Construction method of EIT impedance change of each unit of lung, equipment and program product

Abstract

The invention relates to the field of intelligent medical treatment, in particular to a construction method and device for EIT impedance changes of all lung units and a program product. Comprising the following steps: acquiring EIT pulmonary ventilation image data of a testee, including N complete respiratory cycles and distances from the unit center of each unit in a lung area to left and right bronchial mouths; calculating a maximum time constant and a minimum time constant of the lung area based on the pulmonary ventilation image data; obtaining the latest change starting point and the earliest change starting point of the lung region impedance based on the maximum time constant and the minimum time constant of the lung region; based on the distance from the unit center of each unit in the lung region to the left and right bronchus, the maximum time constant and the minimum time constant of the lung region, and the latest change starting point and the earliest change starting point of the impedance of the lung region, performing function relation fitting calculation to obtain the time constant and the impedance change starting point of each unit in the lung region; and obtaining the impedance change of each unit in the lung region based on the time constant of the unit given by the lung region and the impedance change starting point. The application has good clinical value.

Description

Technical Field

[0001] The present application relates to the field of intelligent medical care, and specifically to a method, device, program product, and computer-readable storage medium for constructing EIT impedance changes of each lung unit. Background Art

[0002] Electrical impedance tomography (EIT) is an imaging technique based on the electrical impedance characteristics of biological tissue. It measures the surface voltage by applying a safe current to specific areas of the body through surface electrodes. Image reconstruction algorithms are then used to calculate the relative or absolute distribution of electrical conductivity within the body. EIT offers numerous advantages for monitoring pulmonary ventilation status: its unique core advantage of continuous real-time imaging, its safety advantages of being radiation-free and non-invasive, and its portability and low cost. These advantages have attracted numerous researchers to promote the research and application of EIT in pulmonary ventilation monitoring, and the technology has made significant progress. EIT has a complete electromagnetic theory system, encompassing both forward and inverse problems. The goal of various EIT inverse problem algorithms is to accurately estimate the conductivity distribution within the human body. The completeness of the forward problem is a key foundation for inverse problem algorithm research. For example, the Gauss-Newton algorithm estimates the conductivity distribution by minimizing the difference between the measured boundary voltage and the boundary voltage generated by the forward problem. GREIT generates a large amount of training data using the forward problem model to obtain a reconstruction matrix, which estimates the conductivity distribution. The BP back-projection algorithm uses potential equipotential lines obtained from the forward problem as projection paths to construct a projection matrix, thereby estimating the conductivity distribution. However, in monitoring pulmonary ventilation, current forward problem simulations are limited to the end-expiratory and end-inspiratory states. While current clinical attention focuses on the conductivity differences between these two states, it is important to note that EIT monitoring of pulmonary ventilation is inherently a continuous process. Focusing solely on the end-expiratory and end-inspiratory states does not fully capture the entire pulmonary ventilation process. Therefore, studying conductivity changes throughout the entire pulmonary ventilation process holds significant clinical significance.

[0003] However, the lack of a full-scale simulation of pulmonary EIT signals has hindered further research into the inverse EIT problem. Without a complete simulation, research on continuous-time signals can only be conducted based on actual measured signals. However, these signals contain difficult-to-remove noise and difficult-to-separate blood perfusion signals, particularly the perfusion signal, which has a significant impact on pulmonary ventilation. Furthermore, the lack of a pure pulmonary ventilation signal as a standard makes it difficult to evaluate the quality of pulmonary EIT signals. Summary of the Invention

[0004] To address the above issues, the present invention provides a method for constructing EIT impedance changes of each lung unit, which specifically includes:

[0005] Obtain the test subject's EIT lung ventilation image data, including N complete respiratory cycles, where N is a natural number greater than 1, and the distance from the unit center of each unit in the lung region to the left and right bronchial orifices;

[0006] Calculating a maximum time constant and a minimum time constant of a lung region based on the lung ventilation image data;

[0007] The latest change starting point and the earliest change starting point of the lung impedance are obtained based on the maximum time constant and the minimum time constant of the lung area;

[0008] Based on the distance from the unit center of each unit in the lung region to the left and right bronchi, the maximum time constant and the minimum time constant of the lung region, and the latest change starting point and the earliest change starting point of the lung region impedance, a functional relationship fitting calculation is performed to obtain the time constant and the impedance change starting point of each unit in the lung region;

[0009] The impedance change of each unit in the lung area is obtained based on the time constant of the unit in the lung area and the starting point of the impedance change.

[0010] The maximum time constant and the minimum time constant of the lung region are calculated as follows: based on the lung ventilation image data, a time constant of impedance change of each pixel in the lung region and a time constant of impedance change of all pixels in a complete respiratory cycle are obtained, and a ratio of the time constant of impedance change of each pixel in the lung region to the time constant of impedance change of all pixels is calculated to obtain N groups of time constant ratios;

[0011] Calculating the average of the N groups of time constant ratios by cycle number to obtain the maximum time constant and the minimum time constant of the lung region;

[0012] Optionally, the functional relationship between the time constant of each unit and the distance is calculated as follows:

[0013]

[0014] in, represents the unit time constant, a and b represent the ratio of the minimum and maximum time constants of the lung unit to the time constant of the overall impedance change, aT represents the maximum time constant of the lung unit, d represents the normalized distance between the unit center and the left and right main bronchial orifices, and T represents the time constant of the overall impedance change;

[0015] Optionally, the functional relationship between the starting point of each unit change and the distance is calculated as follows:

[0016]

[0017] Wherein, s represents the starting point of unit change, d represents the normalized distance between the unit center and the left and right main bronchial orifices, T represents the time constant of overall impedance change, and b represents the ratio of the maximum time constant of the lung unit to the time constant of overall impedance change.

[0018] Optionally, the normalized distance d between the unit center and the left and right bronchi is Or s is a nonlinear relationship, and the d constraint is obtained by the nonlinear function to obtain the constrained normalized d, and the constrained normalized d is equal to Or s to perform functional relationship fitting calculation.

[0019] The nonlinear function is constrained by a hyperbolic tangent function to obtain a constrained normalized d;

[0020] Optionally, the process of constraining d is: obtaining d data, nonlinearizing function f(x);

[0021] Normalize the function f(x) to obtain the normalized function g(x);

[0022] The d data is calculated by the normalized function g(x) to obtain the constrained normalized d;

[0023] Optionally, the nonlinearization function is constructed based on a hyperbolic tangent function and is expressed as:

[0024]

[0025] in, and are the upper and lower limits of the domain, and x represents the independent variable of the function;

[0026] Optionally, the process of constraining d further includes adjusting the degree of nonlinearity, and the nonlinearity adjustment is performed by adjusting the domain of f(x), including: defining any two parameters p0 and p1, wherein the two parameters satisfy a constraint condition; converting the domain of f(x) based on the constraint condition to obtain a converted domain, and performing nonlinearity adjustment in the converted domain by the spacing between p0 and p1; the constraint condition is expressed as:

[0027]

[0028] in, for and The minimum spacing;

[0029] The function f(x) is normalized after domain conversion to obtain a normalized function;

[0030] Optionally, the nonlinear function performs d constraint to adjust the distance distribution of different units, and the distribution adjustment of the time constants and the starting points of the changes of different units is obtained by adjusting the distance distribution.

[0031] The distances from the unit center of each unit in the lung region to the left and right bronchial openings include the distance from the unit center of the left lung unit to the left bronchial opening and the distance from the unit center of the right lung unit to the bronchial opening; the time constant and impedance change starting point of each unit in the lung region are obtained by performing fitting calculation based on the distance from the unit center of the left lung unit to the left bronchial opening, the distance from the unit center of the right lung unit to the bronchial opening, the maximum time constant and the minimum time constant of the lung region, and the latest change starting point and the earliest change starting point of the lung region impedance;

[0032] Optionally, the lung region is a pixel region of the end-inspiratory image of the lung ventilation image data having a pixel value greater than a preset threshold;

[0033] The method further includes constructing an impedance change of the left lung region or the right lung region, and obtaining a distance from a unit center of the left or right lung unit to the left or right bronchial orifice;

[0034] Based on the lung ventilation image data, the maximum time constant and the minimum time constant of the left or right lung area are obtained;

[0035] The latest change starting point and the earliest change starting point of the left or right lung region are obtained based on the maximum time constant and the minimum time constant of the left or right lung region;

[0036] The impedance change of the left or right lung area is obtained by fitting and calculating based on the distance from the unit center of the left or right lung unit to the left or right bronchial orifice, the maximum time constant and the minimum time constant of the left or right lung area, the latest change starting point and the earliest change starting point of the left or right lung area;

[0037] Optionally, the lung units include lungs, heart, spine, ribs, sternum, and trachea.

[0038] The present invention aims to provide a method for constructing a lung EIT signal simulation model, comprising:

[0039] Obtain human chest imaging and conductivity change curve data of blood perfusion;

[0040] Constructing a three-dimensional model based on the chest image to obtain a finite element three-dimensional chest model;

[0041] Performing excitation measurement on the finite element three-dimensional chest model to obtain a sensitivity matrix;

[0042] Based on the above-mentioned method for constructing the EIT impedance change of each lung unit and the finite element three-dimensional chest model, the electrical impedance change of each lung unit and the total electrical impedance change of the first lung area are obtained;

[0043] performing data processing on the conductivity change curve data of the blood perfusion based on the total electrical impedance change of the first lung area to obtain a processed impedance change of the blood perfusion;

[0044] performing impedance addition on the electrical impedance change of each lung unit based on the impedance change of the processed blood perfusion to obtain the electrical impedance change of each unit after the addition;

[0045] The lung EIT simulation signal is calculated based on the sensitivity matrix and the change in electrical impedance of each unit after addition.

[0046] The impedance addition is to add an impedance change to the electrical impedance change of each unit that is opposite to the change in blood perfusion impedance and has an amplitude of half the amplitude of the change in blood perfusion impedance;

[0047] Optionally, the lung heart unit further includes a heart blood perfusion impedance change, and the processed blood perfusion impedance change is added to the heart unit to obtain a processed heart unit impedance change, and then the impedance addition is performed on the processed heart unit impedance change to obtain a post-addition heart unit electrical impedance change;

[0048] Optionally, the data processing is to convert the conductivity of the blood perfusion into the blood perfusion impedance, and then adjust the variation of the blood perfusion impedance to 5% of the variation of the total electrical impedance to obtain the processed blood perfusion impedance variation;

[0049] Optionally, the 5 percent of the variation amplitude is 5 percent of the product of the lung conductivity variation and the lung volume;

[0050] Optionally, the total electrical impedance change of the first lung area is obtained by multiplying the electrical impedance change of each unit by the volume of each unit and then summing them up;

[0051] Optionally, the finite element three-dimensional model construction includes chest model reconstruction, setting EIT electrodes, and conductivity configuration, wherein a three-dimensional chest model is obtained after three-dimensional reconstruction of chest images, EIT electrodes are set in the three-dimensional chest model, and conductivity is configured for different tissue regions in the three-dimensional chest model to obtain a finite element three-dimensional chest model;

[0052] Optionally, performing excitation measurement on the EIT electrode sheet in the finite element three-dimensional chest model to obtain a sensitivity matrix;

[0053] Optionally, the chest model reconstruction is performed by dividing the chest into different tissue regions, reconstructing the different tissue regions to obtain a chest model, and then setting EIT electrodes in the chest model to obtain a three-dimensional chest model;

[0054] Optionally, the different tissue regions include one or more of the following: chest, lungs, heart, spine, ribs, sternum, trachea;

[0055] Optionally, the different tissue regions further include other regions, and the other regions are treated as a whole, and the electrical conductivity of the other regions is uniformly configured;

[0056] Optionally, the EIT simulation signal includes one or more of the following: a pure lung ventilation EIT signal, a pure blood perfusion EIT signal, and a mixed lung EIT signal.

[0057] The object of the present invention is to provide a lung EIT signal simulation model, wherein the EIT signal simulation model is obtained by the above-mentioned lung EIT signal simulation model construction method.

[0058] The object of the present invention is to provide a computer program product comprising a computer program or instructions, which are executed by a processor to implement the above-mentioned method for constructing the EIT impedance changes of each lung unit, or the method for constructing a lung EIT signal simulation model.

[0059] An object of the present invention is to provide a computer device comprising a memory, a processor, and a computer program or instructions stored on the memory, wherein the computer program or instructions are executed by the processor to implement the above-mentioned method for constructing the EIT impedance changes of each lung unit, or the method for constructing a lung EIT signal simulation model.

[0060] An object of the present invention is to provide a computer-readable storage medium having a computer program or instructions stored thereon, wherein the computer program or instructions are executed by a processor to implement the above-mentioned method for constructing the EIT impedance changes of each lung unit, or the method for constructing a lung EIT signal simulation model.

[0061] Advantages of the present invention:

[0062] 1. To address the lack of a complete simulated signal, the actual measured signal contains difficult-to-remove noise and difficult-to-separate blood perfusion signals, resulting in a lack of a pure lung ventilation signal. A chest model was developed that includes complete and accurate information about the lungs and heart, as well as information about the spine, ribs, sternum, and trachea, which significantly impede chest current flow. By simulating the full-scale conductivity changes of all lung cells, as well as the conductivity changes caused by blood perfusion, a complete lung EIT signal and a pure lung ventilation signal unaffected by blood perfusion were generated.

[0063] 2. Current EIT signal simulations use overall impedance changes. However, in the real world, the impedance changes of different tissue units in the chest vary. To reduce errors and improve accuracy, the present invention simulates the impedance changes of different tissue units, deriving the impedance changes of each tissue unit from their time constants and starting points. Furthermore, the time constants and starting points of impedance changes for different tissue units are related to the distance between the left and right bronchi. Experiments have verified that the impedance changes of different tissue units are derived by fitting a functional relationship between the distance between the left and right bronchi and the time constants and starting points of impedance changes, applying nonlinear constraints.

[0064] 3. Regarding the quality of the EIT real-time signal, the present invention calculates the signal-to-noise ratio threshold through a simulated signal, and evaluates the quality of the real-time signal through the signal-to-noise ratio threshold. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0066] Figure 1 A flow chart of a method for constructing EIT impedance changes of each lung unit provided by an embodiment of the present invention;

[0067] Figure 2 Schematic diagram of a system for constructing EIT impedance changes of each lung unit provided by an embodiment of the present invention;

[0068] Figure 3 Schematic diagram of a device for constructing EIT impedance changes of each lung unit provided by an embodiment of the present invention;

[0069] Figure 4 Pulmonary impedance change simulation curves provided by embodiments of the present invention. (a) Overall impedance change curve; (b) Impedance change curves for different units; (c) Function curve and distribution adjustment histogram of the constructor; (d) Multiple correlation curves included in the constructor and corresponding distribution adjustment histograms; (e) Nonlinear relationship with ; (f) Nonlinear relationship with ;

[0070] Figure 5 The 3D chest finite element model construction process provided in the embodiment of the present invention;

[0071] Figure 6The conductivity distribution changes of the lungs at different breathing moments and their EIT reconstructed images are provided by an embodiment of the present invention. The MAX value is the difference between the end-expiratory and end-inspiratory lung conductivity, 0.1593 S / m.

[0072] Figure 7 EIT simulation signals provided by an embodiment of the present invention. (a) Electrical impedance change curves of blood perfusion, lung ventilation, and intact lung EIT signals; (b) Boundary voltage and curves of lung ventilation and intact lung EIT signals.

[0073] Figure 8 Boundary voltages and curves provided by embodiments of the present invention. (a) Boundary voltages and curves of a pure lung ventilation signal, signal A; (b) Boundary voltages and curves of a lung EIT signal, signal B;

[0074] Figure 9 Figure 1 shows the difference between lung EIT signals and pure lung ventilation signals at different signal-to-noise ratios provided by the present invention. (a) RVDerror changes with signal-to-noise ratio; (b) SD changes with signal-to-noise ratio;

[0075] Figure 10 This is to verify the accuracy of the signal-to-noise ratio estimation based on the simulated signal provided by the embodiment of the present invention. DETAILED DESCRIPTION

[0076] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0077] In some of the processes described in the specification and claims of the present invention and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or may be executed in parallel. The serial numbers of the operations, such as S101, S102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., and do not represent the order of precedence, nor do they limit "first" and "second" to be different types.

[0078] Figure 1 The schematic diagram of the method for constructing the EIT impedance change of each lung unit provided by an embodiment of the present invention specifically includes:

[0079] S1: Obtain the EIT lung ventilation image data of the test subject, including N complete respiratory cycles, where N is a natural number greater than 1 and the distance from the unit center of each unit in the lung area to the left and right bronchial orifices;

[0080] In one embodiment, the distances from the unit center of each unit in the lung region to the left and right bronchial openings include the distance from the unit center of the left lung unit to the left bronchial opening and the distance from the unit center of the right lung unit to the right bronchial opening.

[0081] In one embodiment, the lung region is a pixel region in which the pixel value of the end-inspiratory image of the lung ventilation image data is greater than a preset threshold.

[0082] In one embodiment, the lung units include lungs, heart, spine, ribs, sternum, and trachea, and other units are also included, and the conductivity of the other units is calculated as a whole.

[0083] S2: Calculating a maximum time constant and a minimum time constant of a lung region based on the lung ventilation image data;

[0084] The latest change starting point and the earliest change starting point of the lung impedance are obtained based on the maximum time constant and the minimum time constant of the lung area;

[0085] In one embodiment, the maximum time constant and the minimum time constant of the lung region are calculated by: obtaining, based on the lung ventilation image data, a time constant of impedance change of each pixel in the lung region and a time constant of impedance change of all pixels in a complete respiratory cycle, and calculating a ratio of the time constant of impedance change of each pixel in the lung region to the time constant of impedance change of all pixels to obtain N sets of time constant ratios;

[0086] The N groups of time constant ratios are averaged according to the number of cycles to obtain the maximum time constant and the minimum time constant of the lung region.

[0087] S3: Based on the distance from the unit center of each unit in the lung region to the left and right bronchi, the maximum time constant and the minimum time constant of the lung region, and the latest change starting point and the earliest change starting point of the lung region impedance, a functional relationship fitting calculation is performed to obtain the time constant and impedance change starting point of each unit in the lung region;

[0088] In one embodiment, a fitting calculation is performed based on the distance from the unit center of the left lung unit to the left bronchial opening, the distance from the unit center of the right lung unit to the right bronchial opening, the maximum time constant and the minimum time constant of the lung area, and the latest change starting point and the earliest change starting point of the lung area impedance to obtain the time constant and the impedance change starting point of each unit in the lung area.

[0089] In one embodiment, the functional relationship between the time constant of each unit and the distance is calculated as follows:

[0090]

[0091] in, represents the unit time constant, a and b represent the ratios of the minimum and maximum time constants of the lung unit to the time constant of the overall impedance change, respectively, aT represents the maximum time constant of the lung unit, d represents the normalized distance between the unit center and the left and right main bronchial orifices, and T represents the time constant of the overall impedance change.

[0092] In one embodiment, the functional relationship between the starting point of each unit change and the distance is calculated as follows:

[0093]

[0094] Wherein, s represents the starting point of unit change, d represents the normalized distance between the unit center and the left and right main bronchial orifices, T represents the time constant of the overall impedance change, and b represents the ratio of the maximum time constant of the lung unit to the time constant of the overall impedance change.

[0095] In one embodiment, the normalized distance d between the unit center and the left and right bronchi is Or s is a nonlinear relationship, and the d constraint is obtained by the nonlinear function to obtain the constrained normalized d, and the constrained normalized d is equal to Or s to perform functional relationship fitting calculation.

[0096] In one embodiment, the nonlinear function performs d constraint to adjust the distance distribution of different units, and the time constants and the distribution adjustment of the change starting points of different units are obtained by adjusting the distance distribution.

[0097] In one embodiment, the nonlinear function is constrained by a hyperbolic tangent function to obtain a constrained and normalized d.

[0098] In one embodiment, the process of constraining d is: obtaining d data, nonlinearizing function f(x);

[0099] Normalize the function f(x) to obtain the normalized function g(x);

[0100] The d data is calculated using the normalized function g(x) to obtain the constrained normalized d.

[0101] In one embodiment, the nonlinearization function is constructed based on the hyperbolic tangent function and is expressed as:

[0102]

[0103] in, and are the upper and lower limits of the domain, and x represents the independent variable of the function.

[0104] In one embodiment, the process of constraining d further includes adjusting the degree of nonlinearity, and the nonlinearity adjustment is performed by adjusting the domain of f(x), including: defining any two parameters p0 and p1, wherein the two parameters satisfy a constraint condition; converting the domain of f(x) based on the constraint condition to obtain a converted domain, and performing nonlinearity adjustment in the converted domain by the spacing between p0 and p1; the constraint condition is expressed as:

[0105]

[0106] in, for and The minimum spacing;

[0107] The function f(x) is normalized after domain conversion to obtain a normalized function.

[0108] S4: Obtain the impedance change of each unit in the lung area based on the time constant of the unit in the lung area and the impedance change starting point.

[0109] In one embodiment, the method further comprises constructing an impedance change of the left lung region or the right lung region, and obtaining a distance from a unit center of the left or right lung unit to the left or right bronchial orifice;

[0110] Based on the lung ventilation image data, the maximum time constant and the minimum time constant of the left or right lung area are obtained;

[0111] The latest change starting point and the earliest change starting point of the left or right lung region are obtained based on the maximum time constant and the minimum time constant of the left or right lung region;

[0112] The impedance change of the left or right lung area is obtained by fitting and calculating based on the distance from the unit center of the left or right lung unit to the left or right bronchial orifice, the maximum time constant and the minimum time constant of the left or right lung area, and the latest change starting point and the earliest change starting point of the left or right lung area.

[0113] In a specific embodiment, the change process of lung conductivity is simulated:

[0114] Experimental studies have shown that under normal breathing conditions, the main part of the change in the overall conductivity of the lungs and the main part of the change in the conductivity of each unit in the lungs conform to the law of exponential function change (the overall conductivity change here is the sum of the product of the conductivity change of all units in the lungs and the unit volume).

[0115] The conductivity changes during inspiration and expiration occur over different time periods, but the trends are approximately symmetrical. Therefore, the discussion below focuses solely on the inhalation process, with the exhalation process captured through symmetrical stretching. During inspiration, the lungs' conductivity decreases due to the increase in air volume, inversely matching the change in air volume. For ease of understanding, the conductivity change is negated to align with the air volume trend. Furthermore, it can be shown that when the base conductivity is sufficiently large compared to the variable conductivity (i.e., the base conductivity is relatively stable), the negation of the conductivity change is equivalent to the impedance change, hence the term "impedance change" below.

[0116] The law of change of the overall electrical impedance of the lungs: This law is expressed by the time constant It can be described by the following function about time:

[0117] (1)

[0118] in, for The change in lung impedance at each moment, represents the end of expiration, at which point the lung impedance is The basic impedance, , is the change in lung impedance at the end of inspiration, and is the maximum change in lung impedance. Represents the first 75% of the entire inhalation process, that is, the entire inhalation duration is Obviously, , and the derivative is 0, and according to formula (1), ,and Given the two endpoints and the derivatives of a curve, we can use cubic spline interpolation to get Assuming an I:E ratio of 1:2 and a cycle length of 5 seconds, the global impedance curve of the lung (GIC) is as follows: Figure 4 As shown in a.

[0119] The impedance change law of each lung unit: The impedance change law of each unit is similar to the overall change law, but slightly different. The time constants of different units are different, and the filling start time of the alveoli in different units is different, that is, the starting point of the impedance change is different. For a single unit, there is a certain impedance constant period when the impedance change reaches the maximum and minimum. The physical meaning of the constant period is that the gas volume of the alveoli in the unit remains constant during this time period.

[0120] The human lung is divided into 5 lobes and 18 segments. After air is inhaled, it is diverted from the main bronchi and flows through the small bronchi and bronchioles to the various lung segments of the lungs, filling the alveoli in each segment. During the gas flow, the pressure difference between the pressure inside the lungs and the pressure of the external environment gradually decreases, the resistance gradually increases, and the flow speed slows down. This process can be approximated as follows: the closer the alveoli are to the main bronchus, the earlier and faster they are filled, and they remain in a filled state for a longer time, and the farther they are, the worse it is. That is, the size of the time constant and the early or late starting point of the impedance change are both related to the distance between the unit center and the left and right main bronchial openings (the distance between the left lung unit and the left main bronchial opening, the distance between the right lung unit and the right main bronchial opening, uniformly recorded as vector ) is related to the size of

[0121] During inspiration, alveoli close to the bronchi fill faster and earlier, and remain filled for a longer time. Therefore, the time constant is shorter, the starting point of change is earlier, and the constant period when the impedance is maximum is longer. Conversely, alveoli far from the bronchi have a longer time constant, a later starting point of change, and a longer constant period when the impedance is minimum. The impedance curve of lung elements (EIC) is as follows: Figure 4 As shown in b.

[0122] The overall inspiratory impedance change curve is known, and the inspiratory impedance change of each unit must occur within this cycle. Therefore, by determining the values ​​of any two variables among the three variables of time constant, change starting point and constant period, the value of the other variable can be obtained. Therefore, by confirming the time constant and change starting point of each unit, the impedance change law of each unit in the lung can be confirmed. Although there is an apparent correlation, the specific relationship is unknown.

[0123] For units near the bronchial orifices, their time constants are the smallest, and their starting point of change is the overall onset of inspiration. For units farthest from the bronchial orifices, their time constants are the largest, their starting point of change is the latest, and their end point of change is the overall end of inspiration. Determining the maximum time constant also determines the latest starting point of change. Therefore, determining the minimum and maximum time constants for units in a lung region defines the range of time constants and starting points of change for all units.

[0124] The present invention uses actual EIT measurement data to obtain minimum and maximum time constants: under a calm breathing state, the EIT device is used to monitor the lung ventilation status of a normal adult male for approximately 3 minutes. The pixel area in the end-inspiratory image that is greater than 25% of the maximum pixel value is recorded as the lung area. The time constant of the impedance change of each pixel in the lung area during all complete inspiratory cycles of 3 minutes is calculated and compared with the time constant of the impedance change of all pixels. The data are averaged by the number of cycles to eliminate errors, and the minimum and maximum time constant ratios are obtained, which are 0.75 and 0.96, respectively.

[0125] The ratio of the minimum and maximum time constants of the lung unit to the time constant of the overall impedance change is expressed as , The time constant of the overall impedance change during the inhalation process is , the minimum and maximum time constants are recorded as, , At this time, the earliest change starting point and the latest change starting point are 0 respectively. .

[0126] First, assume that the time constant and distance of the unit , the starting point and distance of the unit change All are linear relationships,

[0127] (2)

[0128] (3)

[0129] in, 、 are the time constant and starting point of the unit, respectively. is the normalized distance between the unit center and the left and right main bronchial orifices, Based on formulas (2) and (3), the time constant and starting point of each lung unit can be determined. Accordingly, the impedance change curve of each unit can be obtained. By multiplying all unit impedance change curves by the unit volume and summing them up, a new set of overall impedance change curves (GIC-0) can be drawn, as shown in Figure 2. Figure 4 Obviously, the new overall impedance change curve is quite different from the actual overall impedance change curve, indicating that and and It's not a simple linear relationship.

[0130] This paper proposes the following method to seek and , and Related relationships:

[0131] The following function is constructed based on the hyperbolic tangent function:

[0132] (4)

[0133] in, The lower limit of the definition domain is set independently. , because the derivative is , close enough to 0 to meet the requirements of subsequent solutions. This function is obtained by the hyperbolic tangent function After segmented splicing and normalization, the function curve is as follows Figure 4 As shown in c, Is a monotonically increasing piecewise bounded function. Set two parameters and , satisfying the following constraints,

[0134] (5)

[0135] in for and The minimum spacing is taken in this paper , by choosing different and The values ​​of New upper and lower bounds of the domain, and Normalize again to generate a new function ,right Perform a scale transformation so that , and then ,at this time Still satisfied , resubstitute into formula (2) or (3) to achieve or and Nonlinearity of relationships.

[0136] Figure 4 c and Figure 4 It can be noticed that when and When both are in the interval [-2.5, 0] or [2.5, 5], the nonlinear relationship is characterized by a concave function. and When both are in [0,2.5], the nonlinear relationship is characterized by a convex function. and When they are on both sides of 0, the nonlinear relationship is close to linear. and In other positions, the nonlinear relationship is a combination of the above three relationships. , The choice of spacing can achieve the weakening or deepening of the above-mentioned different degrees of nonlinearity. right The change is to adjust the distance distribution of different units. Figure 4 c and Figure 4 The histogram of the side axis of d is the corresponding Right and The distribution adjustment is performed after the scale of a group of numbers evenly distributed between the two is changed to 0 and 1. By adjusting the distance distribution, the distribution adjustment of the time constants and the starting point of the change of different units is achieved, thereby achieving the approximation of the overall impedance curve.

[0137] by and As parameters, based on and The error between the calculated overall impedance change curve (GIC-1) and the actual overall impedance change curve is the loss function. The global optimal solution algorithm (the present invention uses the pattern search algorithm, pattern search) is used to obtain the solution that minimizes the loss function. and . and , and The two groups of correlations correspond to the two groups and , based on the constructed finite element model (including the finite element chest three-dimensional model and other chest finite element three-dimensional models below), the solution results are , ; , The calculated overall impedance change curve (GIC-1) is as follows: Figure 4 As shown in a, the error is reduced by 93.5% compared with the overall impedance change curve (GIC-1) generated based on the linear relationship. and , and The actual relationships are as follows Figure 4 e and Figure 4 As shown in f.

[0138] In one embodiment, the chest is the imaging area, and the finite element chest model discretizes the area into K tetrahedral finite element units, where K is a natural number greater than 1. The parameters of the finite element units include conductivity, and the unit conductivity changes based on time configuration to achieve impedance change simulation.

[0139] An embodiment of the present invention provides a method for constructing a lung EIT signal simulation model, comprising:

[0140] Obtain human chest imaging and conductivity change curve data of blood perfusion;

[0141] Constructing a three-dimensional model based on the chest image to obtain a finite element three-dimensional chest model;

[0142] Performing excitation measurement on the finite element three-dimensional chest model to obtain a sensitivity matrix;

[0143] Based on the above-mentioned method for constructing the EIT impedance change of each lung unit and the finite element three-dimensional chest model, the electrical impedance change of each lung unit and the total electrical impedance change of the first lung area are obtained;

[0144] performing data processing on the conductivity change curve data of the blood perfusion based on the total electrical impedance change of the first lung area to obtain a processed impedance change of the blood perfusion;

[0145] performing impedance addition on the electrical impedance change of each lung unit based on the impedance change of the processed blood perfusion to obtain the electrical impedance change of each unit after the addition;

[0146] The lung EIT simulation signal is calculated based on the sensitivity matrix and the change in electrical impedance of each unit after addition.

[0147] In one embodiment, the total electrical impedance change of the first lung region is obtained by multiplying the electrical impedance change of each unit by the volume of each unit and then summing them up.

[0148] In one embodiment, the electrical impedance change of each lung unit and the total electrical impedance change of the first lung region are obtained based on the EIT impedance change process of each unit of the above-mentioned method for constructing the EIT impedance change of each lung unit and the finite element three-dimensional chest model.

[0149] In one embodiment, the impedance addition is to add an impedance change to the electrical impedance change of each unit, which is opposite to the blood perfusion impedance change and has an amplitude of half the amplitude of the blood perfusion impedance change.

[0150] In one embodiment, the processed impedance change of the blood perfusion is added to the heart unit to obtain the processed impedance change of the heart unit, and the pulmonary EIT simulation signal is calculated based on the sensitivity matrix, the processed impedance change of the heart unit and the added impedance change of each unit.

[0151] In one embodiment, the lung units in the impedance addition based on the electrical impedance changes of the lung units do not include the heart unit.

[0152] In one embodiment, the impedance change of the processed blood perfusion is added to the heart unit to obtain the processed heart unit impedance change. The lung units in the impedance addition of the electrical impedance change of each lung unit do not include the heart unit. The lung EIT simulation signal is calculated based on the sensitivity matrix, the processed heart unit impedance change and the electrical impedance change of each unit after addition.

[0153] In one embodiment, the data processing is to convert the conductivity of blood perfusion into blood perfusion impedance, and then adjust the change amplitude of blood perfusion impedance to 5% of the total electrical impedance change amplitude to obtain the processed blood perfusion impedance change.

[0154] In one embodiment, the five percent of the magnitude of the change is five percent of the product of the lung conductivity change and the lung volume.

[0155] In one embodiment, the finite element three-dimensional model construction includes chest model reconstruction, setting EIT electrodes, and conductivity configuration. A three-dimensional chest model is obtained by three-dimensional reconstruction of chest images, EIT electrodes are set in the three-dimensional chest model, and conductivity is configured for different tissue areas in the three-dimensional chest model to obtain a finite element three-dimensional chest model.

[0156] In one embodiment, excitation measurement is performed on the EIT electrode pads in the finite element three-dimensional chest model to obtain a sensitivity matrix.

[0157] In one embodiment, the chest model is reconstructed by dividing the chest into different tissue regions, reconstructing the different tissue regions to obtain a chest model, and then setting EIT electrodes in the chest model to obtain a three-dimensional chest model.

[0158] In one embodiment, the different tissue regions include one or more of the following: chest, lung, heart, spine, ribs, sternum, and trachea.

[0159] In one embodiment, the different tissue regions further include other regions, and the electrical conductivity of the other regions is uniformly configured as a whole.

[0160] In one embodiment, the EIT simulation signal includes one or more of the following: a pure lung ventilation EIT signal, a pure blood perfusion EIT signal, and a mixed lung EIT signal.

[0161] In a specific embodiment, accurate simulation of lung EIT signals requires the construction of an accurate 3D chest finite element model, accurate simulation of the conductivity changes of all lung cells, and the addition of blood perfusion signals.

[0162] Constructing an accurate 3D chest model:

[0163] Constructing a complete 3D chest finite element model requires distinguishing different tissue regions, setting up EIT electrodes, and configuring the conductivity of different tissue regions of the model.

[0164] Differentiating Different Chest Tissue Regions: Differentiating different tissue regions requires a comprehensive consideration of both accuracy and model complexity. To ensure accuracy, differentiation of all different chest tissue regions should be achieved as much as possible. To reduce model complexity, numerous tissue regions with similar conductivity, which have little impact on conductivity changes during respiration, should be comprehensively considered and not differentiated. In the chest model, the lungs and heart are the primary regions of conductivity variation and are a key focus for EIT. Bone has significantly lower conductivity than other chest tissues due to its high mineral content, which significantly impedes current flow. Similarly, the trachea, being permanently filled with gas, has much lower conductivity than other chest tissues. Both regions significantly impede the current applied by the EIT device, and their impact on boundary voltages is non-negligible. Therefore, to ensure model accuracy and reliability, the 3D chest finite element model includes six major regions: the lungs, heart, spine, ribs, sternum, and trachea.

[0165] Setting up EIT electrodes: To obtain EIT simulation signals, 16 equidistant electrodes were placed on the horizontal plane between the fourth and fifth intercostal spaces, starting from the frontal midline of the human body. These electrodes consist of electrode patches and conductive paste, which adheres tightly to the surface of the chest wall. The electrode patches are used for current excitation and voltage measurement, and the conductive paste is used to set contact impedance.

[0166] Configuring the conductivity of different tissue regions in the model: 50 kHz is a common excitation current frequency for lung EIT. Table 1 shows the conductivity of different tissues at this frequency. In the chest model, excluding the six major regions of the lungs, heart, spine, ribs, sternum, and trachea, all other tissue regions are considered as a unified whole. The conductivity of glands such as lymphatic tissue, pancreas, thymus, and thyroid glands within these regions is 0.53395 S / m, while the conductivity of major organs such as the esophagus and small intestine is 0.53369 S / m. Therefore, the conductivity of this region in the chest model is uniformly set to 0.534 S / m. In addition, the material of the electrode patch is copper, which has a conductivity of 59600000 S / m. The normal reference value of the contact impedance between the EIT electrode and the human body is usually in the range of 100 ohms to 300 ohms. In the present invention, it is set to about 125 ohms. Among them, the thickness of the conductive paste is 1 mm, and the bottom radius is 5 mm, which is approximately a cylinder. According to the resistance law, the conductivity of the conductive paste is set to 0.1 S / m.

[0167] Table 1. Electrical conductivity of different chest tissues

[0168]

[0169] Construct a 3D chest finite element model: Based on the real human chest plain scan CT image, use the software MIMICS to extract the regional boundaries of each tissue (chest, lungs, heart, spine, ribs, sternum and trachea), and export it in the stl file format. Then import the chest regional boundaries into the software SOLIDWORKS, and by adding reference lines and reference surfaces and other functions in the software, locate the electrode position and create 16 equidistant electrode drawing platforms tangent to the chest surface, and also export it in the stl file format. Import the regional boundaries and electrode drawing platforms of all tissues into the software COMSOL, create 16 equidistant electrodes based on the electrode drawing platform, and combine the various parts to form a union to complete the chest model construction. Finally, perform finite element segmentation on the model in COMSOL to obtain the final 3D finite element model. Figure 5 The model building process is shown. The finite element model contains 34,228 nodes and 183,614 elements. In addition, a fine mesh is defined on each electrode to better simulate the local current distribution and sensitivity.

[0170] In one embodiment, the conductivity change of blood perfusion is added:

[0171] The impact of conductivity changes caused by blood perfusion on pulmonary ventilation monitoring is significant. Changes in cardiac volume reflect changes in blood perfusion. Therefore, the present invention utilizes the known ventricular volume change curve to simulate the conductivity change curve caused by blood perfusion. Previous experimental studies have shown that the impedance change amplitude due to blood perfusion does not exceed 5% of the impedance change amplitude due to pulmonary ventilation. In the present invention, the impedance change amplitude for blood perfusion is set at 5% of the impedance change amplitude due to pulmonary ventilation. Compared to the lungs, the heart is smaller, and the blood perfusion signal amplitude is smaller and changes more rapidly. Therefore, an approximation can be made, setting the impedance change uniformly for all cells within the entire cardiac region. It should be noted that the 5% change amplitude set in the present invention does not represent 5% of the maximum change in conductivity of the lung cells during pulmonary ventilation, but rather 5% of the product of the change in lung conductivity and lung volume. Therefore, when setting the conductivity change for the heart region cells, it is necessary to divide by the heart volume. Furthermore, each time the heart pumps blood, approximately 50% of the blood flow enters the lungs through the pulmonary artery, participating in the pulmonary circulation. Therefore, all cells in the lung region simultaneously add an impedance change that is opposite to the blood perfusion and has an amplitude of 50% of the blood perfusion impedance change.

[0172] In one embodiment, the blood perfusion impedance change is added to the lung electrical impedance change, and the lung electrical impedance change is calculated based on the above-mentioned electrical impedance changes of each lung tissue area; the blood perfusion impedance is added to the electrical impedance changes of each tissue unit in any one or several lung areas, and the electrical impedance changes of each tissue unit are obtained by the above-mentioned method of constructing the EIT impedance changes of each lung unit.

[0173] In one embodiment, a lung EIT simulation signal is generated:

[0174] EIT applies a safe current to a specific part of the human body through surface electrodes, and measures the surface potential caused by the current excitation. The process can be described mathematically as follows:

[0175] (6)

[0176] in, represents the change in conductivity, , and represent the current conductivity distribution and the reference conductivity distribution respectively; represents the boundary voltage change, and Represent the current boundary voltage vector and the reference boundary voltage vector respectively; is the Jacobian matrix or sensitivity matrix, , calculated based on a given finite element model and excitation measurement pattern.

[0177] Using the finite element model constructed in the above part, different excitation and measurement modes are adopted to calculate the corresponding sensitivity matrix , using the above-mentioned construction method of the EIT impedance change of each lung unit to generate the electrical impedance change, after taking the inverse, we can get the conductivity change , For time The total number of frames collected in the lung is generated using formula (6). .

[0178] Compared with the actual measurement method, the simulation method has the advantages of being noise-free and having separable signals. Using the simulation method, pure lung ventilation EIT signals, pure blood perfusion EIT signals, and a mixed lung EIT signal can be obtained.

[0179] In one specific embodiment, the lung EIT signal generated by the aforementioned lung EIT signal simulation model construction (simulation) method employed an opposing excitation adjacent measurement mode. Each EIT signal frame contained 192 measurement values, with a respiratory cycle of 5 seconds, a cardiac cycle of 0.8 seconds, a sampling rate of 20 Hz, and a 25-second acquisition period. The inverse problem reconstruction algorithm (GREIT) was used, and the noise figure (NF) was 0.5, a key parameter of the GREIT inverse problem model. The results are as follows:

[0180] Figure 6 It shows the distribution changes of lung ventilation conductivity during a respiratory cycle. , and the corresponding EIT reconstructed image .

[0181] Figure 7 The time-varying blood perfusion, lung ventilation and the sum of the total electrical impedance changes of the two are shown respectively. , and the lung ventilation and lung EIT signals generated using formula (6), The sum of the absolute values ​​of The change in blood perfusion impedance was 5% of the change in lung ventilation impedance, causing significant fluctuations in both the impedance change and the lung EIT signal. This indicates that the impact of conductivity changes caused by blood perfusion cannot be ignored during EIT monitoring of lung ventilation.

[0182] In a specific embodiment, the current problem with lung EIT simulation is that it focuses on the two states of end-expiration and end-inspiration, and can only obtain state information but not continuous-time signals. In fact, EIT monitoring of lung ventilation is a continuous process, and continuous-time signals contain a wealth of undiscovered pathological information. However, the actual measured lung EIT signal inevitably contains blood perfusion signals, and there is currently no effective method to completely separate them, which limits the continuity analysis of lung ventilation signals. The full-process simulation method for lung EIT signals proposed in this article generates continuous-time signals while obtaining pure lung ventilation signals, solving the current problem.

[0183] In one embodiment, the EIT simulation signal is based on the above-mentioned construction method of the EIT impedance change of each lung unit to obtain the EIT impedance change process of each unit (the change law, fitted as a mathematical function). In the finite element chest model, each unit changes based on the EIT impedance change process, and the impedance of each EIT unit is converted into conductivity change. The EIT electrode sheet obtains a sensitivity matrix by stimulating the measured boundary voltage, and is obtained through the sensitivity matrix and conductivity change.

[0184] The disclosed embodiments of the present invention further provide a computer program product or system, including a computer program, which, when executed by a processor, implements the above-mentioned method steps for constructing the EIT impedance changes of each lung unit or the method steps for constructing a lung EIT signal simulation model.

[0185] In one embodiment, the method further includes evaluating the quality of a real-time EIT signal, calculating a preset signal-to-noise ratio threshold using a pure lung ventilation EIT signal and a mixed lung EIT signal, obtaining a real-time EIT signal, performing decomposition filtering, and then calculating the signal-to-noise ratio to obtain a real-time signal-to-noise ratio, and comparing the real-time signal-to-noise ratio with the preset signal-to-noise ratio threshold to obtain an evaluation result indicating high or low quality.

[0186] In one embodiment, the process of constructing the preset threshold is as follows:

[0187] Obtain the EIT pure signal data set, recorded as A signal, and the noise-free lung ventilation signal data set with added blood perfusion signal, recorded as B signal;

[0188] Adding different levels of electronic measurement noise to the B signal to generate B signal data sets with different signal-to-noise ratios;

[0189] Calculate parameters under different signal-to-noise ratio signals based on signal A and signal B;

[0190] The signal-to-noise ratio threshold is obtained based on the variation characteristics of the parameters with the signal-to-noise ratio;

[0191] Optionally, for each signal-to-noise ratio signal, the signal-to-noise ratio is fixed, S groups of different random noises are added, S groups of parameter values ​​are calculated, and the S groups of parameter values ​​are averaged to obtain a parameter of the current signal-to-noise ratio; S is a natural number greater than or equal to 100;

[0192] Optionally, the preset threshold range is 8-12 dB, and the comparison evaluation process is: when the signal-to-noise ratio of the real-time EIT signal calculated after decomposition filtering is greater than 12 dB, it is determined to be a high-quality signal; when the signal-to-noise ratio of the real-time EIT signal calculated after decomposition filtering is less than 8 dB, it is determined to be a low-quality signal;

[0193] Optionally, the comparative evaluation process further includes an early warning, where when the signal-to-noise ratio is continuously within a preset threshold interval of 10dB-12dB and the duration thereof is greater than the window width, it is determined to be a high-quality early warning signal; when the signal-to-noise ratio is continuously within a preset threshold interval of 8dB-10dB and the duration thereof is greater than the window width, it is determined to be a low-quality early warning signal;

[0194] Optionally, the preset threshold is 10 dB, and the comparison evaluation process is replaced by: when the signal-to-noise ratio of the real-time EIT signal calculated after decomposition filtering is greater than 10 dB, it is determined to be a high-quality signal; when the signal-to-noise ratio of the real-time EIT signal calculated after decomposition filtering is less than 10 dB, it is determined to be a low-quality signal.

[0195] In one embodiment, the method extracts the actual respiratory cycle length from the EIT real-time signal by adaptively adjusting the window width to perform signal quality assessment;

[0196] Optionally, the process of adaptively adjusting the window width to extract the actual respiratory cycle length from the EIT real-time signal is as follows:

[0197] S1 sets the initial respiratory value c0 seconds based on the respiratory cycle C of normal people, the window width W includes N respiratory cycles, and the initial window width w0 is Nc0 seconds;

[0198] S2 calculates the signal-to-noise ratio of the signal within the initial window width to obtain an estimated value, which is recorded as snr0;

[0199] S3 detects the actual respiratory cycle of the signal within the initial window width to obtain the actual respiratory cycle C1;

[0200] S4 updates the estimated respiratory cycle to the signal-to-noise ratio on the time axis 0-C1;

[0201] S5 updates the breathing cycle to C1, the initial window width to W1, and W1 is equal to NC1;

[0202] S6 performs a sliding window with a step size of C1 and repeats S2-S5 to obtain the actual respiratory cycle length and signal-to-noise ratio;

[0203] Optionally, the adaptive adjustment window width of the EIT real-time signal is multi-channel parallel processing, and the signal-to-noise ratio of each channel is averaged to obtain the overall signal-to-noise ratio of the current window width.

[0204] In one specific embodiment, EIT signal quality determines imaging quality. A high-quality lung EIT signal is essential for accurately reflecting lung ventilation. Due to the complex clinical measurement environment, actual EIT signals often contain a certain amount of noise, making signal quality difficult to guarantee. Commonly used low-pass filtering can remove most high-frequency noise, but when the signal is contaminated to a certain degree by noise, even after low-pass filtering, the signal will be severely distorted and fail to accurately represent lung ventilation. In this case, the examiner often needs to promptly change the measurement strategy or use other denoising methods to obtain a valid EIT signal. However, it is difficult to distinguish the signal quality at this time using only the global or regional impedance change curves provided by the EIT device. Therefore, it is necessary to propose a real-time lung EIT signal quality assessment method that provides timely feedback on EIT signal quality changes. The present invention classifies lung EIT signal quality into two categories: high and low. Using the pure lung ventilation signal and EIT signal generated by the above method, the EIT signal is constructed into lung EIT signals with different signal-to-noise ratios. By analyzing the error magnitude of the lung ventilation EIT image caused by signals with different signal-to-noise ratios, the signal-to-noise ratio classification threshold for determining the current signal quality is determined. At the same time, a real-time SNR estimation method for lung EIT signals is performed based on wavelet decomposition. The real-time SNR is compared with the SNR classification threshold to achieve real-time quality assessment of lung EIT signals.

[0205] In a specific embodiment, determining the signal-to-noise ratio classification threshold of the lung EIT signal requires a sufficient lung EIT signal dataset and a parameter reflecting the EIT image error:

[0206] Dataset: Using the simulated lung EIT signal generated by the aforementioned method, the required pure lung ventilation signal, denoted as Signal A, was directly generated as the baseline signal. A noise-free lung ventilation signal with the blood perfusion signal added was denoted as Signal B, serving as the noise-free lung EIT signal. Under resting conditions, the normal adult cardiac cycle ranges from 0.6s to 1s, the I / E ratio ranges from 1 / 2 to 1 / 1.5, and the respiratory cycle ranges from 3s to 5s. To avoid random results caused by a single signal, different cardiac cycles, I / E ratios, and respiratory cycles were set within the normal range. Multiple pure lung ventilation and lung EIT signals, namely Signals A and B, were generated. Each signal length was 10 times the corresponding number of cycles. Signals A and B with the same cardiac cycle, I / E ratio, and respiratory cycle were grouped together. Different degrees of electronic measurement noise (the noise type was set to uniform Gaussian noise, which is reasonable as a general signal-to-noise ratio evaluation method) were added to signal B in each group to generate a series of lung EIT signals with different signal-to-noise ratios. Signal A in the same group served as the benchmark signal for lung EIT signals with different signal-to-noise ratios within the group.

[0207] Parameters: Clinical parameters and image parameters of EIT lung ventilation are used to reflect the impact of different levels of noise on EIT lung ventilation images.

[0208] Clinical parameters of EIT lung ventilation: The regional ventilation delay index (RVD) of the lung area,

[0209] (7)

[0210] (8)

[0211] Where L represents the lung area (the pixel area greater than 25% of the maximum pixel value in the end-inspiratory EIT image is recorded as the lung area), t i,40% is the time for the i-th pixel in the lung area to reach 40% of the maximum impedance change of the pixel, T inspiration,global is the inspiratory time of the overall impedance change curve; N A is the total number of pixels in the lung area of ​​the end-inspiratory EIT image generated by signal A, RVD i and RVD A,i are the current signal and signal A’s RVD is the ventilation delay index of each pixel, both of which are calculated based on the lung area identified by signal A. erroris the lung area ventilation delay index error, which is used to evaluate the difference in ventilation distribution between the current signal and signal A.

[0212] EIT lung ventilation image parameters: shape deformation (SD),

[0213] (9)

[0214] Among them, L and L A are the lung region vectors for the EIT images generated by the current signal and signal A, respectively. The lung region vector is a binary vector where the value of the corresponding element in the vector for pixels within the lung region is 1, and the value for pixels outside the lung region is 0. SD is used to evaluate the difference between the current signal and signal A on the end-inspiratory EIT image.

[0215] RVD error The error in the EIT image process and the error in the EIT tidal image are analyzed from two aspects: the error in the EIT image process and the error in the EIT tidal image.

[0216] According to clinical routine, the noisy lung EIT signal is first low-pass filtered to remove noise and blood perfusion signals, and then the RVD of the signal is calculated. error In order to reduce the randomness of the results caused by the randomness of the noise, for each signal-to-noise ratio signal, its signal-to-noise ratio is fixed, and 100 groups of different random noises are added respectively. Correspondingly, 100 groups of parameter values ​​are obtained and averaged as the RVD of the current lung EIT signal at this signal-to-noise ratio. error and SD. By analyzing the RVD of signals with different signal-to-noise ratios error The signal-to-noise ratio threshold is determined based on the variation characteristics of SD with the signal-to-noise ratio.

[0217] In a specific embodiment, the signal-to-noise ratio of the lung EIT signal is estimated in real time:

[0218] Classic signal quality assessment methods typically use a fixed-length, fixed-speed sliding window to process physiological signals. Specifically, a fixed-length time segment is selected from a continuous physiological signal stream and the signal within this time segment is analyzed and calculated. Although pulmonary EIT signals exhibit distinct periodic characteristics, the length of the respiratory cycle and the stability of the respiratory signal amplitude vary significantly, depending on individual breathing habits and the respiratory state of the same individual. Using a fixed-length sliding window can lead to variations in the target signal loss caused by wavelet decomposition, reducing the stability and accuracy of the signal-to-noise ratio estimation.

[0219] Aiming at the characteristics of lung EIT signals, this paper proposes a signal quality assessment method based on wavelet decomposition, which monitors the respiratory cycle in real time and adaptively adjusts the window width to achieve better stability. The method is as follows: based on the normal respiratory cycle range, the initial value of the respiratory cycle C is set to c0 seconds. At the same time, the signal analysis window width W is always set to N respiratory cycles, that is, the length is NC. At this time, the initial window width w0 is Nc0 seconds. First, the signal-to-noise ratio of the signal within the w0 window width is estimated to obtain an estimated value, which is recorded as snr0. Then, the signal within the w0 window width is detected to obtain the actual average period c1 of this part of the signal. At this time, the signal-to-noise ratio of the signal from 0 to c1 on the time axis is estimated to be snr0, the respiratory cycle C is updated to c1, and the window width W is updated to w1 = Nc1. Then, the new window slides forward with a step size of c1, that is, the sliding step size is c1. The above steps are repeated to continuously update the respiratory cycle C and window width W to achieve real-time signal-to-noise ratio estimation of lung EIT signals. In this paper, the initial value of the respiratory cycle C is set to 4 seconds. To minimize time delay, the window width W is kept constant at 2 respiratory cycles. (A longer window width encompasses more cycles and improves the stability of the signal-to-noise ratio estimation, but also increases time delay.) Two points should be noted about this method. First, because EIT signals are multi-channel, signal analysis employs multi-channel parallel processing. This involves using wavelet decomposition to process the EIT signal in each channel within the window width, separating the target signal from the noise. The signal-to-noise ratio (SNR) for each channel is then calculated and averaged to form the overall SNR for the signal within the window width. Second, due to the window width, this method introduces a certain delay in the SNR estimation.

[0220] In one specific embodiment, multiple sets of clean lung ventilation signals and lung EIT signals (signal A and signal B) are used to determine the signal-to-noise ratio classification threshold for the lung EIT signal. The cardiac cycle is set to 0.8s, the inspiration-expiration ratio is set to 1 / 1.5 and 1 / 2, and the respiratory cycle is set to 3s, 4s, and 5s. Six sets of clean lung ventilation signals and lung EIT signals are generated, namely six sets of signal A and signal B. The length of each set of signals is 10 times the corresponding number of cycles. The boundary voltage changes of all signals are The sum of the absolute values ​​of ; ,like Figure 8 shown.

[0221] Add different levels of noise to signal B to generate 50 sets of lung EIT signals with uniformly increased signal-to-noise ratios ranging from 0 to 25dB. Calculate the clinical parameter RVD of EIT lung ventilation error and image parameter SD, reflecting the impact of different degrees of noise on lung ventilation signals.

[0222] Table 2 first gives the RVD of 6 groups of signal B errorand SD, which are used to characterize the effect of blood perfusion signal on lung ventilation imaging in the absence of noise. The results in the table show that RVD is greatly affected by blood perfusion signal. error The SD is between 3% and 5%; the SD is relatively small, reaching a maximum of only 2%. However, the SD results vary widely. This is because if the difference in the blood perfusion signal at end-expiration and end-inspiration is large, the differential imaging result at these two moments will be significantly affected by the blood perfusion signal, resulting in a larger SD. Conversely, the SD will be smaller. The blood perfusion signal has a significant and unstable influence on lung ventilation monitoring. Therefore, in actual clinical applications, low-pass filtering is often used to remove the blood perfusion signal to ensure stable EIT imaging.

[0223] Table 2. Differences between noise-free lung EIT signals and pure lung ventilation signals

[0224]

[0225] According to clinical routine operation, the noisy lung EIT signal was low-pass filtered with a cutoff frequency of 1 Hz to filter out noise and blood perfusion signals, and then the RVD of the signal was calculated. error In order to reduce the randomness of the results caused by the randomness of the noise, for each signal-to-noise ratio signal, its signal-to-noise ratio is fixed, and 100 groups of different random noises are added respectively. Correspondingly, 100 groups of parameter values ​​are obtained and averaged as the RVD of the current lung EIT signal at this signal-to-noise ratio. error RVD of 50 groups of signals with different signal-to-noise ratios in the range of 0~25dB error The variation of SD with signal-to-noise ratio is shown in Figure 9 shown.

[0226] Figure 9 Displays the overall RVD of the minimum period and minimum I / O ratio signal error The curve is at the lowest position, the maximum cycle and the maximum inhalation-exhalation ratio signal the overall RVD error The curve is at its highest position, indicating that SD is less affected by cycle and I / O ratio, and the curve is concentrated. The minimum cycle and minimum I / O ratio signals reflect the maximum error range, so when comprehensively analyzing the trends of all signals, focus on these signals. Figure 9 In the signal with a signal-to-noise ratio of 10dB, RVD error The maximum values ​​of RVD and SD are 6.58% and 1.84% respectively; for signals with a signal-to-noise ratio higher than 12dB, error and SD are small and change slowly; for signals with a signal-to-noise ratio lower than 8dB, RVD error Therefore, this paper recommends using 10dB as the SNR threshold, which can ensure the RVD of high-quality signals. errorThe SNR and SD should not exceed 6.58% and 1.84% respectively. A 4dB-wide buffer zone, 8-12dB, is defined with a center of 10dB. One benefit of setting a buffer zone is that, due to the instability of human respiratory motion, the breathing amplitude of actual measurement signals fluctuates, often causing fluctuations in the SNR estimate. Setting a buffer zone can enhance the stability of signal quality assessment results. When the signal-to-noise ratio (SNR) is greater than 12dB, the signal quality is considered high, and when it is less than 8dB, the signal quality is considered low. If the SNR remains between 8-12dB for a prolonged period, a warning message is issued.

[0227] In a specific embodiment, this section uses simulated signals and actual measured signals to verify the accuracy of the signal-to-noise ratio estimation method.

[0228] Simulation signal: Based on the EIT signal simulation method, a lung EIT signal is generated with a constant respiratory cycle of 4 seconds, an inspiratory-expiratory ratio of 1 / 2, and a constant cardiac cycle of 0.8 seconds. The entire signal is divided into three parts, and noise is added to make the signal-to-noise ratio of the three parts 0dB, 10dB, and 20dB respectively. The adaptive window width method is used to estimate the signal-to-noise ratio. The results are as follows: Figure 10 As shown in the figure. When the signal-to-noise ratio (SNR) changes suddenly, the estimated curve has a transition band of a certain length. This transition band, caused by window sliding, does not exceed the window width and is within an acceptable range for clinical applications. Under stable estimation conditions, the maximum error in the estimated value is 1.227dB when the SNR reference value is 0dB; 0.5888dB when the SNR reference value is 10dB; and 0.990dB when the SNR reference value is 20dB.

[0229] Actual measurement signals: In a laboratory environment, using EIT equipment, we collected EIT signals from the lungs of healthy adults. These signals are considered noise-free and include various breathing states, including calm breathing, deep breathing, and rapid breathing.

[0230] The measured signal under calm breathing conditions was divided into three segments. Random noise was added to adjust the signal-to-noise ratios to 0dB, 10dB, and 20dB, respectively. The adaptive window width method was used to estimate the signal-to-noise ratio. Because the actual measured signal has significant variability, the estimation error is slightly larger than that of the simulated signal. Under stable estimation conditions, the maximum error in the estimated value was 2.860dB when the reference signal-to-noise ratio was 0dB; 1.392dB when the reference signal-to-noise ratio was 10dB; and 1.475dB when the reference signal-to-noise ratio was 20dB. The overall trend was correct.

[0231] A lung EIT signal containing various respiratory states, including deep breathing, calm breathing, and rapid breathing, was added to maintain a uniform signal-to-noise ratio (SNR) of 10 dB for each EIT signal. The adaptive window width method was used to estimate the SNR. The estimated SNR was stable and unaffected by changes in the respiratory cycle, with a maximum overall error of 1.862 dB.

[0232] Figure 2 The schematic diagram of the system for constructing the EIT impedance changes of each lung unit provided by an embodiment of the present invention specifically includes:

[0233] Acquisition unit: Acquire the subject's EIT lung ventilation image data, including N complete respiratory cycles, where N is a natural number greater than 1, and the distance from the unit center of each unit in the lung area to the left and right bronchial orifices;

[0234] A calculation unit: calculates a maximum time constant and a minimum time constant of a lung region based on the lung ventilation image data;

[0235] The latest change starting point and the earliest change starting point of the lung impedance are obtained based on the maximum time constant and the minimum time constant of the lung area;

[0236] Relationship unit: Based on the distance from the unit center of each unit in the lung region to the left and right bronchi, the maximum time constant and minimum time constant of the lung region, and the latest change starting point and earliest change starting point of the lung region impedance, a functional relationship fitting calculation is performed to obtain the time constant and impedance change starting point of each unit in the lung region;

[0237] Impedance unit: The impedance change of each unit in the lung area is obtained based on the time constant of the unit in the lung area and the starting point of the impedance change.

[0238] An embodiment of the present invention provides a system for constructing a lung EIT signal simulation model, specifically comprising:

[0239] Acquisition module: acquires human chest images and conductivity change curve data of blood perfusion;

[0240] Construction module: constructing a three-dimensional model based on the chest image to obtain a finite element three-dimensional chest model;

[0241] Excitation module: performing excitation measurement on the finite element three-dimensional chest model to obtain a sensitivity matrix;

[0242] Impedance module: Based on the above-mentioned method for constructing the EIT impedance changes of each lung unit and the finite element three-dimensional chest model, the electrical impedance changes of each lung unit and the total electrical impedance changes of the first lung region are obtained;

[0243] A blood flow module: processing the conductivity change curve data of the blood perfusion based on the total electrical impedance change of the first lung area to obtain a processed impedance change of the blood perfusion;

[0244] performing impedance addition on the electrical impedance change of each lung unit based on the impedance change of the processed blood perfusion to obtain the electrical impedance change of each unit after the addition;

[0245] Simulation module: based on the sensitivity matrix and the change of electrical impedance of each unit after addition, a lung EIT simulation signal is calculated.

[0246] Figure 3 A schematic diagram of a computer device provided by an embodiment of the present invention specifically includes:

[0247] A memory and a processor; the memory is used to store program instructions; the processor is used to call the program instructions, and when the program instructions are executed, any one of the above-mentioned methods for constructing the EIT impedance changes of each lung unit or the lung EIT signal simulation model is executed.

[0248] The disclosed embodiments of the present invention also provide a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, any one of the above-mentioned methods for constructing the EIT impedance changes of each lung unit or constructing a lung EIT signal simulation model is implemented.

[0249] The validation results of this validation example demonstrate that assigning inherent weights to indications can improve the performance of the present method compared to the default settings. Those skilled in the art will readily appreciate that, for ease of description and brevity, the specific operating processes of the systems, devices, and units described above can be referenced to the corresponding processes in the aforementioned method embodiments and will not be further elaborated upon here. It should be understood that the disclosed systems, devices, and methods can be implemented in other ways within the several embodiments provided herein. For example, the device embodiments described above are merely illustrative. For example, the division of units described is merely a logical functional division. In actual implementation, other divisions may be employed, such as combining or integrating multiple units or components into another system, or omitting or disabling certain features. Furthermore, the coupling, direct coupling, or communication connection shown or discussed may be through interfaces, indirect coupling, or communication connection between devices or units, and may be electrical, mechanical, or other forms. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the units may be selected to achieve the objectives of the present embodiment as needed. In addition, the functional units in the various embodiments of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated units may be implemented in the form of hardware or in the form of software functional units. Those skilled in the art will understand that all or part of the steps in the various methods of the above-mentioned embodiments may be completed by instructing the relevant hardware through a program, and the program may be stored in a computer-readable storage medium, which may include: a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0250] Those skilled in the art will understand that all or part of the steps in the above-mentioned embodiment method can be implemented by instructing the relevant hardware through a program, and the program can be stored in a computer-readable storage medium. The above-mentioned medium storage can be a read-only memory, a disk or an optical disk, etc.

[0251] The above is a detailed introduction to a computer device provided by the present invention. For those skilled in the art, according to the concept of the embodiments of the present invention, there may be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for constructing EIT impedance changes of each lung unit, characterized in that: include: Obtain the test subject's EIT lung ventilation image data, including N complete respiratory cycles, where N is a natural number greater than 1, and the distance from the unit center of each unit in the lung region to the left and right bronchial orifices; Calculating a maximum time constant and a minimum time constant of a lung region based on the lung ventilation image data; The latest change starting point and the earliest change starting point of the lung impedance are obtained based on the maximum time constant and the minimum time constant of the lung area; Based on the distance from the unit center of each unit in the lung region to the left and right bronchi, the maximum time constant and the minimum time constant of the lung region, and the latest change starting point and the earliest change starting point of the lung region impedance, a functional relationship fitting calculation is performed to obtain the time constant and the impedance change starting point of each unit in the lung region; The impedance change of each unit in the lung area is obtained based on the time constant and the impedance change starting point of each unit in the lung area.

2. The method for constructing the EIT impedance change of each lung unit according to claim 1, characterized in that: The maximum time constant and the minimum time constant of the lung region are calculated as follows: based on the lung ventilation image data, a time constant of impedance change of each pixel in the lung region and a time constant of impedance change of all pixels in a complete respiratory cycle are obtained, and a ratio of the time constant of impedance change of each pixel in the lung region to the time constant of impedance change of all pixels is calculated to obtain N groups of time constant ratios; Calculating the average of the N groups of time constant ratios by cycle number to obtain the maximum time constant and the minimum time constant of the lung region; Optionally, the functional relationship between the time constant of each unit and the distance is calculated as follows: in, represents the unit time constant, a and b represent the ratio of the minimum and maximum time constants of the lung unit to the time constant of the overall impedance change, aT represents the maximum time constant of the lung unit, d represents the normalized distance between the unit center and the left and right main bronchial orifices, and T represents the time constant of the overall impedance change; Optionally, the functional relationship between the starting point of each unit change and the distance is calculated as follows: Where s represents the starting point of unit change, d represents the normalized distance between the unit center and the left and right main bronchial orifices, T represents the time constant of overall impedance change, and b represents the ratio of the maximum time constant of the lung unit to the time constant of overall impedance change. Optionally, the normalized distance d between the unit center and the left and right bronchi is Or s is a nonlinear relationship, and the d constraint is obtained by the nonlinear function to obtain the constrained normalized d, and the constrained normalized d is equal to Or s to perform functional relationship fitting calculation.

3. The method for constructing the EIT impedance change of each lung unit according to claim 2, characterized in that: The nonlinear function is constrained by a hyperbolic tangent function to obtain a constrained normalized d; Optionally, the process of constraining d is: obtaining d data, nonlinearizing function f(x); Normalize the function f(x) to obtain the normalized function g(x); The d data is calculated by the normalized function g(x) to obtain the constrained normalized d; Optionally, the nonlinearization function is constructed based on a hyperbolic tangent function and is expressed as: in, and are the upper and lower limits of the domain, and x represents the independent variable of the function; Optionally, the process of constraining d further includes adjusting the degree of nonlinearity, and the nonlinearity adjustment is performed by adjusting the domain of f(x), including: defining any two parameters p0 and p1, wherein the two parameters satisfy a constraint condition; converting the domain of f(x) based on the constraint condition to obtain a converted domain, and performing nonlinearity adjustment in the converted domain by the spacing between p0 and p1; the constraint condition is expressed as: in, for and The minimum spacing; The function f(x) is normalized after domain conversion to obtain a normalized function; Optionally, the nonlinear function performs d constraint to adjust the distance distribution of different units, and the distribution adjustment of the time constants and the starting points of the changes of different units is obtained by adjusting the distance distribution.

4. The method for constructing the EIT impedance change of each lung unit according to claim 1, characterized in that: The distances from the unit center of each unit in the lung region to the left and right bronchial openings include the distance from the unit center of the left lung unit to the left bronchial opening and the distance from the unit center of the right lung unit to the bronchial opening; the time constant and impedance change starting point of each unit in the lung region are obtained by performing fitting calculation based on the distance from the unit center of the left lung unit to the left bronchial opening, the distance from the unit center of the right lung unit to the bronchial opening, the maximum time constant and the minimum time constant of the lung region, and the latest change starting point and the earliest change starting point of the lung region impedance; Optionally, the lung region is a pixel region of the end-inspiratory image of the lung ventilation image data having a pixel value greater than a preset threshold; The method further includes constructing an impedance change of the left lung region or the right lung region, and obtaining a distance from a unit center of the left or right lung unit to the left or right bronchial orifice; Based on the lung ventilation image data, the maximum time constant and the minimum time constant of the left or right lung area are obtained; The latest change starting point and the earliest change starting point of the left or right lung region are obtained based on the maximum time constant and the minimum time constant of the left or right lung region; The impedance change of the left or right lung area is obtained by fitting and calculating based on the distance from the unit center of the left or right lung unit to the left or right bronchial orifice, the maximum time constant and the minimum time constant of the left or right lung area, the latest change starting point and the earliest change starting point of the left or right lung area; Optionally, the lung units include lungs, heart, spine, ribs, sternum, and trachea.

5. A method for constructing a lung EIT signal simulation model, characterized in that: include: Obtain human chest imaging and conductivity change curve data of blood perfusion; Constructing a three-dimensional model based on the chest image to obtain a finite element three-dimensional chest model; Performing excitation measurement on the finite element three-dimensional chest model to obtain a sensitivity matrix; Based on the method for constructing the EIT impedance change of each lung unit and the finite element three-dimensional chest model described in any one of claims 1 to 4, the electrical impedance change of each lung unit and the total electrical impedance change of the first lung region are obtained; performing data processing on the conductivity change curve data of the blood perfusion based on the total electrical impedance change of the first lung area to obtain a processed impedance change of the blood perfusion; performing impedance addition on the electrical impedance change of each lung unit based on the impedance change of the processed blood perfusion to obtain the electrical impedance change of each unit after the addition; The lung EIT simulation signal is calculated based on the sensitivity matrix and the change in electrical impedance of each unit after addition.

6. The method for constructing a lung EIT signal simulation model according to claim 5, characterized in that: The impedance addition is to add an impedance change to the electrical impedance change of each unit that is opposite to the change in blood perfusion impedance and has an amplitude of half the amplitude of the change in blood perfusion impedance; Optionally, the processed impedance change of the blood perfusion is added to the heart unit to obtain a processed heart unit impedance change, and a pulmonary EIT simulation signal is calculated based on the sensitivity matrix, the processed heart unit impedance change, and the added impedance change of each unit; Optionally, the data processing is to convert the conductivity of the blood perfusion into the blood perfusion impedance, and then adjust the variation of the blood perfusion impedance to 5% of the variation of the total electrical impedance to obtain the processed blood perfusion impedance variation; Optionally, the 5 percent of the variation amplitude is 5 percent of the product of the lung conductivity variation and the lung volume; Optionally, the total electrical impedance change of the first lung area is obtained by multiplying the electrical impedance change of each unit by the volume of each unit and then summing them up; Optionally, the construction of the finite element three-dimensional chest model includes chest model reconstruction, setting EIT electrodes, and conductivity configuration, wherein a three-dimensional chest model is obtained after three-dimensional reconstruction of chest images, EIT electrodes are set in the three-dimensional chest model, and conductivity is configured for different tissue regions in the three-dimensional chest model to obtain the finite element three-dimensional chest model; Optionally, performing excitation measurement on the EIT electrode sheet in the finite element three-dimensional chest model to obtain a sensitivity matrix; Optionally, the chest model reconstruction is performed by dividing the chest into different tissue regions, reconstructing the different tissue regions to obtain a chest model, and then setting EIT electrodes in the chest model to obtain a three-dimensional chest model; Optionally, the different tissue regions include one or more of the following: chest, lungs, heart, spine, ribs, sternum, trachea; Optionally, the different tissue regions further include other regions, and the other regions are treated as a whole, and the electrical conductivity of the other regions is uniformly configured; Optionally, the EIT simulation signal includes one or more of the following: a pure lung ventilation EIT signal, a pure blood perfusion EIT signal, and a mixed lung EIT signal.

7. A lung EIT signal simulation model, characterized in that: The EIT signal simulation model is obtained by the method for constructing a lung EIT signal simulation model according to any one of claims 5-6.

8. A computer program product comprising a computer program or instructions, characterized in that: The computer program or instructions are executed by a processor to implement the method for constructing the EIT impedance changes of each lung unit as described in any one of claims 1-4, or to execute the method for constructing the lung EIT signal simulation model as described in any one of claims 5-6.

9. A computer device comprising a memory, a processor, and a computer program or instruction stored in the memory, wherein: The computer program or instructions are executed by a processor to implement the method for constructing the EIT impedance changes of each lung unit as described in any one of claims 1-4, or to execute the method for constructing the lung EIT signal simulation model as described in any one of claims 5-6.

10. A computer-readable storage medium having a computer program or instruction stored thereon, characterized in that: The computer program or instructions are executed by a processor to implement the method for constructing the EIT impedance changes of each lung unit as described in any one of claims 1-4, or to execute the method for constructing the lung EIT signal simulation model as described in any one of claims 5-6.

Images