These are the scale parameter and the displacement parameter, respectively.

[0088] By setting a noise threshold, the wavelet coefficients are processed to remove the detail coefficients affected by noise.

[0089] ,

[0090] in, The detail coefficient after noise reduction. The noise threshold is set.

[0091] The denoised signal is reconstructed by performing inverse wavelet transform using the processed wavelet coefficients.

[0092] ,

[0093] in, This represents the continuous pulse signal of the denoised sample.

[0094] In this embodiment, the calculation formula for normalizing the noise-reduced continuous pulse signal is as follows:

[0095] ,

[0096] in, The pulse signal after normalization. This is the denoised pulse signal. The minimum value in the original pulse signal data. This is the maximum value in the original pulse signal data.

[0097] Normalization is a data mapping method that transforms data into the range (-1, 1). Since different evaluation indicators have different dimensions, data normalization is used to make them dimensionless for comprehensive analysis. Normalization can also solve numerical problems caused by excessively large values.

[0098] The signal evaluation module is used to obtain the amplitude and power spectral density of continuous pulse signals from the preprocessed dataset, calculate the average and variance of the amplitude of the continuous pulse signals and the median of the low-to-high frequency ratio, calculate the pulse quality score of each continuous pulse signal, filter the continuous pulse signals according to the pulse quality score, and use the filtered continuous pulse signals as the analysis dataset.

[0099] In this embodiment, the formulas for calculating the average value and variance of the continuous pulse signal amplitude are as follows:

[0100] ,

[0101] ,

[0102] in, For continuous pulse signals, the first The amplitude of the main peak wave, The number of main peak waves in a continuous pulse signal. The average amplitude of the main peak wave of the continuous pulse signal. This represents the variance of the amplitude of the main peak wave of the continuous pulse signal.

[0103] A clean pulse signal has a stable and regular waveform, while a pulse signal subjected to a lot of noise has a large amplitude fluctuation. The average value and variance of the signal amplitude are significantly different between clean and noisy pulse signals. Therefore, the average value and variance of the signal amplitude can be used as characteristics to evaluate the quality of the pulse signal.

[0104] The formula for calculating the median low-to-high frequency ratio is:

[0105] ,

[0106] in, The median of the low-to-high frequency ratio, For frequency The power spectral density at that point These represent the frequency range of the continuous pulse signal. These represent the low-frequency range of the continuous pulse signal.

[0107] Since the frequency range of a pulse signal is 0-20Hz, with most of its energy concentrated in the 1-10Hz frequency range, and considering the types of interference in the pulse signal, the low-to-high frequency ratio (LPR) is defined as the ratio of the energy in the 1-2.25Hz frequency range to the energy in the 0-8Hz frequency range. This invention calculates the LPR of each pulse in a pulse signal segment and uses the median LPR as a characteristic for evaluating signal quality.

[0108] The formula for calculating the continuous pulse quality score is as follows:

[0109] ,

[0110] in, For continuous pulse quality scoring, These represent the weights of the average amplitude, variance, and median low-to-high frequency ratio of the continuous pulse signal on the quality score.

[0111] The continuous pulse quality score reflects the overall signal quality of the continuous pulse signal. Although noise reduction can improve the clarity of the signal, there may still be quality differences between different signals. These differences may come from different conditions during signal acquisition, individual differences, equipment performance, and other factors. Selecting high-quality signals can reduce these effects, thereby reducing the errors that may be introduced during the analysis process.

[0112] The feature extraction module is used to perform time-domain feature analysis on the continuous pulse signal in the analysis dataset, and then extract the frequency-domain features of the pulse signal through fast Fourier transform, and form a feature dataset by combining the time-domain features and the frequency-domain features.

[0113] Extracting time-domain and frequency-domain features from pulse signals as a basis for assessing cardiac function can enhance the representativeness of data and improve the accuracy and adaptability of cardiac function assessment in many ways. Time-domain features reveal the behavior and characteristics of the signal in the time dimension, such as heart rate and waveform morphology, which are important indicators for directly assessing cardiac function. Frequency-domain features provide information on the frequency distribution of the signal, such as heart rate variability and spectral energy distribution. These features can reflect the autonomic nervous regulation of the heart under physiological and pathological conditions. The information provided by time-domain and frequency-domain analysis is complementary to some extent, and their combined use can provide a more comprehensive understanding of the health status of cardiac function.

[0114] In this embodiment, the process of time-domain feature analysis is as follows: by periodically segmenting the continuous pulse signal, the time-domain features of the pulse signal in each period are obtained. A window of constant width is set according to the length and characteristics of the pulse signal. The new sampling point is placed at the highest position of the window. The average value and standard deviation of the continuous pulse signal period are calculated. The segmentation period of the continuous pulse signal is obtained through the average value and standard deviation of the period. The specific calculation formula is as follows.

[0115] ,

[0116] ,

[0117] ,

[0118] in, The average value of the continuous pulse signal is the periodic value. The standard deviation of the period of the continuous pulse signal. To set a constant window width, To analyze continuous pulse signals in the sample, This represents the position of the last point of the window. For the first one inside the window The signal of the point, The segmentation period of a continuous pulse signal.

[0119] The specific formula for extracting the frequency domain features of the pulse signal using the Fast Fourier Transform is as follows;

[0120] ,

[0121] in, The frequency domain signal of a periodic pulse signal One portion, The first in a continuous pulse signal One sample, The total number of samples, Representing the first frequency domain One frequency component, , It is the imaginary unit.

[0122] The Fast Fourier Transform (FFT) algorithm can transform a signal from the time domain to the frequency domain, revealing the amplitude and phase characteristics of each frequency component. In cardiac function assessment, performing FFT on pulse signals and analyzing their frequency domain characteristics is highly beneficial because it provides a non-invasive, intuitive, and efficient way to understand the health status of cardiac function. Transforming the pulse signal from the time domain to the frequency domain allows us to see the different frequency components within the signal. Each heartbeat produces specific frequency patterns in the pulse signal. For example, acceleration and deceleration of the heart rate affect the amplitude and distribution of these frequencies. Analyzing these frequency components can effectively identify cardiac symptoms such as arrhythmias and abnormal heart rates.

[0123] The cardiac function assessment module is used to establish a cardiac function assessment model through a neural network. The model is trained by using a feature dataset of continuous pulse signals, along with age and gender, as inputs and cardiac function assessment results as outputs. The pulse signals to be assessed, along with age and gender, collected by the acquisition module are input into the trained assessment model to complete the assessment of cardiac function.

[0124] Long Short-Term Memory (LSTM) networks are a special type of recurrent neural network that is well-suited for processing and predicting long-term dependencies in time-series data. In applications using pulse signals to assess cardiac function, LSTM networks can effectively capture the complex dynamics of cardiac health over time. Pulse signals are a typical type of time-series data, continuously recording changes in heartbeats and blood flow over time. Cardiac pathologies often form time-dependent patterns in pulse signals. Due to its internal gating mechanisms, such as forget gates, input gates, and output gates, LSTM networks can effectively learn these long-term dependencies in time series data. Through its multi-layered structure and gating mechanisms, it can simultaneously process and integrate these complex signal features, providing a more accurate assessment of cardiac function.

[0125] In long-term health monitoring, particularly for patients with chronic heart disease, Long Short-Term Memory (LSTM) networks can process data over long periods and learn the evolution of changes in cardiac function. This ability makes LTM networks particularly suitable for predicting health trends in cardiac function and potential acute events.

[0126] By combining long short-term memory networks and pulse signals, a powerful cardiac function assessment model can be constructed. This model can learn from continuous physiological data and predict cardiac health, providing accurate and personalized cardiac function assessment results.

[0127] In this embodiment, the specific method for establishing a cardiac function assessment model through a neural network is as follows:

[0128] The cardiac function assessment model established by the neural network is based on the long short-term memory neural network. The neural network model includes an input gate, a forgetting gate, and an output gate.

[0129] The equation for the forgetting gate is:

[0130] ,

[0131] in, The current open / closed state of the forget gate. This represents the sigmoid function. For the weight of the forget gate, Forget gate bias parameters, This is the hidden state from the previous moment. Input switch for the current data;

[0132] Determine how many cell states from the previous time step need to be retained in the current time step;

[0133] The equation for the input gate is:

[0134] ,

[0135] in, To input the current open / closed state of the gate, This represents the sigmoid function. For the input gate weights, For the input gate bias parameters, This is the output hidden state from the previous time step. The features include pulse characteristics and age and gender information in the feature dataset;

[0136] ,

[0137] in, Candidate cell state To represent the hyperbolic tangent function, For candidate cell weights, These are the candidate cell state bias parameters. This is the hidden state from the previous moment. Input switch for the current data;

[0138] Determine how much of the network's input data at the current moment needs to be saved to the cell state;

[0139] The updated equation is:

[0140] ,

[0141] in, Current cell state This represents the cell state at the previous moment;

[0142] The equation for the output gate is:

[0143] ,

[0144] ,

[0145] in, To output the current open / closed state of the gate. This represents the sigmoid function. For the output gate weights, These are the output gate bias parameters. This is the hidden state from the previous moment. The current data input switch, The results of the cardiac function assessment;

[0146] Control how much of the current cell state needs to be output to the current output value;

[0147] The mathematical expressions for the sigmoid function and the hyperbolic tangent function are:

[0148] ,

[0149] ,

[0150] in, For the sigmoid function, For the input of the sigmoid function, It is the hyperbolic tangent function. For function input.

[0151] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0152] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0153] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0154] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A continuous pulse signal processing and cardiac function assessment system, characterized in that, include: The data acquisition module is used to acquire continuous pulse signal samples from samplers of different ages and genders, evaluate the cardiac function of each person, establish a sample dataset, and map the continuous pulse signals, age, gender and timestamps in the sample dataset one by one. At the same time, it acquires the continuous pulse signals and age and gender information of the person to be evaluated. The preprocessing module is used to analyze the noise characteristics of the continuous pulse signal in the sample dataset, set a noise threshold for filtering and noise reduction, filter out the noise caused by the device acquisition, and then normalize the noise-reduced continuous pulse signal to form a preprocessed dataset. The signal evaluation module is used to obtain the amplitude and power spectral density of the continuous pulse signal in the preprocessed dataset, calculate the average value and variance of the amplitude of the continuous pulse signal and the median of the low-frequency ratio, calculate the pulse quality score of each continuous pulse signal, filter the continuous pulse signals according to the pulse quality score, and use the filtered continuous pulse signals as the analysis dataset. The feature extraction module is used to perform time-domain feature analysis on the continuous pulse signal in the analysis dataset, and then extract the frequency-domain features of the pulse signal through fast Fourier transform, and form a feature dataset by combining the time-domain features and the frequency-domain features. The cardiac function assessment module is used to establish a cardiac function assessment model through a neural network. The model is trained by using a feature dataset of continuous pulse signals and age and gender as inputs, and the cardiac function assessment results as outputs. The pulse signals to be assessed, along with age and gender, collected by the acquisition module are input into the trained assessment model to complete the assessment of cardiac function. The formulas for calculating the average value and variance of the amplitude of the continuous pulse signal are as follows: in, For continuous pulse signals, the first The amplitude of the main peak wave, The number of main peak waves in a continuous pulse signal. The average amplitude of the main peak wave of the continuous pulse signal. The variance of the amplitude of the main peak wave of the continuous pulse signal; The formula for calculating the median low-to-high frequency ratio is: in, The median of the low-to-high frequency ratio, For frequency The power spectral density at that point These represent the frequency range of the continuous pulse signal. These represent the low-frequency range of the continuous pulse signal; The formula for calculating the continuous pulse quality score is as follows: in, For continuous pulse quality scoring, These represent the weights of the average amplitude, variance, and median low-to-high frequency ratio of the continuous pulse signal in relation to the quality score.

2. The continuous pulse signal processing and cardiac function assessment system according to claim 1, characterized in that: The specific steps for filtering and noise reduction by setting a noise threshold are as follows: First, wavelet transform is performed on the continuous pulse signal to decompose it into a series of approximation coefficients and detail coefficients of different frequencies; in, For continuous pulse signals in the sample dataset, These are approximate coefficients. For detail coefficients, For scaling function, For wavelet functions, These are the acquisition time points for continuous pulse signals. These are the scale parameters and displacement parameters, respectively. By setting a noise threshold, the wavelet coefficients are processed to remove the detail coefficients affected by noise. in, The detail coefficient after noise reduction. The noise threshold is set; The denoised signal is reconstructed by performing inverse wavelet transform using the processed wavelet coefficients. in, This represents the continuous pulse signal of the denoised sample.

3. The continuous pulse signal processing and cardiac function assessment system according to claim 1, characterized in that: The calculation formula for normalizing the noise-reduced continuous pulse signal is as follows: in, The normalized pulse signal This is the denoised pulse signal. The minimum value in the original pulse signal data. This is the maximum value in the original pulse signal data.

4. The continuous pulse signal processing and cardiac function assessment system according to claim 1, characterized in that: The specific method for establishing a cardiac function assessment model using a neural network is as follows: The cardiac function assessment model established by the neural network is based on the long short-term memory neural network. The neural network model includes an input gate, a forgetting gate, and an output gate. The equation for the forgetting gate is: in, The current open / closed state of the forget gate. This represents the sigmoid function. For the weight of the forget gate, Forget gate bias parameters, This is the hidden state from the previous moment. Input switch for the current data; Determine how many cell states from the previous time step need to be retained in the current time step; The equation for the input gate is: in, To input the current open / closed state of the gate, This represents the sigmoid function. For the input gate weights, For the input gate bias parameters, This is the output hidden state from the previous time step. The features include pulse characteristics and age and gender information in the feature dataset; in, Candidate cell state To represent the hyperbolic tangent function, For candidate cell weights, These are the candidate cell state bias parameters. This is the hidden state from the previous moment. Input switch for the current data; Determine how much of the network's input data at the current moment needs to be saved to the cell state; The updated equation is: in, Current cell state This represents the cell state at the previous moment; The equation for the output gate is: in, To output the current open / closed state of the gate. This represents the sigmoid function. For the output gate weights, These are the output gate bias parameters. This is the hidden state from the previous moment. The current data input switch, The results of the cardiac function assessment; Control how much of the current cell state needs to be output to the current output value; The mathematical expressions for the sigmoid function and the hyperbolic tangent function are: in, For the sigmoid function, For the input of the sigmoid function, It is the hyperbolic tangent function. For function input.

Images