WO2020114191A1

WO2020114191A1 - Blood oxygen saturation measurement confidence calculation method, system, and storage medium

Info

Publication number: WO2020114191A1
Application number: PCT/CN2019/116485
Authority: WO
Inventors: 林霖; 王涛
Original assignee: 深圳技术大学
Priority date: 2018-12-05
Filing date: 2019-11-08
Publication date: 2020-06-11
Also published as: CN109512392A

Abstract

Provided are a blood oxygen saturation measurement confidence calculation method, system, and storage medium, said method comprising: recording light intensity information by means of photoplethysmography, the light intensity information comprising red light and infrared light transmission intensity, and infrared light reflection intensity; establishing a linear regression model of a pulse blood oxygen signal characteristic value according to the light intensity information; calculating the confidence of a linear regression line in the linear regression model; according to the confidence of the linear regression line, obtaining the confidence level of the blood oxygen saturation measurement result. The light intensity information obtained by means of photoplethysmography is used for constructing a linear regression model; finally, the confidence of the blood oxygen saturation measurement result is obtained by means of a linear regression model, thereby achieving the reliability evaluation of the measured value, avoiding missed and false detections caused by interference factors such as motion pseudo-error and noise, and improving the reliability of blood oxygen saturation measurement results; the invention is more scientific and can be widely used in the field of signal processing technology.

Description

Calculation method, system and storage medium for blood oxygen saturation measurement confidence

Technical field

The invention relates to the technical field of signal processing, in particular to a calculation method, system and storage medium for blood oxygen saturation measurement confidence.

Background technique

The relaxation and contraction of the heart drives blood flow through the lungs, combining oxygen with reduced hemoglobin (Hb) to form oxyhemoglobin (Oxyhemoglobin, HbO2). Oxygen is released through blood to capillaries. Sufficient oxygen is the material basis to realize the metabolism of human tissue cells and maintain life activities. Blood oxygen saturation is an important physiological parameter that reflects the oxygen content in the blood, and it has a direct relationship with the respiratory system, circulatory system and cardiopulmonary function. At present, blood oxygen saturation is widely used in intensive care, home health care and high-risk occupations such as firefighters, pilots, etc. physical examination.

The detection methods of blood oxygen saturation can be divided into two types: invasive detection and non-invasive detection. Among them, the invasive blood oxygen saturation test mainly uses the Van Slyke pressure test method and the oxygen electrode method. The main method of noninvasive detection is photoplethysmography (PPG). The blood volume of the blood vessel changes with the diastole and contraction of the heart, resulting in different absorption rates of light, and the intensity of the reflected or transmitted light also changes in a pulsating cycle. The pulse wave oximeter uses photoelectric volume pulse wave tracing to record the intensity of the reflected or transmitted light of red light at 660 nm and infrared light at 940 nm, and then calculate the hemorrhagic oxygen saturation according to Lambert-Beer law. In actual measurement, the accurate calculation of the characteristic value R of the pulse blood oxygen signal is the key to the noninvasive detection of blood oxygen saturation based on the photoelectric volume pulse wave tracing method.

The traditional R value extraction method needs to decompose the pulse wave into two components: DC component, which reflects the absorption of light by HbO2 and Hb in the blood, and DC component reflects the non-blood tissues such as muscle, bone, fat, and water in the fingertips. Equal absorption of light. The calculation of the AC component usually uses the peak-valley method, that is, the difference between the peak value and the valley value within a pulse cycle is approximated to be the amplitude of the AC component. In the process of measurement and AC/DC decomposition, the interference and random noise introduced and generated will affect the accuracy of the R value calculated by the peak-to-valley method. It is usually necessary to use the average of multiple periodic peak-to-valley values to improve the accuracy, thereby affecting the real-time performance of the calculation . Someone has proposed a method of calculating the R value using a linear regression model. This method makes full use of the data of all sampling points, instead of relying solely on the peak and valley values of the pulse wave, which improves the stability of the calculation results. However, photoelectric detection is susceptible to the influence of the external light environment and the artifact of fingertip movements to cause changes in blood filling conditions and light transmission paths. The above factors will cause distortion of the measurement results, leading to missed detection and false detection. At present, the blood oxygen saturation tester lacks a scientific analysis of the reliability of the measurement results. How to reasonably evaluate the reliability of the measured values is an urgent problem to be solved in the industry.

Summary of the invention

In order to solve the above technical problems, an object of the present invention is to provide a method, system and storage medium for calculating blood oxygen saturation measurement confidence with high reliability.

The technical solution adopted in one aspect of the present invention is:

The calculation method of blood oxygen saturation measurement confidence includes the following steps:

Recording light intensity information by photoplethysmography, the light intensity information includes red light, infrared light transmitted light intensity and infrared light reflected light intensity;

According to the light intensity information, establish a linear regression model of pulse blood oxygen signal eigenvalues;

Calculate the confidence of the linear regression line in the linear regression model;

According to the confidence of the linear regression line, the confidence of the blood oxygen saturation measurement result is obtained.

Further, the step of recording light intensity information by photoplethysmography includes the following steps:

Obtain light intensity information analog signal by photoelectric volume pulse wave tracing method;

Convert light intensity information analog signal into light intensity information digital signal;

Wherein, the analog signal of light intensity information is a continuous signal, and the digital signal of light intensity information is a discrete signal.

Further, the step of establishing a linear regression model of the pulse blood oxygen signal characteristic value based on the light intensity information includes the following steps:

Calculate intermediate variables based on discrete light intensity information digital signals;

Establish linear regression equations based on intermediate variables;

Use the least square method for linear fitting, and calculate the parameters of the linear regression equation;

According to the calculated parameters, a linear regression model of pulse blood oxygen signal characteristic values is established.

Further, the step of calculating the confidence of the linear regression line in the linear regression model includes the following steps:

Through the linear regression model, calculate the sum of squared deviations from the total deviation;

Through the linear regression model, calculate the sum of squared residuals;

Calculate the regression sum of squares based on the sum of squared deviations from the total deviation and the sum of squared residuals;

Calculate the confidence of the linear regression line based on the sum of squared regressions and the sum of squared deviations from the mean deviation.

Further, the step of obtaining the confidence of the blood oxygen saturation measurement result according to the confidence of the linear regression line is specifically as follows:

Determine whether the confidence of the linear regression line is greater than the preset threshold. If it is, determine that the measurement result of blood oxygen saturation is reliable; otherwise, determine that the measurement result of blood oxygen saturation is not reliable.

The technical solution adopted by another aspect of the present invention is:

Calculation system for the confidence level of blood oxygen saturation measurement, including:

The recording module is used to record light intensity information by photoelectric volume pulse wave tracing, the light intensity information includes red light, infrared light transmitted light intensity and infrared light reflected light intensity;

The model building module is used to establish a linear regression model of the pulse blood oxygen signal characteristic value based on the light intensity information;

Calculation module, used to calculate the confidence of the linear regression line in the linear regression model;

The judgment module is used to obtain the confidence of the blood oxygen saturation measurement result according to the confidence of the linear regression line.

Further, the recording module includes:

The acquisition module is used to acquire the analog signal of light intensity information by photoelectric volume pulse wave tracing method;

The conversion module is used to convert the analog signal of the light intensity information into the digital signal of the light intensity information;

Further, the calculation module includes:

The first calculation unit is used to calculate the sum of squared deviations of the total deviation from the linear regression model;

The second calculation unit is used to calculate the sum of squared residuals through a linear regression model;

The third calculation unit is used to calculate the regression sum of squares based on the sum of squared deviations from the total deviation and the sum of squared residuals;

The fourth calculation unit is used to calculate the confidence of the linear regression line based on the sum of squared regressions and the sum of squared deviations of the total deviations.

The technical solution adopted by another aspect of the present invention is:

At least one processor;

At least one memory for storing at least one program;

When the at least one program is executed by the at least one processor, the at least one processor implements the calculation method of the blood oxygen saturation measurement confidence level.

The technical solution adopted by another aspect of the present invention is:

A storage medium in which instructions executable by a processor are stored, and when the instructions executable by the processor are executed by the processor, are used to perform the calculation method of the blood oxygen saturation measurement confidence level.

The beneficial effects of the present invention are: the present invention constructs a linear regression model through the light intensity information obtained by photoplethysmography, and finally obtains the confidence of the blood oxygen saturation measurement result through the linear regression model, thereby achieving the reliability of the measured value The sexual evaluation avoids the situation of missed detection and false detection caused by the influence of interference factors such as motion artifacts and noise, and improves the reliability of the blood oxygen saturation measurement result and is more scientific.

BRIEF DESCRIPTION

FIG. 1 is a flowchart of steps in an embodiment of the present invention;

2 is a schematic diagram of red light recorded according to an embodiment of the present invention;

3 is a schematic diagram of infrared light recorded in an embodiment of the present invention;

4 is a first schematic diagram of calculating an intermediate variable according to an embodiment of the present invention;

5 is a second schematic diagram of calculating an intermediate variable according to an embodiment of the present invention;

6 is a data scatter diagram of an embodiment of the present invention;

7 is a schematic diagram of a characteristic value of a pulse blood oxygen signal according to an embodiment of the present invention;

FIG. 8 is a schematic diagram of confidence coefficients according to an embodiment of the present invention.

detailed description

The present invention will be further explained and explained below with reference to the accompanying drawings and specific embodiments. The step numbers in the embodiments of the present invention are set for the convenience of explanation and description, and the order between the steps is not limited, and the execution order of the steps in the embodiments can be performed according to the understanding of those skilled in the art Adaptive adjustment.

Referring to FIG. 1, an embodiment of the present invention provides a method for calculating a blood oxygen saturation measurement confidence level, including the following steps:

S1. Record photointensity information by photoplethysmography, the photointensity information includes red light, infrared light transmitted light intensity and infrared light reflected light intensity;

As a further preferred embodiment of step S1, the step of recording light intensity information by photoplethysmography includes the following steps:

S11. Obtain light intensity information analog signal by photoelectric volume pulse wave tracing method;

S12. Convert the analog signal of light intensity information into a digital signal of light intensity information;

Specifically, in this embodiment, the transmission or reflected light intensity I _ir (t) of red light I _rd (t) having a wavelength of 660 nm and infrared light of 940 nm is recorded. The invention adopts photoelectric volume pulse wave tracing method, which generates red light with a wavelength of 660nm and infrared light with a wavelength of 940nm by the hardware system. The red light I _rd (t) recorded for this embodiment; FIG. 3 is the infrared light I _ir (t) recorded for this embodiment, where the light intensity signals shown in FIGS. 2 and 3 show periodic changes, which is It is caused by the difference in light absorption rate due to the expansion and contraction of the blood vessel blood volume.

S2. According to the light intensity information, establish a linear regression model of the characteristic value of the pulse blood oxygen signal;

As a further preferred embodiment of step S2, the step of establishing a linear regression model of the characteristic value of the pulse blood oxygen signal based on the light intensity information includes the following steps:

S21. Calculate the intermediate variable according to the discrete light intensity information digital signal;

In this embodiment, considering that I _rd (n) and I _ir (n) are discrete digital signals, the intermediate variable x(n)=I _rd (n)(I _ir (n)-I _ir ( n-1)) and y(n)=I _rd (n)(I _ir (n)-I _ir (n-1)), for convenience of writing, they are also denoted as x _n and y _n , and all time points The value of is written as a vector in the form of X=[x ₁ ,x ₂ ,...,x _n ] and Y=[y ₁ ,y ₂ ,...,y _n ].

S22. Establish a linear regression equation according to the intermediate variables;

In this embodiment, X is an independent variable and Y is a dependent variable, assuming that the linear regression equation is Y = b ₁ + b ₂ X, where b ₁ and b ₂ are undetermined parameters.

S23. Use the least square method to perform linear fitting and calculate the parameters of the linear regression equation;

In this embodiment, the least square method is used to linearly fit the data, and the parameters b ₁ and b _{2 are} calculated:

Among them, E(*) is the mean value, and the pulse blood oxygen signal characteristic value R is the parameter b ₂ .

S24. According to the calculated parameters, establish a linear regression model of the characteristic value of the pulse blood oxygen signal.

Specifically, the intermediate variables calculated in this embodiment are:

with

among them,

Is a derivative of I _rd (t),

Is the derivative of I _ir (t). Considering the actual hardware system, the light intensity information will be converted into a discrete digital signal after analog-to-digital conversion. In the present invention, vectors X and Y are used to represent the set of intermediate variables at n time points (discrete points), namely X = [x ₁ , x ₂ , L, x _n ], Y=[y ₁ , y ₂ , L, y _n ]. In addition, the characteristic value of pulse oximetry signal

If you define an intermediate variable

with

Then y=Rx. In order to solve the characteristic value of the pulse blood oxygen signal, the present invention makes full use of the obtained light intensity information, uses linear regression theory to fit the intermediate variables x and y, and calculates the confidence of the linear regression line according to the analysis of variance. As shown in Figure 4 according to the formula

The calculated intermediate variable x (denoted as X in Figure 4), as shown in Figure 5 is based on the formula

The calculated intermediate variable y (denoted as Y in Figure 5). Then, draw the data scatterplot with x as the abscissa and y as the ordinate, as shown in Figure 6. This embodiment assumes that the linear regression equation is y = b ₁ + b ₂ x, and the data is fitted according to the least squares rule to obtain b ₁ = 0.01, b ₂ = 0.55, and the pulse blood oxygen signal characteristic value R is the slope of the regression equation b ₂ .

S3. Calculate the confidence of the linear regression line in the linear regression model;

As a further preferred embodiment of step S3, the step of calculating the confidence of the linear regression line in the linear regression model includes the following steps:

S31. Calculate the sum of squared deviations of the total deviations through the linear regression model;

S32. Calculate the sum of squared residuals through a linear regression model;

S33. Calculate the sum of squared regressions based on the sum of squared deviations from the total deviation and the sum of squared residuals

S34. Calculate the confidence of the linear regression line based on the sum of squared regressions and the sum of squared deviations of the total deviations.

In this embodiment, the variance analysis method is used to calculate the confidence C of the linear regression line as a measure of the quality of the fit. The specific implementation method is as follows: the regression equation Y = b ₁ + b ₂ X is introduced to estimate the average level of Y corresponding to the X value, that is, Y represents the actual measured value.

First, calculate the sum of squared deviations from the total deviation of the dependent variable Y

This reflects the total variation of the dependent variable Y.

Next, calculate the sum of squared residuals of the dependent variable Y

This reflects the difference caused by X worth

Variation caused by the difference, in which

Represents the regression equation

The calculated regression value.

Then, calculate the regression square sum of the dependent variable Y. The regression square sum is the difference between the sum of the squared deviations of the total deviation and the sum of the squared residuals, that is, SSE=SST-SSR.

Finally, a confidence coefficient is calculated, the confidence coefficient C being the ratio between the sum of squared regressions and the sum of squared deviations from the total deviation, that is, C=1-SSE/SST. The confidence coefficient C takes a value between 0 and 1, which indicates the actual effect of the regression analysis, and further reflects the reliability of the pulse blood oxygen signal characteristic value R based on the linear regression model. The closer to 1, the more reliable the result.

S4. According to the confidence of the linear regression line, obtain the confidence of the blood oxygen saturation measurement result.

As a further preferred embodiment, the step of obtaining the confidence of the blood oxygen saturation measurement result according to the confidence of the linear regression line is specifically as follows:

In this embodiment, a threshold value is preset for the confidence coefficient C. If the threshold value is exceeded, the regression analysis is considered to be credible. The confidence threshold value in this embodiment is 0.95.

As shown in FIG. 7, the characteristic value R of the pulse oximetry signal calculated according to the linear regression theory of this embodiment is shown. As shown in FIG. 8, the confidence C of the linear regression line is calculated using the analysis of variance method of this embodiment. The change rule of the pulse blood oxygen signal characteristic value R and confidence C shown in 8 can be seen, when the amount of data is small, the values of R and C change drastically with time. As time goes on, the data used for fitting increases, and the values of R and C gradually converge and tend to be stable. As shown in Figure 8, after about 0.4s, C converges above 0.95, which satisfies the preset threshold, which can be considered as corresponding The R value is credible.

Corresponding to the method of FIG. 1, an embodiment of the present invention also provides a calculation system for the confidence level of blood oxygen saturation measurement, including:

The recording module is used for recording light intensity information by photoelectric volume pulse wave tracing, the light intensity information includes red light, infrared light transmitted light intensity and infrared light reflected light intensity;

As a further preferred embodiment, the recording module includes:

The acquisition module is used to obtain the analog signal of the light intensity information through the photoelectric volume pulse wave tracing method;

As a further preferred embodiment, the calculation module includes:

The first calculation unit is used to calculate the sum of squared deviations from the total deviation through a linear regression model;

At least one processor;

At least one memory for storing at least one program;

The contents of the above method embodiments are all applicable to this system embodiment, and the specific functions implemented by this system embodiment are the same as the above method embodiments, and the beneficial effects achieved are also the same as the beneficial effects achieved by the above method embodiments.

In addition, an embodiment of the present invention further provides a storage medium in which instructions executable by a processor are stored, and the instructions executable by the processor are used to execute the blood oxygen saturation measurement confidence when executed by the processor Degree calculation method.

In summary, in the process of extracting the characteristic value R of the pulse blood oxygen signal using the method of linear regression curve fitting, the present invention proposes to use the confidence coefficient C based on the analysis of variance theory as a measure to measure the quality of fitting. The confidence coefficient indicates the actual effect of regression analysis, and thus reflects the credibility of the pulse blood oxygen signal characteristic value R based on the linear regression model. Therefore, the present invention avoids the situation of missed detection and false detection caused by interference factors such as motion artifacts and noise. Therefore, by using the method of the present invention, while calculating the characteristic value of the pulse blood oxygen signal, the confidence of the detection is evaluated, and the accuracy of the measurement is greatly improved.

The above is a detailed description of the preferred implementation of the present invention, but the present invention is not limited to the described embodiments, and those skilled in the art can make various equivalent modifications or replacements without violating the spirit of the present invention. These equivalent variations or replacements are included in the scope defined by the claims of the present application.

Claims

The calculation method of blood oxygen saturation measurement confidence level is characterized by the following steps:

Record light intensity information by photoplethysmography, the light intensity information includes red light, infrared light transmitted light intensity and infrared light reflected light intensity;

According to the light intensity information, establish a linear regression model of pulse blood oxygen signal eigenvalues;

Calculate the confidence of the linear regression line in the linear regression model;

According to the confidence of the linear regression line, the confidence of the blood oxygen saturation measurement result is obtained.
The method for calculating the confidence level of blood oxygen saturation measurement according to claim 1, wherein the step of recording light intensity information by photoplethysmography includes the following steps:

Obtain light intensity information analog signal by photoelectric volume pulse wave tracing method;

Convert light intensity information analog signal into light intensity information digital signal;

Wherein, the analog signal of light intensity information is a continuous signal, and the digital signal of light intensity information is a discrete signal.
The method for calculating the confidence level of blood oxygen saturation measurement according to claim 2, wherein the step of establishing a linear regression model of the characteristic value of the pulse blood oxygen signal based on the light intensity information includes the following steps:

Calculate intermediate variables based on discrete light intensity information digital signals;

Establish linear regression equations based on intermediate variables;

Use the least square method for linear fitting, and calculate the parameters of the linear regression equation;

According to the calculated parameters, a linear regression model of pulse blood oxygen signal characteristic values is established.
The calculation method of the blood oxygen saturation measurement confidence level according to claim 1, wherein the step of calculating the confidence level of the linear regression line in the linear regression model includes the following steps:

Through the linear regression model, calculate the sum of squared deviations from the total deviation;

Through the linear regression model, calculate the sum of squared residuals;

Calculate the regression sum of squares based on the sum of squared deviations from the total deviation and the sum of squared residuals;

Calculate the confidence of the linear regression line based on the sum of squared regressions and the sum of squared deviations from the mean deviation.
The method for calculating the confidence level of blood oxygen saturation measurement according to claim 1, wherein the step of obtaining the confidence level of the blood oxygen saturation measurement result according to the confidence of the linear regression line is specifically:

Determine whether the confidence of the linear regression line is greater than the preset threshold. If it is, determine that the measurement result of blood oxygen saturation is reliable; otherwise, determine that the measurement result of blood oxygen saturation is not reliable.
The calculation system of the blood oxygen saturation measurement confidence level is characterized by:

The recording module is used for recording light intensity information by photoelectric volume pulse wave tracing, the light intensity information includes red light, infrared light transmitted light intensity and infrared light reflected light intensity;

The model building module is used to establish a linear regression model of the pulse blood oxygen signal characteristic value based on the light intensity information;

Calculation module, used to calculate the confidence of the linear regression line in the linear regression model;

The judgment module is used to obtain the confidence of the blood oxygen saturation measurement result according to the confidence of the linear regression line.
The calculation system of blood oxygen saturation measurement confidence according to claim 6, wherein the recording module comprises:

The acquisition module is used to obtain the analog signal of the light intensity information through the photoelectric volume pulse wave tracing method;

The conversion module is used to convert the analog signal of the light intensity information into the digital signal of the light intensity information;

Wherein, the analog signal of light intensity information is a continuous signal, and the digital signal of light intensity information is a discrete signal.
The calculation system of blood oxygen saturation measurement confidence according to claim 6, wherein the calculation module comprises:

The first calculation unit is used to calculate the sum of squared deviations from the total deviation through a linear regression model;

The second calculation unit is used to calculate the sum of squared residuals through a linear regression model;

The third calculation unit is used to calculate the regression sum of squares based on the sum of squared deviations from the total deviation and the sum of squared residuals;

The fourth calculation unit is used to calculate the confidence of the linear regression line based on the sum of squared regressions and the sum of squared deviations of the total deviations.
The calculation system of the blood oxygen saturation measurement confidence level is characterized by:

At least one processor;

At least one memory for storing at least one program;

When the at least one program is executed by the at least one processor, the at least one processor implements the blood oxygen saturation measurement confidence calculation method according to any one of claims 1-5.
A storage medium in which instructions executable by a processor are stored, wherein the instructions executable by the processor are used to execute the blood according to any one of claims 1-5 when executed by the processor Calculation method of confidence in oxygen saturation measurement.