Psychological stress assessment method and system based on skin electric signal image coding

Abstract

The invention provides a psychological stress assessment method and system based on skin electric signal image coding, relates to the technical field of physiological signal analysis, and aims to solve the problem that in the prior art, the calculation process of complex nonlinear features in the time domain and the frequency domain of skin electric signals is complex and tedious. Non-linear changes of signals in a pressure state cannot be captured only depending on linear features, psychological pressure recognition precision is low, and robustness is poor. The method comprises the following steps: acquiring skin electric signals, carrying out preprocessing and Fourier transform to obtain a time sequence and a frequency spectrum sequence, coding the time sequence and the frequency spectrum sequence by adopting multiple coding methods to obtain multiple different coded images, extracting texture features of the coded images, selecting, constructing an optimal feature set, inputting the optimal feature set into a support vector machine, and carrying out image fusion. And obtaining an assessment result of the psychological stress state. Potential complex nonlinear characteristics in the time domain and the frequency domain of the skin electric signals are efficiently and comprehensively excavated, and the accuracy of psychological stress recognition is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of physiological signal analysis, and in particular relates to a method and system for evaluating psychological stress based on skin electrical signal image coding. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] In modern society, psychological stress has become an increasingly common and serious phenomenon. Long-term psychological stress can lead to a series of health problems, such as mental illness, obesity, and cardiovascular disease, and may even lead to sudden death, placing a series of burdens on the global health care system. Therefore, it is crucial to objectively and effectively detect and evaluate psychological stress.

[0004] Physiological studies have shown that when the human body perceives stress, the sympathetic nerves in the autonomic nervous system become abnormally excited, leading to increased sweat gland activity, which in turn increases skin conductance. Skin electrical signals measure changes in skin surface conductivity, which are closely related to the level of psychological stress. They have the advantages of simple acquisition and easy continuous monitoring, and can be applied to real-time assessment of psychological stress status. To use skin electrical signals for stress assessment, it is first necessary to extract indicators that can characterize the characteristics of signal changes under stress from different aspects of the signal, such as the time domain and frequency domain. However, the calculation process of nonlinear features is complex and cumbersome, and relying solely on linear features cannot effectively capture the nonlinear changes of signals under stress, thereby limiting the accuracy and robustness of psychological stress recognition. Therefore, how to efficiently and comprehensively mine the potential complex nonlinear characteristics (such as dynamic changes, chaos, and subtle details) in the time and frequency domains of skin electrical signals to improve the accuracy of psychological stress recognition has become an urgent problem to be solved. Summary of the invention

[0005] In order to overcome the shortcomings of the above-mentioned prior art, the present invention provides a method and system for evaluating psychological stress based on skin electrical signal image coding. Four advanced coding methods, namely Markov transition field, Gram angle sum field, Gram angle difference field and recursive graph, are used to convert the time domain sequence and spectrum sequence of skin electrical signal into two-dimensional images, and by combining multiple texture feature extraction techniques, the implicit patterns in the image are mined from different angles, so as to more accurately evaluate the psychological stress state.

[0006] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions:

[0007] The first aspect of the present invention discloses a method for evaluating psychological stress based on skin electrical signal image coding, comprising:

[0008] Collect the skin electrical signals of the test subject, preprocess the skin electrical signals to obtain the time series of the skin electrical signals;

[0009] Perform Fourier transform on the time series of the skin electrical signals to obtain the frequency spectrum sequence of the skin electrical signals;

[0010] Use Markov transition field, Gram angle sum field, Gram angle difference field and recurrence plot to encode the time series and frequency spectrum sequence of the skin electrical signals respectively to obtain a variety of different encoded images;

[0011] Adopt texture feature extraction technology to extract texture features from a variety of different encoded images to obtain multiple texture features;

[0012] Perform feature selection on multiple texture features, construct an optimal feature set, and input the optimal feature set into a support vector machine to obtain the evaluation result of the mental stress state.

[0013] As a further technical solution, preprocess the skin electrical signals, specifically:

[0014] Perform downsampling on the skin electrical signals to obtain the downsampled skin electrical signals;

[0015] Adopt a psychophysiological signal processing model to perform component decomposition and noise removal on the downsampled skin electrical signals.

[0016] As a further technical solution, use the Markov transition field to encode the time series and frequency spectrum sequence of the skin electrical signals. The specific process is as follows:

[0017] Discretize the time and frequency spectrum sequences of the skin electrical signals respectively using quantile intervals;

[0018] Calculate the transition probability between each interval to obtain the Markov transition matrix;

[0019] Based on the Markov transition matrix, by adding the correlation between each quantile and the time step, obtain the Markov transition field matrix;

[0020] Perform image encoding on the time series and frequency spectrum sequence of the skin electrical signals according to the Markov transition field matrix.

[0021] As a further technical solution, use the Gram angle sum field and the Gram angle difference field to encode the time series and frequency spectrum sequence of the skin electrical signals. The specific process is as follows:

[0022] Scale the time series and frequency spectrum sequence of the skin electrical signals respectively and convert them to the polar coordinate system;

[0023] By calculating the cosine value of the sum of angles and the sine value of the difference of angles between points in the polar coordinate system, the Gram angle sum field and the Gram angle difference field matrix are obtained;

[0024] According to the Gram angle sum field and the Gram angle difference field matrix, image coding is performed on the time series and the frequency spectrum series of the skin electrical signal.

[0025] As a further technical solution, a recurrence plot is used to code the time series and the frequency spectrum series of the skin electrical signal. The specific process is as follows:

[0026] The time series and the frequency spectrum series of the skin electrical signal are mapped into a high-dimensional phase space, and each point in the series is mapped into a high-dimensional vector;

[0027] The distance between every two points in the high-dimensional space is calculated to obtain a recurrence matrix,

[0028] According to the recurrence matrix, image coding is performed on the time series and the frequency spectrum series of the skin electrical signal.

[0029] As a further technical solution, the texture feature extraction technology includes: local binary pattern feature extraction, gray level co-occurrence matrix feature extraction, and first-order statistic feature extraction.

[0030] As a further technical solution, feature selection is performed on multiple texture features to construct an optimal feature set, and the optimal feature set is input into a support vector machine to obtain an evaluation result of the psychological stress state. The specific process is as follows:

[0031] The recursive feature elimination method based on the support vector machine is adopted. By repeatedly constructing a support vector machine classification model, the feature with the lowest score is removed in each iteration, and finally the optimal feature set is selected;

[0032] The optimal feature set is input into the support vector machine to obtain an evaluation result of the psychological stress state.

[0033] In a second aspect, a psychological stress evaluation system based on skin electrical signal image coding is disclosed, including:

[0034] A signal processing module, configured to collect the skin electrical signal of a tested person, preprocess the skin electrical signal to obtain a time series of the skin electrical signal; perform Fourier transform on the time series of the skin electrical signal to obtain a frequency spectrum series of the skin electrical signal;

[0035] An image coding module for sequences, configured to respectively use a Markov transition field, a Gram angle sum field, a Gram angle difference field, and a recurrence plot to code the time series and the frequency spectrum series of the skin electrical signal to obtain a variety of different coded images;

[0036] An image feature extraction module, which is used to extract texture features from a variety of different encoded images by using texture feature extraction technology to obtain multiple texture features;

[0037] A feature selection and classification module, which is used to perform feature selection on multiple texture features, construct an optimal feature set, and input the optimal feature set into a support vector machine to obtain an evaluation result of the psychological stress state.

[0038] The third aspect of the present invention discloses a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the method described in the first aspect are implemented.

[0039] The fourth aspect of the present invention discloses a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the steps of the method described in the first aspect are executed.

[0040] The above one or more technical solutions have the following beneficial effects:

[0041] (1) In this embodiment, a variety of image coding technologies are used to process the time-domain sequence and frequency-domain sequence of the skin electrical signal, comprehensively mining its advanced complex characteristics in the time domain and frequency domain, such as dynamic changes, chaos, and fine details. Compared with traditional feature extraction technologies, the present invention can optimize the capture of time-domain and frequency-domain information of the skin electrical signal, provide a richer information source for the recognition of psychological stress, and thus effectively improve the robustness of psychological stress assessment.

[0042] (2) In this embodiment, a variety of texture features are extracted from the encoded images, including local binary pattern, gray-level co-occurrence matrix, and first-order statistics, which can comprehensively characterize the texture features in the images from different dimensions. Through this multi-angle feature extraction method, richer information is obtained from the time-domain and frequency-domain encoded images, providing a comprehensive and detailed feature basis for psychological stress assessment.

[0043] (3) In this embodiment, by converting the time-domain sequence and frequency-spectrum sequence of the skin electrical signal into two-dimensional images and further extracting effective information in combination with various texture feature extraction technologies, the problems of cumbersome calculation and errors that may occur in the traditional feature extraction process are avoided. This processing method has stronger robustness, can effectively reduce the risk of errors, and greatly reduce the time and effort required for manual calculation, making the whole process more efficient.

[0044] (4) In this embodiment, the characteristics of simple, non-invasive skin electrical signal acquisition and easy long-term continuous monitoring make it an ideal signal for real-time monitoring of psychological stress. Through the combination of modern computer processing technology, the method of the present invention can achieve rapid and efficient real-time assessment of psychological stress, and is applicable to various actual application scenarios, such as personal health management, workplace stress monitoring, etc., and has broad application prospects.

[0045] Advantages of additional aspects of the present invention will be given in part in the following description, become apparent in part from the following description, or be learned through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The accompanying drawings forming a part of this specification are used to provide a further understanding of the present invention. The schematic embodiments and descriptions thereof of the present invention are used to explain the present invention and do not constitute an improper limitation to the present invention.

[0047] Figure 1 It is a schematic flow chart of a method for psychological stress assessment based on skin electrical signal image coding according to Embodiment 1 of the present invention;

[0048] Figure 2 It is a result diagram of different coded images of the time domain and frequency spectrum sequences of the skin electrical signal according to Embodiment 1 of the present invention;

[0049] Figure 3 It is a schematic structural diagram of a system for psychological stress assessment based on skin electrical signal image coding according to Embodiment 2 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0050] It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

[0051] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present invention.

[0052] In the case of no conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

[0053] Embodiment 1

[0054] This embodiment discloses a method for psychological stress assessment based on skin electrical signal image coding.

[0055] To more clearly illustrate this embodiment, the implementation process of psychological stress assessment based on skin electrical signal image coding can be specifically described as follows:

[0056] A psychological stress assessment method based on skin electro-signal image coding is provided, including:

[0057] S1. Collect the skin electro-signal of the tested person, preprocess the skin electro-signal, and obtain the time series of the skin electro-signal;

[0058] S2. Perform Fourier transform on the time series of the skin electro-signal to obtain the frequency spectrum sequence of the skin electro-signal;

[0059] S3. Respectively use Markov transition field, Gram angle and field, Gram angle difference field and recurrence plot to code the time series and frequency spectrum sequence of the skin electro-signal, and obtain a variety of different coded images;

[0060] S4. Adopt texture feature extraction technology to extract texture features from a variety of different coded images, and obtain multiple texture features;

[0061] S5. Perform feature selection on multiple texture features, construct an optimal feature set, and input the optimal feature set into a support vector machine to obtain the evaluation result of the psychological stress state.

[0062] As Figure 1 shown, in step S1, collect the skin electro-signal of the tested person, preprocess the skin electro-signal, and obtain the time series of the skin electro-signal.

[0063] Collect the skin electro-signal of the tested person, and preprocess the skin electro-signal specifically as follows:

[0064] (1) Downsample the skin electro-signal to obtain the downsampled skin electro-signal.

[0065] In this embodiment, downsampling is adopted to reduce the sampling frequency of the skin electro-signal to 2 Hz.

[0066] (2) Adopt a psychophysiological signal processing model to perform component decomposition and noise removal on the downsampled skin electro-signal.

[0067] In this embodiment, the cvxEDA model is adopted to perform component decomposition and noise removal on the skin electro-signal, retain the phase component therein, and after this processing, the time series of the skin electro-signal for subsequent analysis is obtained.

[0068] The cvxEDA model simply but physiologically reasonably represents the skin electro-signal (y) as the sum of three components: the main frequency component (t), the phase component (r), and an independent and identically distributed zero-mean Gaussian noise term (ε) containing all measurement errors and artifacts. The formula is:

[0069] y = r + t + ε (1)

[0070] The cvxEDA model is inspired by physiology and comprehensively explains electrodermal signals through methods based on Bayesian statistics, convex optimization, and sparsity. Compared with traditional filtering methods, cvxEDA can strictly and robustly decompose the signal into the main frequency component and the phase component, automatically eliminate noise, and overcome common problems related to phase response overlap.

[0071] As Figure 1 shown, in step S2, the time series of the electrodermal signal is Fourier-transformed to obtain the frequency spectrum sequence of the electrodermal signal.

[0072] In this embodiment, time-domain sequence analysis provides a direct representation of the signal over time.

[0073] Frequency-domain sequence analysis, on the other hand, transforms the electrodermal signal to the frequency domain through Fourier transform to analyze the frequency components of the signal, revealing the structure of the frequency distribution in the signal, and is used to analyze and understand the frequency-domain change pattern of the electrodermal signal under psychological stress conditions.

[0074] As Figure 1 shown, in step S3, the Markov transition field, Gram angle and field, Gram angle difference field, and recurrence plot are respectively used to encode the time series and frequency spectrum sequence of the electrodermal signal to obtain a variety of different encoded images.

[0075] In this embodiment, the Markov transition field, Gram angle and field, Gram angle difference field, and recurrence matrix are first constructed, and then the time series and frequency spectrum sequence of the electrodermal signal are encoded according to the Markov transition field, Gram angle and field, Gram angle difference field, and recurrence matrix.

[0076] S3-1. Construct the Markov transition field matrix.

[0077] In this embodiment, the Markov transition field matrix is first constructed, and then the time series and frequency spectrum sequence of the electrodermal signal are encoded according to the Markov transition field matrix.

[0078] Constructing the Markov transition field based on the time series and frequency spectrum sequence of the electrodermal signal, the specific process is as follows:

[0079] (1) Discretize the time and frequency spectrum sequences of the electrodermal signal using quantile intervals respectively.

[0080] For the sequence X = {x1, x2, …, x n}, map each value x t in the sequence to the corresponding quantile interval q i so that the Q quantile intervals have the same area under the Gaussian curve.

[0081] Among them, sequence X generally refers to the time series or spectrum sequence of electrodermal signals, and the two sequences are operated using the same steps.

[0082] (2) Calculate the transition probability between each interval to obtain the Markov transition matrix.

[0083] Calculating the transition probability of each quantile interval can obtain a Markov transition matrix W with a dimension of Q×Q. The formula is:

[0084]

[0085] w ij = p{x t ∈q i ||x t-1 ∈q j}; (2)

[0086] Among them, Q represents the number of quantile intervals, w ij represents the element in the i-th row and j-th column of the Markov transition matrix, q i and q j represent two different quantile intervals, x t-1 and x t represent two adjacent points in the sequence. w ij can be calculated as the probability that the point currently in the interval q i will appear in the interval q j at the next moment. w ij satisfies

[0087] (3) Based on the Markov transition matrix, by adding the correlation between each quantile and the time step, obtain the Markov transition field;

[0088] According to formula (2), by adding the correlation between each quantile and the time step, further obtain the Markov transition field M. The formula is:

[0089]

[0090] Among them, n is the length of the sequence, w ij represents the transition probability between two quantile intervals q i and q j , q i and q j represent two different quantile intervals.

[0091] After the above steps, the Markov transition field can clearly express the state transition patterns in the time series and spectrum sequence of electrodermal signals, and capture information such as state changes and transition rules in the sequence.

[0092] S3-2. Construct the Gram Angular Field (GAF) and Gram Angular Difference Field (GADF) matrices.

[0093] In this embodiment, the GAF matrix is first constructed, and then the time series and frequency spectrum series of the skin electrical signals are encoded based on the GAF.

[0094] The construction of the GAF based on the time series and frequency spectrum series of the skin electrical signals is as follows:

[0095] (1) Scale and transform the time series and frequency spectrum series of the skin electrical signals into the polar coordinate system respectively.

[0096] For the sequence X = {x1, x2,..., x n}, scale it to the interval [-1, 1] using the formula:

[0097]

[0098] After scaling the sequence transform it to the polar coordinate system using the formula:

[0099]

[0100] where

[0101] represents the i-th point in the sequence , θ i represents the angle of this point transformed into the polar coordinate system, r i represents the radial distance, t i is the timestamp, and n is a constant factor, usually taking the value of the length of the encoded sequence, so as to transform it into the unit circle.

[0102] (2) Calculate the cosine values of the sum of angles and the sine values of the difference of angles between points in the polar coordinate system to obtain the GAF and GADF.

[0103] The formulas for the GAF and GADF are respectively:

[0104]

[0105] where GASF and GADF represent the GAF and GADF respectively.

[0106] After the above steps, the GAF and GADF can generate images through the angular information of the time series and frequency domain series of the skin electrical signals in the polar coordinate system, which can capture the dynamic complex details in the sequence while maintaining the time correlation and amplitude information of the sequence without loss.

[0107] S3-3. Construct the recurrence matrix.

[0108] In this embodiment, a recurrence matrix is first constructed, and then the time series and frequency spectrum series of the electro-dermal signals are encoded according to the recurrence matrix.

[0109] Constructing a recurrence matrix based on the time series and frequency spectrum series of the electro-dermal signals, the specific process is as follows:

[0110] (1) Map the time series and frequency spectrum series of the electro-dermal signals into a high-dimensional phase space, and each point in the series is mapped into a high-dimensional vector;

[0111] Map the sequence X = {x1, x2,..., x n} into an m-dimensional reconstructed phase space, and the formula is:

[0112]

[0113] where m is the embedding dimension and τ is the time delay.

[0114] (2) Calculate the distance between every two points in the high-dimensional space to obtain the recurrence matrix.

[0115] Calculate the distance between two points X i and X j in the m-dimensional space to obtain the proximity of the two points X i and X j . According to the proximity, calculate the recurrence matrix D, and the formula is:

[0116] D i,j = ∥X i - X j ∥(8)

[0117] where ∥·∥ is the norm, and usually the Euclidean distance is taken.

[0118] After the above steps, the recurrence graph captures the complex dynamic patterns and non-linear dependencies of the sequence through the recurrence relationship, is applicable to the analysis of complex and non-stationary sequences, and can be used to explore the periodicity, hierarchy and chaos of the time series and frequency spectrum series of the electro-dermal signals.

[0119] S3-4. Image-encode the time series and frequency spectrum series of the electro-dermal signals according to the Markov transition field, Gram angle field, Gram angle difference field and recurrence matrix.

[0120] In the matrices of the four encoding methods obtained according to formulas (3), (6) and (8), the value of each point can be regarded as a pixel point. As Figure 2 shown, scale the values of the elements in the matrix to 0-255 so that they correspond to the pixel values in the image, thereby converting it into a grayscale image, realizing the encoding of a one-dimensional sequence into a two-dimensional image, and 8 image encodings can be obtained.

[0121] As shown Figure 1 in FIG., in step S4, texture feature extraction technology is used to extract texture features from a variety of different encoded images, obtaining multiple texture features.

[0122] In this embodiment, eight kinds of image encodings are used to extract texture features by using a variety of texture feature extraction technologies, and a total of 256 kinds of texture features are obtained.

[0123] Specifically, in this embodiment, three texture feature extraction technologies, local binary pattern, gray-level co-occurrence matrix, and first-order statistics, are used to extract texture features of images from different angles.

[0124] After the above steps, the encoded images of the skin electrical signal are processed by a variety of texture feature extraction technologies to comprehensively mine the advanced complex features in the time-domain and frequency-domain encoded images. It can optimize the capture of the time-domain and frequency-domain information of the skin electrical signal, provide a richer information source for the recognition of psychological stress, and thus effectively improve the robustness of psychological stress assessment.

[0125] S4-1. Use the local binary pattern to extract the texture features of the image.

[0126] The local binary pattern (LBP) is a local texture feature description operator that locally binaryizes the difference between the gray value of the central pixel and the gray values of its adjacent pixels. Given a pixel point (x c , y c ), its LBP code can be expressed as:

[0127]

[0128] where g c is the gray value of the central pixel point (x c , y c ), g p is the gray value of the neighborhood pixels on the circumference with a radius of R, P is the number of neighborhood pixels, and s(·) is the sign function.

[0129] To achieve rotation invariance, the rotation invariant pattern of LBP is expressed as:

[0130]

[0131] where U(LBP P,R ) ≤ 2 means that the binary number corresponding to the LBP code contains at most two jumps from 1 to 0 or from 0 to 1.

[0132] From LBP P,R to The mapping reduces the number of LBP eigenvalue types to P + 2 types.

[0133] Specifically, set P to 8. Through the local binary pattern, each encoded image can obtain 10 LBP features, and a total of 80 features are obtained from the 4 encoded images of the skin electrical signal time and frequency spectrum sequences.

[0134] After the above steps, the local binary pattern describes the texture through the gray-scale changes in the local neighborhood, can effectively capture the local texture features in the encoded image, and can effectively express the subtle changes in the texture. At the same time, the LBP method has rotational invariance, so it has strong robustness to image rotation.

[0135] S4-2. Extract the texture features of the image using the gray-level co-occurrence matrix.

[0136] The gray-level co-occurrence matrix (GLCM) is a method for analyzing image texture based on pixels. The GLCM is obtained by calculating the number of occurrences of the gray-scale difference between two pixel points in a specific direction θ and distance d in the image, and then features are extracted from the matrix to characterize the image texture.

[0137] Specifically, to improve the robustness to small-scale changes, the distance parameter d is set to 2. To comprehensively capture the texture characteristics in different directions, θ is selected as 4 directions: 0°, 45°, 90°, and 135°. Calculate the contrast, correlation, energy, and homogeneity of the GLCM in 4 different directions respectively. The formulas are as follows:

[0138] Contrast=∑ i,j |i - j| 2 p(i,j),

[0139]

[0140] Energy=∑ i,j p(i,j) 2 ,

[0141]

[0142] where p(i,j) represents the value at (i,j) in the GLCM, and μ and σ represent the mean and standard deviation respectively.

[0143] After passing through the gray-level co-occurrence matrix, a total of 128 GLCM features are obtained from the 4 encoded images of the time series and frequency spectrum sequence of the skin electrical signal.

[0144] After the above steps, the gray-level co-occurrence matrix extracts the global texture information of the encoded image by calculating the spatial relationship between pixel gray-level pairs in the image, including texture contrast, homogeneity, energy, etc. It can comprehensively describe more complex texture patterns, especially the directionality and roughness of the texture.

[0145] S4-3. Extract the texture features of the image using first-order statistics.

[0146] The first-order statistics (FOS) features are basic statistical metrics calculated from the pixel values of the image without considering the spatial relationship between pixels, and can be used to describe the gray-level distribution of the image texture.

[0147] Specifically, six features including mean, variance, kurtosis, skewness, energy, and entropy were calculated for the gray-level matrix of each encoded image, and a total of 48 FOS features were obtained from the four types of time-domain and frequency-domain sequence encoded images.

[0148] After the above steps, the first-order statistics can capture the overall gray-level distribution of the image, can well represent the global information such as the brightness and contrast of the encoded image, and is simple and fast to calculate.

[0149] In summary, the combination of the three texture feature extraction methods can comprehensively characterize the texture features in the image from different dimensions. Through this multi-angle feature extraction method, more abundant information is obtained from the time-domain and frequency-domain encoded images, providing a comprehensive and detailed feature basis for psychological stress assessment.

[0150] As Figure 1 shown, in step S5, multiple texture features are selected, an optimal feature set is constructed, and the optimal feature set is input into the support vector machine to obtain the evaluation result of the psychological stress state.

[0151] The specific process is as follows: (1) Use the recursive feature elimination method based on the support vector machine. By repeatedly constructing the support vector machine classification model, the feature with the lowest score is removed in each iteration, and finally the optimal feature set is selected;

[0152] (2) Input the optimal feature set into the support vector machine to obtain the evaluation result of the psychological stress state.

[0153] When constructing the psychological stress assessment model based on the support vector machine, 5-fold cross-validation is used to evaluate the performance of the model to ensure the stability and reliability of the results.

[0154] The image encoding method in this embodiment is used for pressure assessment comparison with the traditional feature extraction method. Table 1 shows the evaluation results. It can be seen from Table 1 that the image encoding method in this embodiment is far superior to the traditional feature extraction method in terms of accuracy, sensitivity, evaluation precision, and F1 score.

[0155] Table 1. Pressure assessment results using the image encoding method and traditional feature extraction method of the present invention

[0156]

[0157] Through the above steps, rapid and efficient real-time assessment of psychological pressure can be achieved, which is applicable to various actual application scenarios, such as personal health management, workplace stress monitoring, etc., and has broad application prospects.

[0158] Embodiment 2

[0159] The purpose of this embodiment is to provide a psychological pressure assessment system based on skin electrical signal image encoding, including:

[0160] A signal processing module, which is used to collect the skin electrical signal of the tested person, preprocess the skin electrical signal to obtain the time series of the skin electrical signal; perform Fourier transform on the time series of the skin electrical signal to obtain the frequency spectrum series of the skin electrical signal;

[0161] An image encoding module for the sequence, which is used to encode the time series and frequency spectrum series of the skin electrical signal respectively using Markov transition field, Gram angle and field, Gram angle difference field, and recurrence plot to obtain a variety of different encoded images;

[0162] An image feature extraction module, which is used to extract texture features from a variety of different encoded images using texture feature extraction technology to obtain multiple texture features;

[0163] A feature selection and classification module, which is used to perform feature selection on multiple texture features, construct an optimal feature set, and input the optimal feature set into a support vector machine to obtain the evaluation result of the psychological pressure state.

[0164] Based on providing a psychological pressure assessment system based on skin electrical signal image encoding, the method steps in Embodiment 1 are implemented.

[0165] Embodiment 3

[0166] The purpose of this embodiment is to provide a computing device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the above method are implemented.

[0167] Embodiment 4

[0168] The purpose of this embodiment is to provide a computer-readable storage medium.

[0169] A computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, it executes the steps of the above method.

[0170] Each step involved in the device of the above embodiment corresponds to the first method embodiment, and the specific implementation manner can refer to the relevant description part of the first embodiment. The term "computer-readable storage medium" should be understood to include a single medium or multiple media including one or more instruction sets; it should also be understood to include any medium that can store, encode, or carry an instruction set for execution by a processor and enable the processor to execute any method in the present invention.

[0171] Those skilled in the art should understand that the above-mentioned modules or steps of the present invention can be implemented by a general-purpose computer device. Optionally, they can be implemented by program codes executable by a computing device, so that they can be stored in a storage device and executed by the computing device, or they can be separately fabricated into individual integrated circuit modules, or multiple modules or steps among them can be fabricated into a single integrated circuit module for implementation. The present invention is not limited to any specific combination of hardware and software.

[0172] Although the specific implementation manner of the present invention has been described above in conjunction with the drawings, it is not a limitation to the protection scope of the present invention. Those skilled in the art should understand that based on the technical solution of the present invention, various modifications or deformations that can be made without creative efforts by those skilled in the art are still within the protection scope of the present invention.

Claims

1. A psychological stress assessment method based on skin electrical signal image coding, characterized in that: include: Collecting the test subject's skin electrical signals, preprocessing the skin electrical signals, and obtaining a time series of the skin electrical signals; Performing Fourier transform on the time series of the skin electrical signal to obtain the spectrum series of the skin electrical signal; Markov transition field, Gram angle sum field, Gram angle difference field and recursive graph are used to encode the time series and spectrum series of skin electrical signals respectively, and various encoded images are obtained; The texture feature extraction technology is used to extract texture features from a variety of different coded images to obtain multiple texture features; Feature selection is performed on multiple texture features to construct the optimal feature set, which is then input into a support vector machine to obtain the evaluation result of the psychological stress state.

2. A method for evaluating psychological stress based on skin electrical signal image coding as claimed in claim 1, characterized in that: Preprocess the skin electrical signals, specifically: Downsampling the skin electrical signal to obtain a downsampled skin electrical signal; The psychophysiological signal processing model is used to decompose the components and remove noise from the downsampled skin electrical signals.

3. The method for psychological stress assessment based on skin electrical signal image coding according to claim 1, characterized in that: The Markov transition field is used to encode the time series and spectrum series of the skin electrical signal. The specific process is as follows: The temporal and spectral series of the skin electrical signal are discretized using quantile intervals respectively; Calculate the transition probability between each interval and obtain the Markov transition matrix; Based on the Markov transition matrix, the Markov transition field matrix is ​​obtained by adding the correlation between each quantile and the time step; Image encoding is performed on the time series and spectrum series of skin electrical signals according to the Markov transition field matrix.

4. The method for psychological stress assessment based on skin electrical signal image coding according to claim 1, characterized in that: The time series and spectrum series of the skin electrical signal are encoded using the Gram angle sum field and the Gram angle difference field. The specific process is as follows: The time series and spectrum series of the skin electrical signal are scaled and converted into polar coordinate system respectively; By calculating the cosine of the angle sum and the sine of the angle difference between each point in the polar coordinate system, the Gram angle sum field and the Gram angle difference field matrices are obtained; Image encoding is performed on the time series and spectrum series of skin electrical signals according to the Gram angle sum field and Gram angle difference field matrices.

5. The method for psychological stress assessment based on skin electrical signal image coding according to claim 1, characterized in that: The recursive graph is used to encode the time series and spectrum series of the skin electrical signal. The specific process is as follows: The time series and spectrum series of skin electrical signals are mapped into a high-dimensional phase space, and each point in the sequence is mapped into a high-dimensional vector; Calculate the distance between every two points in high-dimensional space and get the recursive matrix. According to the recursive matrix, the time series and spectrum series of the skin electrical signal are image encoded.

6. The method for psychological stress assessment based on skin electrical signal image coding according to claim 1, characterized in that: Texture feature extraction techniques include: local binary pattern feature extraction, gray-level co-occurrence matrix feature extraction and first-order statistics feature extraction.

7. The method for psychological stress assessment based on skin electrical signal image coding according to claim 1, characterized in that: Feature selection is performed on multiple texture features to construct the optimal feature set, which is then input into the support vector machine to obtain the evaluation results of the psychological stress state. The specific process is as follows: The recursive feature elimination method based on support vector machine is adopted. By repeatedly building the support vector machine classification model, the feature with the lowest score is removed in each iteration, and finally the optimal feature set is screened out; The optimal feature set is input into the support vector machine to obtain the evaluation results of the psychological stress state.

8. A psychological stress assessment system based on skin electrical signal image coding, characterized in that: include: A signal processing module is used to collect the skin electrical signals of the tested person, pre-process the skin electrical signals, and obtain a time series of the skin electrical signals; Performing Fourier transform on the time series of the skin electrical signal to obtain the spectrum series of the skin electrical signal; A sequence image encoding module is used to encode the time series and spectrum series of the skin electrical signal using Markov transition field, Gram angle sum field, Gram angle difference field and recursive graph to obtain a variety of different encoded images; An image feature extraction module is used to extract texture features from a variety of different coded images using texture feature extraction technology to obtain multiple texture features; The feature selection and classification module is used to select multiple texture features, construct an optimal feature set, and input the optimal feature set into a support vector machine to obtain an evaluation result of the psychological stress state.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method described in any one of claims 1 to 7 are performed.

Images