HRV-based nonlinear chaotic feature extraction algorithm

Abstract

The invention provides a nonlinear chaotic feature extraction algorithm based on HRV. The nonlinear chaotic feature extraction algorithm comprises the steps of obtaining a scalar time sequence corresponding to a heart rate system data observation value; calculating a two-dimensional scatter diagram corresponding to the single rotation angle based on the scalar time sequence; based on the coordinate values corresponding to the scatter points in the two-dimensional scatter diagram, the parting scale of the sandbox to be cut is determined; based on the coordinate values corresponding to the scatter points in the two-dimensional scatter diagram and the parting scale of the sandbox to be cut, the first number of times of equal cutting of the sandbox to be cut is determined; based on the first cutting times, determining a scatter number statistical array; judging whether rotation angles which do not participate in calculation exist in the first preset number of rotation angles or not; if not, obtaining a first preset number of scattered point number statistical array, and executing S7; and based on the first preset number of scatter number statistical array, determining a heart rate system chaos state judgment slope. The algorithm provided by the invention is simple in operation, does not need complex phase space reconstruction, and does not occupy too many operation resources.

Description

Technical Field

[0001] The present application relates to the field of heart rate data analysis, and in particular to a nonlinear chaotic feature extraction algorithm based on HRV. Background Art

[0002] The existing system dimensions for characterizing heart rate variability (HRV) are divided into time domain indicators, frequency domain indicators, time-frequency domain indicators and nonlinear indicators. Nonlinear indicators include: Poincare map, approximate entropy, sample entropy, multi-scale entropy, fuzzy entropy, Shannon entropy, fractal dimension, correlation dimension, etc. HRV indicators can evaluate and monitor fatigue.

[0003] There are many existing HRV nonlinear analysis methods, but some nonlinear analysis methods (dimensionality, some entropy indicators) have limitations in feature extraction of HRV and are not very effective in fatigue classification. At the same time, some nonlinear analysis methods are relatively complex to calculate. Summary of the invention

[0004] In order to overcome the above technical defects, the present application provides a calculation method.

[0005] To achieve the above objectives, this application is implemented according to the following technical solutions:

[0006] A nonlinear chaotic feature extraction algorithm based on HRV, characterized by comprising:

[0007] Step S1: Obtain a scalar time series corresponding to the heart rate system data observation value;

[0008] Step S2: Based on the scalar time series, a two-dimensional scatter plot corresponding to a single rotation angle is calculated, wherein the single rotation angle is any rotation angle of a first preset number of rotation angles, and the first preset number of rotation angles are selected at equal intervals from (π / 5, 4π / 5);

[0009] Step S3: based on the coordinate values ​​corresponding to the scattered points in the two-dimensional scatter plot, determining the sandbox to be cut and the type scale of the sandbox to be cut; wherein the sandbox to be cut is a sandbox covering all the scattered points in the two-dimensional scatter plot;

[0010] Step S4: determining a first number of cuts for evenly cutting the sandbox to be cut based on the coordinate values ​​corresponding to the scattered points in the two-dimensional scatter plot and the type scale of the sandbox to be cut;

[0011] Step S5: Based on the first cutting times, determining a scatter point number statistical array corresponding to the first cutting times; wherein the scatter point number statistical array is an array containing the number of scatter points in the sandbox after the cutting;

[0012] Step S6: determining whether there are any rotation angles not involved in the calculation among the first preset number of rotation angles;

[0013] If yes, repeat the above steps S2-S5;

[0014] If not, a first preset number of scatter point statistics array is obtained, and S7 is executed;

[0015] Step S7: Determine the chaotic state judgment slope of the heart rate system based on the first preset number of scatter point statistical arrays.

[0016] Optionally, determining a two-dimensional scatter plot corresponding to a single rotation angle based on the scalar time series includes:

[0017] Based on the scalar time series and the single rotation angle; then calculating a plurality of pq trajectory points corresponding to the single rotation angle;

[0018] Based on the plurality of pq trajectory points, a two-dimensional scattergram corresponding to a single rotation angle is determined.

[0019] Optionally, the method of determining the sandbox to be cut and the type scale of the sandbox to be cut based on the coordinate values ​​corresponding to the scattered points in the two-dimensional scatter plot comprises:

[0020] Based on the scattered points in the two-dimensional scatter plot, determining the maximum and minimum values ​​of the p coordinates and the maximum and minimum values ​​of the q coordinates in the pq trajectory points corresponding to the two-dimensional scatter plot;

[0021] Based on the maximum and minimum values ​​of the p coordinate, determine the length of the sandbox parting scale to be cut;

[0022] Based on the maximum and minimum values ​​of the q coordinate, determine the width of the sandbox parting scale to be cut;

[0023] Based on the minimum value of the p coordinate and the minimum value of the q coordinate, determining the starting point of the sandbox to be cut in the two-dimensional scatter plot;

[0024] The sandbox to be cut is determined based on the starting point of the sandbox to be cut in the two-dimensional scatter plot, the length of the parting scale of the sandbox to be cut, and the width of the parting scale of the sandbox to be cut.

[0025] Optionally, determining a first number of cutting times for cutting the sandbox to be cut based on the coordinate values ​​corresponding to the scattered points in the two-dimensional scatter plot and the type scale of the sandbox to be cut includes:

[0026] Based on the pq trajectory points corresponding to the two-dimensional scatter plot, determining a first minimum distance value between a plurality of p coordinates in a p-coordinate set and a second minimum distance value between a plurality of p coordinates in a q-coordinate set;

[0027] Based on a first ratio of the length of the parting scale of the sandbox to be cut and the first minimum distance value, and a second ratio of the width of the parting scale of the sandbox to be cut and the second minimum distance value, a first number of cutting times for evenly cutting the sandbox to be cut is determined.

[0028] Optionally, the step of determining a scatter point number statistical array corresponding to the first number of cutting times based on the first number of cutting times includes:

[0029] Taking the first number of cutting times as the first number of cutting times, the sandbox to be cut is cut equally in sequence in descending order with 1 as the tolerance, and the number of scattered points contained in a single sandbox after each equal cutting is counted respectively to obtain a statistical array of the number of scattered points corresponding to the first number of cutting times.

[0030] Optionally, the determining the chaotic state judgment slope of the heart rate system based on the first preset number of scatter point number statistical array includes:

[0031] Based on the first preset number of scattered point count arrays, a two-dimensional matrix corresponding to the first preset number of rows is obtained;

[0032] Based on the median of multiple elements in each column of the two-dimensional matrix, a chaotic state judgment slope of the heart rate system is determined.

[0033] Optionally, determining the chaotic state judgment slope of the heart rate system based on the median of a plurality of elements in each column of the two-dimensional matrix includes:

[0034] Based on the median of multiple elements in each column of the two-dimensional matrix, determine the first number of cuts corresponding to the median and the length of the sandbox classification scale to be cut corresponding to the number of rows where the median is located;

[0035] Taking logarithms of the median and the ratio of the length of the sandbox partitioning scale to be cut to the first number of cutting times, respectively, to obtain the logarithm-taking median and the logarithm-taking ratio of the length of the sandbox partitioning scale to be cut to the first number of cutting times;

[0036] The least square method is adopted to fit the median of the logarithm and the ratio of the length of the sandbox type scale to be cut of the logarithm and the first cutting number to determine the chaotic state judgment slope of the heart rate system.

[0037] In a second aspect, the present application provides a computer device, comprising a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, at least one program, a code set or an instruction set is loaded and executed by the processor to implement the HRV-based nonlinear chaotic feature extraction algorithm as described in any one of the first aspects.

[0038] In a third aspect, the present application provides a computer-readable storage medium, wherein the readable storage medium stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by a processor to implement the HRV-based nonlinear chaotic feature extraction algorithm as described in any one of the first aspects.

[0039] This application has the following beneficial effects:

[0040] The algorithm provided in this application provides a new perspective for HRV analysis. Different from phase space reconstruction, it constructs a HRV nonlinear index based on the idea of ​​box counting dimension by performing a series of rotations and translations on one-dimensional time series data to quantify the degree of chaos of the system, thereby extracting the chaotic characteristics of HRV. The algorithm is simple to operate, does not require complex phase space reconstruction, does not occupy too many computing resources, and has a certain improvement in classification accuracy.

[0041] In addition to the above-described purposes, features and advantages, the present application has other purposes, features and advantages. The present application will be further described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0043] Figure 1 It is a schematic diagram of a nonlinear chaotic feature extraction algorithm process based on HRV provided in an embodiment of the present application.

[0044] Figure 2 This is a trajectory diagram corresponding to a 0-1test method provided in an embodiment of the present application.

[0045] Figure 3 It is a schematic diagram of a two-dimensional scatter plot provided in an embodiment of the present application.

[0046] Figure 4 It is a schematic diagram of a sandbox to be cut provided in an embodiment of the present application.

[0047] Figure 5This is a schematic diagram of a sandbox to be cut after one cutting provided in an embodiment of the present application. DETAILED DESCRIPTION

[0048] The embodiments of the present application are described in detail below with reference to the accompanying drawings; however, the present application can be implemented in many different ways as defined and covered by the claims.

[0049] There are many existing HRV nonlinear analysis methods, but some nonlinear analysis methods (dimensionality, some entropy indicators) have limitations in feature extraction of HRV and are not very effective in fatigue classification. At the same time, some nonlinear analysis methods are relatively complex to calculate.

[0050] In order to solve the above technical problems, Figure 1 As shown, the present application proposes a nonlinear chaotic feature extraction algorithm based on HRV, including:

[0051] Step S1: Obtain a scalar time series corresponding to the heart rate system data observation value;

[0052] The scalar time series corresponding to the heart rate system data observation value refers to the sequence formed by the heart rate observation values ​​corresponding to each moment recorded in sequence over time in a specific heart rate monitoring system. Among them, each observation value is a separate scalar, representing the heart rate value at that moment. These values ​​are arranged in chronological order to form a time series, which can be used to analyze the changing trend, rhythm and other characteristics of the heart rate to assist medical diagnosis and health monitoring. The scalar time series corresponding to the heart rate system data observation value can be obtained using corresponding medical equipment or wearable devices, and this article does not make specific restrictions on this.

[0053] It should be noted that the heart rate system data here is for a single person.

[0054] Step S2: Based on the scalar time series, a two-dimensional scatter plot corresponding to a single rotation angle is calculated, wherein the single rotation angle is any rotation angle of a first preset number of rotation angles, and the first preset number of rotation angles are selected at equal intervals from (π / 5, 4π / 5);

[0055] This step in the present application mainly adopts the 0-1test method. Therefore, a brief introduction to the 0-1test method is given here: The basic idea of ​​0-1Test is to construct a two-dimensional Euclidean group SE(2). The elements of SE(2) form rigid displacements, i.e., translations and rotations, in the two-dimensional Euclidean plane. In principle, this plane has no topological relationship with the state space of the system dynamics. However, the parameters used to characterize the translation and rotation are always determined by the current state of the system. Therefore, there is a certain equivalence relationship between the dynamic evolution described by the SE(2) elements in the plane and the heart rate variability dynamics system under our study.

[0056] The 0-1Test method requires a scalar time series of N observations, s(n), where n = 1, 2, ... N, where s(n) is the one-dimensional observable value of the underlying dynamical system. The following formula (1) is used to obtain s(n):

[0057] S(n)=RR(n+1)-RR(n) (1)

[0058] Where RR represents the interval between two adjacent R peaks extracted from the ECG signal, in ms, and n=1, 2, ...N, where N=total number of samples - 1.

[0059] The 0-1 method constructs the SE(2) group element as shown in formula (2):

[0060] p n+1 =p n +s(n)cos cn,q n+1 =q n +s(n)sincn (2)

[0061] The evolution of any point in the (p,q) plane describes a trajectory, which is called the auxiliary trajectory. It reproduces the indirect or supplementary evolution of the real dynamics observed in the system. Let cn = φ n , the auxiliary trajectory involves an angular rotation φ n , relative to the point (p n ,q n ) is a circle with a radius of s(n) as its center, such as Figure 2 shown.

[0062] The rotation angle mentioned in this application is cn, and the uniform rate of increase is controlled by the c value. The single rotation angle in this application is the first preset number N C Any rotation angle among the rotation angles, the first preset number N C The rotation angle is selected in equal intervals from (π / 5, 4π / 5). The first preset number N in this application C , can be set according to actual needs, this article does not make specific restrictions on this.

[0063] After determining a single rotation angle and a scalar time series, the scalar time series is substituted into formula (2) to calculate a plurality of pq trajectory points corresponding to the single rotation angle. Then, based on the plurality of pq trajectory points, a two-dimensional scatter plot corresponding to the single rotation angle is determined on a two-dimensional plane, as shown in Figure 3 The scattered points in the scatter plot are equivalent to multiple pp trajectory points. In other words, the positions of multiple pq trajectory points are the positions of the scattered points in the scatter plot, and the two correspond one to one.

[0064] Step S3: based on the coordinate values ​​corresponding to the scattered points in the two-dimensional scatter plot, determining the sandbox to be cut and the type scale of the sandbox to be cut; wherein the sandbox to be cut is a sandbox covering all the scattered points in the two-dimensional scatter plot;

[0065] The sandbox to be cut in this application can be understood as a type containing the number of scattered points in a coordinate system with p as the horizontal coordinate and q as the vertical coordinate. The entire type can be regarded as a rectangle. Each time it is not cut, the sandbox to be cut is also a sandbox covering all the scattered points in the two-dimensional scatter plot. After determining the scattered points in the two-dimensional scatter plot, it is also possible to determine the multiple pq coordinate values ​​corresponding to the multiple scattered points in the pq coordinate system. Among the above multiple pq coordinate values, determine the maximum value p of the p coordinate max and the minimum value p min , and the maximum value q of the q coordinate max and the minimum value q min Then we can take the difference between the maximum and minimum values ​​of the p coordinate to get L P , so that L P As the length of the sandbox parting scale to be cut, the maximum and minimum values ​​of the q coordinate are subtracted to obtain L q , then L q As the width of the sandbox to be cut. The point corresponding to the minimum value of the p coordinate and the minimum value of the p coordinate is used as the starting point of the rectangle corresponding to the sandbox to be cut in the pq coordinate system, Lp is used as the length of the rectangle corresponding to the sandbox to be cut, and L q As the width of the rectangle corresponding to the sandbox to be cut, the sandbox to be cut can be determined, such as Figure 4 shown.

[0066] Step S4: determining a first number of cuts for evenly cutting the sandbox to be cut based on the coordinate values ​​corresponding to the scattered points in the two-dimensional scatter plot and the type scale of the sandbox to be cut;

[0067] In this application, the sandbox to be cut needs to be cut equally. The number of cuts is determined by the following calculation method, that is, the first minimum distance value between multiple p coordinates in the element set corresponding to the p coordinate is calculated, which is recorded as d Pmin In other words, find the distance values ​​between multiple p coordinates, and then select the one with the smallest distance value as the first minimum distance value. Then calculate the second minimum distance value between multiple q coordinates in the element set corresponding to the q coordinate, recorded as d qmin Then use the length L of the sandbox parting scale to be cut P Divide by d Pmin , using the width L of the sandbox parting scale to be cut q Divide by d qmin , where the larger value is used as the first cutting number n.

[0068] After determining the first cutting number n, the first cutting number n is used as the first cutting number, and the sandbox to be cut is cut equally in descending order with 1 as the tolerance, that is, .n, n-1, .... 0. Specifically, the length L of the parting scale of the sandbox to be cut is P and width L q , perform equal cutting in descending order of .n, n-1, ....0. Figure 5 As shown, it can be understood that Figure 4 On the basis of the sandbox, make a cut.

[0069] Step S5: Based on the first cutting times, determining a scatter point number statistical array corresponding to the first cutting times; wherein the scatter point number statistical array is an array containing the number of scatter points in the sandbox after the cutting;

[0070] Each time the sandbox is cut equally, 1 / n sandboxes can be obtained. At this time, the minimum value of p and q is used as the starting point, and the number of scattered points contained in each 1 / n sandbox is counted, which is recorded as Num i (n), Num i (n-1),…,Num i (0), i represents the number sequence corresponding to a single rotation angle. That is to say, after obtaining 1 / n cut sandboxes, each cut sandbox has a corresponding starting point p and q, and the number of scattered points contained in the sandbox with the smallest starting point p and q among the 1 / n sandboxes is counted.

[0071] Since the cuts are made n times here, n corresponding scattered point statistics are obtained. At this time, the n corresponding scattered point statistics are temporarily saved as a one-dimensional statistical array.

[0072] Step S6: determining whether there are any rotation angles not involved in the calculation among the first preset number of rotation angles;

[0073] If yes, repeat the above steps S2-S5;

[0074] If not, a first preset number of scatter point statistics array is obtained, and S7 is executed;

[0075] Since all rotation angles in the first preset number of rotation angles need to be calculated in this application, after calculating the scattered point number statistical array of a single rotation angle according to the above steps S2-S6 each time, whether there are rotation angles in the first preset number of rotation angles that are not involved in the calculation is determined to determine the subsequent operation.

[0076] If there is a rotation angle that has not been calculated, the rotation angle that has not been calculated is determined, and the rotation angle that has not been calculated is substituted into the above steps S2-S6 to calculate the scatter point number statistical array corresponding to the rotation angle.

[0077] If there is no rotation angle that is not involved in the calculation among the first preset number of rotation angles, it means that the first preset number of scattered point number statistical arrays are lost at this time, and then subsequent calculations are performed.

[0078] Step S7: Determine the chaotic state judgment slope of the heart rate system based on the first preset number of scatter point statistical arrays.

[0079] After obtaining the first preset number of scattered point count arrays, they are arranged in order to obtain a two-dimensional matrix corresponding to the first preset number as shown in Table 1, that is, the number of rows of the two-dimensional matrix is ​​the first preset number. Each element in the two-dimensional matrix corresponds to the number of scattered points contained in each cut sandbox under the corresponding number of cuts. And the first number of cuts n corresponding to each row is the same, and the length Lp of the corresponding sandbox type scale to be cut is also the same.

[0080] Table 1 Two-dimensional matrix corresponding to the first preset number of rows

[0081]

[0082] After obtaining the above two-dimensional matrix, the median of multiple elements in each column of the corresponding two-dimensional matrix can be determined, recorded as Num(r), Num(r-1)...Num(0). Then determine the first cutting number n corresponding to the median and the length L of the sandbox type scale to be cut corresponding to the row number of the median. P Take the logarithm of the median and get lnNum(n). Then treat the length L of the sandbox type scale to be cut P and the ratio of the first cutting times n, we get ln(L P / n). Using the least square method, for multiple lnNum(n) and ln(LP / n) is fitted to obtain a slope, which is the slope for judging the chaotic state of the heart rate system. Generally speaking, in a single individual, the smaller the slope, the more obvious the chaotic state of the heart rate system of the individual, and the more relaxed the person is; the larger the slope, the more vague the chaotic state of the heart rate system of the individual, and the more tired the person is.

[0083] The algorithm provided in this application provides a new perspective for HRV analysis. Different from phase space reconstruction, it constructs a HRV nonlinear index based on the idea of ​​box counting dimension by performing a series of rotations and translations on one-dimensional time series data to quantify the degree of chaos of the system, thereby extracting the chaotic characteristics of HRV. The algorithm is simple to operate, does not require complex phase space reconstruction, does not occupy too many computing resources, and has a certain improvement in classification accuracy.

[0084] In some embodiments, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor implements a nonlinear chaotic feature extraction algorithm based on HRV provided in the first aspect.

[0085] It will be appreciated by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In a hardware implementation, the division between the functional blocks / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or may be implemented as hardware, or may be implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable storage medium, which may include a computer-readable storage medium (or a non-transitory medium) and a communication medium (or a temporary medium).

[0086] As is known to those of ordinary skill in the art, the term computer-readable storage media includes volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules or other data). Computer-readable storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, it is known to those of ordinary skill in the art that communication media typically contain computer-readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transmission mechanism, and may include any information transfer medium.

[0087] Exemplarily, the computer-readable storage medium may be an internal storage unit of the network management device described in the foregoing embodiment, such as a hard disk or memory of the network management device. The computer-readable storage medium may also be an external storage device of the network management device, such as a plug-in hard disk, a smart memory card (SmartMediaCard, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc., equipped on the network management device.

[0088] In some embodiments, a device is provided, including a processor and a memory, wherein the memory is used to store a computer program; the processor is used to execute the computer program and implement a nonlinear chaotic feature extraction algorithm based on HRV provided in the first aspect of the present application when executing the computer program.

[0089] It should be understood that the processor may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0090] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A nonlinear chaotic feature extraction algorithm based on HRV, characterized in that: include: Step S1: Obtain a scalar time series corresponding to the heart rate system data observation value; Step S2: Based on the scalar time series, a two-dimensional scatter plot corresponding to a single rotation angle is calculated, wherein the single rotation angle is any rotation angle of a first preset number of rotation angles, and the first preset number of rotation angles are selected at equal intervals from (π / 5, 4π / 5); Step S3: based on the coordinate values ​​corresponding to the scattered points in the two-dimensional scatter plot, determining the sandbox to be cut and the type scale of the sandbox to be cut; wherein the sandbox to be cut is a sandbox covering all the scattered points in the two-dimensional scatter plot; Step S4: determining a first number of cuts for evenly cutting the sandbox to be cut based on the coordinate values ​​corresponding to the scattered points in the two-dimensional scatter plot and the type scale of the sandbox to be cut; Step S5: Based on the first cutting times, determining a scatter point number statistical array corresponding to the first cutting times; wherein the scatter point number statistical array is an array containing the number of scatter points in the sandbox after the cutting; Step S6: determining whether there are any rotation angles not involved in the calculation among the first preset number of rotation angles; If yes, repeat the above steps S2-S5; If not, a first preset number of scatter point statistics array is obtained, and S7 is executed; Step S7: Determine the chaotic state judgment slope of the heart rate system based on the first preset number of scatter point statistical arrays.

2. The method according to claim 1, characterized in that The step of determining a two-dimensional scatter plot corresponding to a single rotation angle based on the scalar time series comprises: Based on the scalar time series and the single rotation angle; then calculating a plurality of pq trajectory points corresponding to the single rotation angle; Based on the plurality of pq trajectory points, a two-dimensional scattergram corresponding to a single rotation angle is determined.

3. The method according to claim 2, characterized in that The method of determining the sandbox to be cut and the type scale of the sandbox to be cut based on the coordinate values ​​corresponding to the scattered points in the two-dimensional scatter plot comprises: Based on the scattered points in the two-dimensional scatter plot, determining the maximum and minimum values ​​of the p coordinates and the maximum and minimum values ​​of the q coordinates in the pq trajectory points corresponding to the two-dimensional scatter plot; Based on the maximum and minimum values ​​of the p coordinate, determine the length of the sandbox parting scale to be cut; Based on the maximum and minimum values ​​of the q coordinate, determine the width of the sandbox parting scale to be cut; Based on the minimum value of the p coordinate and the minimum value of the q coordinate, determining the starting point of the sandbox to be cut in the two-dimensional scatter plot; The sandbox to be cut is determined based on the starting point of the sandbox to be cut in the two-dimensional scatter plot, the length of the parting scale of the sandbox to be cut, and the width of the parting scale of the sandbox to be cut.

4. The method according to claim 3, characterized in that The method of determining a first number of cutting times for cutting the sandbox to be cut based on the coordinate values ​​corresponding to the scattered points in the two-dimensional scatter plot and the type scale of the sandbox to be cut includes: Based on the pq trajectory points corresponding to the two-dimensional scatter plot, determining a first minimum distance value between a plurality of p coordinates in a p-coordinate set and a second minimum distance value between a plurality of p coordinates in a q-coordinate set; Based on a first ratio of the length of the parting scale of the sandbox to be cut and the first minimum distance value, and a second ratio of the width of the parting scale of the sandbox to be cut and the second minimum distance value, a first number of cutting times for evenly cutting the sandbox to be cut is determined.

5. The method according to claim 4, characterized in that The step of determining a scatter point number statistical array corresponding to the first cutting times based on the first cutting times comprises: Taking the first number of cutting times as the first number of cutting times, the sandbox to be cut is cut equally in sequence in descending order with 1 as the tolerance, and the number of scattered points contained in a single sandbox after each equal cutting is counted respectively to obtain a statistical array of the number of scattered points corresponding to the first number of cutting times.

6. The method according to claim 5, characterized in that The determining the chaotic state judgment slope of the heart rate system based on the first preset number of scattered point number statistical arrays includes: Based on the first preset number of scattered point count arrays, a two-dimensional matrix corresponding to the first preset number of rows is obtained; Based on the median of multiple elements in each column of the two-dimensional matrix, a chaotic state judgment slope of the heart rate system is determined.

7. The method according to claim 6, characterized in that The method of determining the chaotic state judgment slope of the heart rate system based on the median of multiple elements in each column of the two-dimensional matrix includes: Based on the median of multiple elements in each column of the two-dimensional matrix, determine the first number of cuts corresponding to the median and the length of the sandbox classification scale to be cut corresponding to the number of rows where the median is located; Taking logarithms of the median and the ratio of the length of the sandbox partitioning scale to be cut to the first number of cutting times, respectively, to obtain the logarithm-taking median and the logarithm-taking ratio of the length of the sandbox partitioning scale to be cut to the first number of cutting times; The least square method is adopted to fit the median of the logarithm and the ratio of the length of the sandbox type scale to be cut of the logarithm and the first cutting number to determine the chaotic state judgment slope of the heart rate system.

8. A computer device, characterized in that: The computer device includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, at least one program, a code set or an instruction set is loaded and executed by the processor to implement the HRV-based nonlinear chaotic feature extraction algorithm as described in any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that: The readable storage medium stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by the processor to implement the HRV-based nonlinear chaotic feature extraction algorithm as described in any one of claims 1 to 7.

Images