A psychological stress level assessment system and method based on pulse wave signal
By performing first-order differential and filtering on pulse wave signals, multi-domain heart rate variability characteristics are extracted, and combined with machine learning methods, the problem of inaccurate psychological stress assessment in the existing technology is solved, and a more accurate psychological stress level assessment is achieved.
Patent Information
- Application Number
- CN202211096007.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-09-08
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2042-09-08
AI Technical Summary
The existing psychological stress level assessment methods rely on scales and doctors' subjective judgments, cannot accurately reflect the individual's psychological stress status, and lack continuous monitoring capabilities, so they cannot effectively evaluate psychological stress levels, especially in early identification of diseases such as cardiovascular diseases.
By performing first-order difference, filtering and peak identification processing on pulse wave signals, the heart rate variability characteristics in the time domain, frequency domain, nonlinear and time frequency domain are extracted, and a psychological stress level evaluation model is constructed in combination with machine learning methods to provide more accurate Psychological stress level assessment.
It improves the accuracy and classification performance of psychological stress level assessment, can more comprehensively explore deep-level information, and provide better application value than traditional methods.
Smart Images

Figure CN115363586B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of physiological signal analysis, and in particular relates to a system and method for evaluating psychological stress level based on pulse wave signals. Background Art
[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute the prior art that has been known to a person skilled in the art.
[0003] Long-term exposure to chronic or extreme stress can lead to a series of diseases such as cardiovascular disease, Alzheimer's disease, depression, etc. It is of great significance to accurately assess the stress state and identify the psychological changes that may develop into diseases as early as possible so as to carry out effective intervention in time.
[0004] At present, the commonly used method for assessing the level of psychological stress is still mainly based on the scale. Doctors make a diagnosis by combining the interview with the scale results, but this method is greatly influenced by the doctor's subjective experience, and patients are usually unable to accurately describe their feelings, and may even conceal their condition due to reasons such as stigma. The lack of professional consultants and the inability to continuously monitor the scale survey have further restricted the development of early detection and treatment of psychological stress.
[0005] Pulse wave signals can effectively reflect changes in the regulation state of the autonomic nervous system, and have the advantages of being non-invasive, easy to operate, and low-cost. Conventional methods mostly calculate its simple time domain and frequency domain features, and combine traditional machine learning algorithms to establish the relationship between features and stress levels. They do not consider the nonlinearity and time-frequency domain information unique to the human body as a complex physiological system, and do not fully explore the deep-level effective information related to psychological stress, so they cannot accurately assess the individual's psychological stress level. Summary of the invention
[0006] In order to solve the above problems, the present invention provides a system and method for evaluating psychological stress level based on pulse wave signals, which obtains a cardiac cycle sequence by performing first-order difference, filtering and peak recognition processing on the pulse wave signal, and extracts multi-domain heart rate variability features including time domain features, frequency domain features, nonlinear features and time-frequency domain features from the cardiac cycle sequence for psychological stress level evaluation, so as to fully tap the deep-level effective information related to psychological stress and improve the accuracy of psychological stress level evaluation.
[0007] In order to achieve the above object, the present invention mainly includes the following aspects:
[0008] In a first aspect, an embodiment of the present invention provides a psychological stress level assessment system based on a pulse wave signal, comprising:
[0009] A signal processing module, used for acquiring a pulse wave signal of a target object, performing first-order difference, filtering and peak recognition processing on the pulse wave signal, and obtaining a cardiac cycle sequence of the target object;
[0010] A feature extraction module, used for extracting the heart rate variability features of the cardiac cycle sequence, wherein the heart rate variability features include time domain features, frequency domain features, nonlinear features and time-frequency domain features;
[0011] The level assessment module is used to perform feature selection on the heart rate variability features and construct a feature set, input the feature set into a preset psychological stress level assessment model, and assess the psychological stress level of the target object according to the output results of the model; wherein the psychological stress level assessment model is obtained by training a classifier using a machine learning method.
[0012] In a possible implementation, it further includes a data acquisition module for acquiring a pulse wave signal of a target object; the data acquisition module includes a pulse wave piezoelectric sensor and an amplifying filter that are electrically connected to each other, and the signal acquired by the pulse wave piezoelectric sensor is amplified by the amplifying filter to obtain the pulse wave signal of the target object.
[0013] In a possible implementation, the signal processing module is specifically used to perform A / D conversion on the acquired pulse wave signal, perform first-order difference and filtering on the pulse wave signal after A / D conversion, and obtain a differential sequence; determine the maximum value of each numerical point in the differential sequence, use it as the first identified peak, and use this as a fixed point to slide the window forward or backward to obtain a numerical point that meets a preset threshold as the second peak, so as to construct the cardiac cycle sequence of the target object.
[0014] In a possible implementation, the feature extraction module is used to perform discrete Fourier transform on the cardiac cycle sequence to calculate and extract frequency domain features.
[0015] In one possible implementation, the nonlinear features include but are not limited to one or more of Shannon entropy, approximate entropy, sample entropy, fuzzy entropy, permutation entropy, distribution entropy, Poincare map minor axis, major axis, minor axis / major axis, vector angle index, vector length index, Porta index, and Guzik index.
[0016] In a possible implementation, the feature extraction module is used to perform a wavelet transform of a preset scale on the cardiac cycle sequence using the db4 mother wavelet, and take the mean and standard deviation of the absolute value of the detail coefficient of each layer of the reconstructed signal as the time-frequency domain features.
[0017] In a possible implementation, in the grade evaluation module, when performing feature selection on the heart rate variability features, a recursive elimination method based on a support vector machine is used to screen the features to obtain a feature set.
[0018] In a second aspect, an embodiment of the present invention provides a method for evaluating a psychological stress level based on a pulse wave signal, comprising:
[0019] Acquire a pulse wave signal of a target object, perform first-order difference, filtering and peak recognition processing on the pulse wave signal, and obtain a cardiac cycle sequence of the target object;
[0020] Extracting the heart rate variability features of the cardiac cycle sequence, wherein the heart rate variability features include time domain features, frequency domain features, nonlinear features and time-frequency domain features;
[0021] Feature selection is performed on the heart rate variability features, and a feature set is constructed, the feature set is input into a preset psychological stress level assessment model, and the psychological stress level of the target object is assessed according to the output results of the model; wherein the psychological stress level assessment model is obtained by training a classifier using a machine learning method.
[0022] In a third aspect, an embodiment of the present invention provides a computer device, comprising: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the computer device is running, the processor and the memory communicate via the bus, and when the machine-readable instructions are executed by the processor, the steps of the method for assessing the psychological stress level based on the pulse wave signal as described in the second aspect above are performed.
[0023] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the method for assessing the psychological stress level based on the pulse wave signal as described in the second aspect above are executed.
[0024] One or more of the above technical solutions have the following beneficial effects:
[0025] (1) Different from the existing method of establishing the relationship between features and stress levels based on the time domain features and frequency domain features of the pulse wave signal in combination with the traditional machine learning algorithm, the psychological stress level assessment system proposed in the present invention obtains the cardiac cycle sequence by performing first-order difference, filtering and peak recognition processing on the pulse wave signal, and extracts multi-domain heart rate variability features including time domain features, frequency domain features, nonlinear features and time-frequency domain features from the cardiac cycle sequence for psychological stress level assessment. In this way, the deep-level effective information related to psychological stress can be more comprehensively excavated, and the psychological stress level can be assessed more accurately.
[0026] (2) The extracted multi-domain features of heart rate variability are screened using a cyclic recursive elimination method based on support vector machine, which not only retains the features related to psychological stress to the greatest extent, but also greatly improves the computational efficiency.
[0027] (3) By extracting multi-domain heart rate variability features based on cardiac cycle sequences, the features are input into the psychological stress level assessment model, and the psychological stress level corresponding to the pulse wave signal is evaluated according to its output results; since the multi-domain features can provide a more comprehensive picture of changes in the autonomic nervous system, the present invention can provide better classification performance than existing methods, improve the accuracy of psychological stress level assessment, and effectively improve the application value of heart rate variability analysis based on pulse wave signals in psychological stress level assessment. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.
[0029] Figure 1 is a structural schematic diagram of a psychological stress level assessment system based on pulse wave signals provided in Embodiment 1 of the present invention;
[0030] Figure 2 It is a flow chart of a method for evaluating psychological stress level based on pulse wave signals provided in the second embodiment of the present invention. DETAILED DESCRIPTION
[0031] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.
[0032] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.
[0033] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.
[0034] Embodiment 1
[0035] This embodiment provides a psychological stress level assessment system based on pulse wave signals, including:
[0036] A signal processing module, used for acquiring a pulse wave signal of a target object, performing first-order difference, filtering and peak recognition processing on the pulse wave signal, and obtaining a cardiac cycle sequence of the target object;
[0037] A feature extraction module, used for extracting the heart rate variability features of the cardiac cycle sequence, wherein the heart rate variability features include time domain features, frequency domain features, nonlinear features and time-frequency domain features;
[0038] The level assessment module is used to perform feature selection on the heart rate variability features and construct a feature set, input the feature set into a preset psychological stress level assessment model, and assess the psychological stress level of the target object according to the output results of the model; wherein the psychological stress level assessment model is obtained by training a classifier using a machine learning method.
[0039] The psychological stress level assessment system provided by the embodiment of the present invention includes a signal processing module, a feature extraction module and a level assessment module. The cardiac cycle sequence is obtained by performing first-order difference, filtering and peak recognition processing on the pulse wave signal of the target object, and then the multi-domain heart rate variability features of the cardiac cycle sequence including time domain features, frequency domain features, nonlinear features and time-frequency domain features are extracted. Then, feature selection is performed on the heart rate variability features in the level assessment module, and a feature set is constructed to assess the psychological stress level of the target object. In this way, deep and effective information related to psychological stress can be more comprehensively excavated, and a more accurate assessment of the psychological stress level can be given.
[0040] In the specific implementation, Figure 1 As shown, the psychological stress level assessment system also includes a data acquisition module for collecting the pulse wave signal of the target object; the data acquisition module includes a pulse signal acquisition unit 1, and the pulse signal acquisition unit 1 is specifically a pulse wave piezoelectric sensor and an amplifying filter that are electrically connected to each other. The signal collected by the pulse wave piezoelectric sensor is amplified by the amplifying filter to obtain the pulse wave signal of the target object. In the evaluation process of the pulse wave signal, the data acquisition module is used to collect the pulse wave signal of the target object; in the training process of the psychological stress level assessment model, the data acquisition module is used to collect the pulse wave signal as sample data. In a specific application, the pulse signal acquisition unit 1 is placed at the tip of the index finger of the target object to collect the pulse wave signal.
[0041] The feature extraction module includes an A / D conversion unit 2 and a preprocessing unit 3, wherein the A / D conversion unit 2 is used to perform A / D conversion on the acquired pulse wave signal; the preprocessing unit 3 is used to perform first-order difference and filtering on the pulse wave signal after A / D conversion to obtain a differential sequence, and identify the peak of the first-order differential sequence of the pulse wave to construct a cardiac cycle time series. In a specific application, a fourth-order Butterworth bandpass filter is used for 0.5-10Hz bandpass filtering. The peak recognition method is to determine the maximum value of each numerical point in the differential sequence, take it as the first peak identified, and use it as a fixed point to slide the window forward or backward to obtain a numerical point that meets the preset threshold as the second peak, so as to construct the cardiac cycle sequence of the target object.
[0042] The feature extraction module 4 is used to extract the heart rate variability features in multiple domains such as time domain, frequency domain, nonlinearity and time-frequency domain of the cardiac cycle sequence.
[0043] In a specific implementation, the method for extracting heart rate variability features includes:
[0044] 1. For the preprocessed cardiac cycle sequence, the mean mRR, median medRR, standard deviation SDNN, ratio of standard deviation to mean HRVC, kurtosis St, skewness Kt, cumulative sum Area, 20th quantile Q20, 80th quantile Q80, interquartile difference QD, mean mSD, standard deviation SDSD, root mean square RMSSD of adjacent interval differences, number of adjacent interval differences greater than 50ms NN50, percentage of adjacent interval differences greater than 50ms, 25ms, 20ms and 10ms in the number of sinus beats pNN50, pNN25, pNN20, pNN10, mean mHR and standard deviation stdHR of heart rate, triangular index TRI, triangular interpolation TINN of interval histogram and other 22 statistical features were calculated as the time domain features of the cardiac cycle sequence.
[0045] 2. Perform discrete Fourier transform on the preprocessed cardiac cycle sequence to calculate and extract frequency domain features. The frequency domain features include 19 features, including the mean Freq, standard deviation stdFreq, skewFreq and kurtFreq of the spectrum sequence, very low frequency (0.0033-0.04Hz) power VLF, low frequency (0.04-0.15Hz) power LF, high frequency (0.15-0.4Hz) power HF, total power (0.0033-0.4Hz) TF, normalized low frequency power LFn, normalized high frequency power HFn, low-high frequency power ratio LF / HF, low-high frequency energy ratio LH_Energy, low frequency power ratio Per_LF, high frequency power ratio Per_HF, very low frequency power ratio Per_VLF, coherence ratio CR, frequency peak value VLF_peak in the very low frequency range, frequency peak value LF_peak in the low frequency range, and frequency peak value HF_peak in the high frequency range.
[0046] 3. The nonlinear features of the preprocessed cardiac cycle sequence are calculated. The nonlinear features include 13 features such as Shannon entropy, approximate entropy, sample entropy, fuzzy entropy, permutation entropy, distribution entropy, short axis, long axis, short axis / long axis of Poincare map, vector angle index, vector length index, Porta index and Guzik index.
[0047] Specifically, the process of calculating sample entropy and fuzzy entropy is:
[0048] Step 3.1.1: Normalize the cardiac cycle sequence by subtracting the mean and dividing by the standard deviation;
[0049] Step 3.1.2: For the sequence {u i =u(i), 1≤i≤N}, reconstruct the following space vector Among them, m is the reconstruction dimension and N is the length of the time series.
[0050]
[0051] Step 3.1.3: Definition and The distance d between them is: calculate If the sample entropy is calculated, then A() is the Heaviside step function; if the fuzzy entropy is calculated, then is a Gaussian function.
[0052] Step 3.1.4: Calculation The average value and logarithm are:
[0053] Step 3.1.5: Increase the reconstruction dimension to m+1, repeat steps 3.1.2 to 3.1.4, and calculate the statistical features Depend on The sample entropy and fuzzy entropy are calculated according to the difference of A().
[0054] In this embodiment, the process of calculating the distribution entropy is:
[0055] Step 3.2.1: Normalize the cardiac cycle series by subtracting the mean and dividing by the standard deviation.
[0056] Step 3.2.2: For the sequence {u i =u(i), 1≤i≤N}, reconstruct the following space vector Among them, m is the reconstruction dimension and N is the length of the time series.
[0057]
[0058] Step 3.2.3: Define the matrix for and The distance between A matrix composed of 1≤i≤N-m+1, 1≤j≤N-m+1.
[0059] Step 3.2.4: In the interval [0, 1], use the histogram method to estimate The empirical probability density function of all elements except the main diagonal (i.e., i=j). The number of histograms is predefined as B, and the probability p of an element falling into each histogram is calculated. k (1≤k≤B).
[0060] Step 3.2.5: From the formula Find the entropy of the distribution.
[0061] 4. Calculate the time-frequency domain features of the preprocessed cardiac cycle sequence. Use the db4 mother wavelet to perform a 5-scale wavelet transform on the sequence, and take the mean and standard deviation of the absolute value of the detail coefficient of each layer of the reconstructed signal as 10 time-frequency domain features, denoted as mean and standard deviation. i and stdd i (i=1, 2, ..., 5).
[0062] The level evaluation module 5 is used to select the heart rate variability features and construct a feature set, input the feature set into a preset psychological stress level evaluation model, and evaluate the psychological stress level of the target object according to the output result of the model. The psychological stress level evaluation model uses the recursive elimination method of the support vector machine to select features; and uses the k-fold cross validation method to verify the results.
[0063] The method for generating a psychological stress level assessment model includes: obtaining cardiac cycle sequence samples of pulse wave sample data, extracting heart rate variability features of the cardiac cycle sequence samples including time domain features, frequency domain features, nonlinear features, time-frequency domain features, etc., and performing feature selection on the heart rate variability features to obtain feature selection results; constructing a sample feature set according to the feature selection results, training a classifier with a machine learning method, and obtaining a psychological stress level assessment model.
[0064] In the grade evaluation module, when selecting the heart rate variability feature, the recursive elimination method based on the support vector machine is used to screen the features to obtain a feature set. In the specific implementation, each iteration eliminates the feature with the smallest contribution rate in the feature set, and after multiple cycles of iteration, the optimal feature set of a specified number is obtained. In the obtained optimal feature set: mRR, medRR, HRVC, mSD, Q20, TRI, mHR, stdHR, skewFreq, kurtFreq, Per_VLF, VLF_peak, stdd1, meand2 feature joint analysis has a high recognition performance for the psychological stress group and the healthy group.
[0065] The embodiment of the present invention extracts multi-domain features based on the preprocessed cardiac cycle sequence, and after feature selection, inputs the features into a trained psychological stress level assessment model. The output result of the model provides information reference for psychological stress level assessment.
[0066] Embodiment 2
[0067] like Figure 2 As shown, the embodiment of the present invention also provides a method for evaluating the psychological stress level based on the pulse wave signal, which specifically includes the following steps:
[0068] S201: Acquire a pulse wave signal of a target object, perform first-order difference, filtering and peak recognition processing on the pulse wave signal, and obtain a cardiac cycle sequence of the target object;
[0069] S202: extracting heart rate variability features of the cardiac cycle sequence, wherein the heart rate variability features include time domain features, frequency domain features, nonlinear features, and time-frequency domain features;
[0070] S203: Perform feature selection on the heart rate variability feature and construct a feature set, input the feature set into a preset psychological stress level assessment model, and assess the psychological stress level of the target object according to the output result of the model; wherein the psychological stress level assessment model is obtained by training a classifier using a machine learning method.
[0071] The psychological stress level assessment method based on pulse wave signals provided in this embodiment is obtained based on the aforementioned psychological stress level assessment system based on pulse wave signals. The specific implementation method of the psychological stress level assessment method based on pulse wave signals can be found in the embodiment section of the psychological stress level assessment system based on pulse wave signals in the previous text, and will not be repeated here.
[0072] Embodiment 3
[0073] An embodiment of the present invention further provides a computer device, which includes a processor, a memory, and a bus.
[0074] The memory stores machine-readable instructions executable by the processor. When the computer device is running, the processor communicates with the memory via a bus. When the machine-readable instructions are executed by the processor, the above-mentioned Figure 2 The specific implementation of the steps of the method for evaluating the psychological stress level based on the pulse wave signal in the method embodiment shown can be found in the method embodiment, which will not be repeated here.
[0075] Embodiment 4
[0076] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the method for assessing the psychological stress level based on the pulse wave signal described in the above method embodiment are executed.
[0077] A person skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.
[0078] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A psychological stress level assessment system based on pulse wave signals, characterized in that: include: A signal processing module, used for acquiring a pulse wave signal of a target object, performing first-order difference, filtering and peak recognition processing on the pulse wave signal, and obtaining a cardiac cycle sequence of the target object; A feature extraction module, used for extracting the heart rate variability features of the cardiac cycle sequence, wherein the heart rate variability features include time domain features, frequency domain features, nonlinear features and time-frequency domain features; A level assessment module, used to perform feature selection on the heart rate variability feature and construct a feature set, input the feature set into a preset psychological stress level assessment model, and assess the psychological stress level of the target object according to the output result of the model; wherein the psychological stress level assessment model is obtained by training a classifier using a machine learning method; The signal processing module is specifically used to perform A / D conversion on the acquired pulse wave signal, perform first-order difference on the pulse wave signal after A / D conversion, and use a fourth-order Butterworth bandpass filter to perform 0.5-10 Hz bandpass filtering to obtain a differential sequence; determine the maximum value of each numerical point in the differential sequence, use it as the first wave peak identified, and use this as a fixed point to slide the window forward or backward to obtain a numerical point that meets a preset threshold as the second wave peak, so as to construct the cardiac cycle sequence of the target object; The process of extracting the heart rate variability feature specifically includes: For the preprocessed cardiac cycle sequence, the mean mRR, median medRR, standard deviation SDNN, ratio of standard deviation to mean HRVC, kurtosis St, skewness Kt, cumulative sum Area, 20th quantile Q20, 80th quantile Q80, interquartile difference QD, mean mSD, standard deviation SDSD, root mean square RMSSD of adjacent interval differences, number of adjacent interval differences greater than 50ms NN50, percentage of adjacent interval differences greater than 50ms, 25ms, 20ms and 10ms in the number of sinus beats pNN50, pNN25, pNN20, pNN10, mean mHR and standard deviation stdHR of heart rate, triangular index TRI, triangular interpolation TINN of interval histogram, a total of 22 statistical features were calculated as the time domain features of the cardiac cycle sequence; The preprocessed cardiac cycle sequence was subjected to discrete Fourier transform to calculate and extract frequency domain features, including spectrum sequence mean meanFreq, standard deviation stdFreq, skewness skewFreq and kurtFreq, very low frequency power VLF, low frequency power LF, high frequency power HF, total power TF, normalized low frequency power LFn, normalized high frequency power HFn, low and high frequency power ratio LF / HF, low and high frequency energy ratio LH_Energy, low frequency power proportion Per_LF, high frequency power proportion Per_HF, very low frequency power proportion Per_VLF, coherence ratio CR, frequency peak value VLF_peak in the very low frequency range, frequency peak value LF_peak in the low frequency range, and frequency peak value HF_peak in the high frequency range, a total of 19 features; The nonlinear features of the preprocessed cardiac cycle sequence were calculated, including 13 features: Shannon entropy, approximate entropy, sample entropy, fuzzy entropy, permutation entropy, distribution entropy, Poincare map short axis, long axis, short axis / long axis, vector angle index, vector length index, Porta index and Guzik index; The time-frequency domain features of the preprocessed cardiac cycle sequence were calculated, and the sequence was transformed into a 5-scale wavelet using the db4 mother wavelet. The mean and standard deviation of the absolute value of the detail coefficient of each layer of the reconstructed signal were taken as 10 time-frequency domain features.
2. The psychological stress level assessment system based on pulse wave signal as claimed in claim 1, characterized in that: It also includes a data acquisition module for acquiring the pulse wave signal of the target object; the data acquisition module includes a pulse wave piezoelectric sensor and an amplifying filter that are electrically connected to each other, and the signal acquired by the pulse wave piezoelectric sensor is amplified by the amplifying filter to obtain the pulse wave signal of the target object.
3. The psychological stress level assessment system based on pulse wave signal as claimed in claim 1, characterized in that: In the grade evaluation module, when performing feature selection on the heart rate variability feature, a recursive elimination method based on a support vector machine is used to screen the features to obtain a feature set.
4. A method for evaluating psychological stress level based on pulse wave signals, characterized in that: include: Acquire a pulse wave signal of a target object, perform first-order difference, filtering and peak recognition processing on the pulse wave signal, and obtain a cardiac cycle sequence of the target object; Extracting the heart rate variability features of the cardiac cycle sequence, wherein the heart rate variability features include time domain features, frequency domain features, nonlinear features and time-frequency domain features; Performing feature selection on the heart rate variability feature and constructing a feature set, inputting the feature set into a preset psychological stress level assessment model, and assessing the psychological stress level of the target object according to the output result of the model; wherein the psychological stress level assessment model is obtained by training a classifier using a machine learning method; The method for obtaining the cardiac cycle sequence of the target object specifically includes: performing A / D conversion on the acquired pulse wave signal, performing first-order difference on the pulse wave signal after A / D conversion, and using a fourth-order Butterworth bandpass filter to perform 0.5-10 Hz bandpass filtering to obtain a differential sequence; determining the maximum value of each numerical point in the differential sequence, taking it as the first wave peak identified, and using this as a fixed point to slide the window forward or backward to obtain a numerical point that meets a preset threshold as the second wave peak, so as to construct the cardiac cycle sequence of the target object; The method for extracting the heart rate variability characteristics of the cardiac cycle sequence specifically includes: For the preprocessed cardiac cycle sequence, the mean mRR, median medRR, standard deviation SDNN, ratio of standard deviation to mean HRVC, kurtosis St, skewness Kt, cumulative sum Area, 20th quantile Q20, 80th quantile Q80, interquartile difference QD, mean mSD, standard deviation SDSD, root mean square RMSSD of adjacent interval differences, number of adjacent interval differences greater than 50ms NN50, percentage of adjacent interval differences greater than 50ms, 25ms, 20ms and 10ms in the number of sinus beats pNN50, pNN25, pNN20, pNN10, mean mHR and standard deviation stdHR of heart rate, triangular index TRI, triangular interpolation TINN of interval histogram, a total of 22 statistical features were calculated as the time domain features of the cardiac cycle sequence; The preprocessed cardiac cycle sequence was subjected to discrete Fourier transform to calculate and extract frequency domain features, including spectrum sequence mean meanFreq, standard deviation stdFreq, skewness skewFreq and kurtFreq, very low frequency power VLF, low frequency power LF, high frequency power HF, total power TF, normalized low frequency power LFn, normalized high frequency power HFn, low and high frequency power ratio LF / HF, low and high frequency energy ratio LH_Energy, low frequency power proportion Per_LF, high frequency power proportion Per_HF, very low frequency power proportion Per_VLF, coherence ratio CR, frequency peak value VLF_peak in the very low frequency range, frequency peak value LF_peak in the low frequency range, and frequency peak value HF_peak in the high frequency range, a total of 19 features; The nonlinear features of the preprocessed cardiac cycle sequence were calculated, including 13 features: Shannon entropy, approximate entropy, sample entropy, fuzzy entropy, permutation entropy, distribution entropy, Poincare map short axis, long axis, short axis / long axis, vector angle index, vector length index, Porta index and Guzik index; The time-frequency domain features of the preprocessed cardiac cycle sequence were calculated, and the sequence was transformed into a 5-scale wavelet using the db4 mother wavelet. The mean and standard deviation of the absolute value of the detail coefficient of each layer of the reconstructed signal were taken as 10 time-frequency domain features.
5. A computer device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the computer device is running, the processor and the memory communicate via the bus, and when the machine-readable instructions are executed by the processor, the steps of the method for assessing the psychological stress level based on the pulse wave signal as described in claim 4 are performed.
6. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the steps of the method for evaluating the psychological stress level based on the pulse wave signal as claimed in claim 4 are executed.
