Human erythrocyte health status monitoring method based on artificial intelligence and optical measurement

Abstract

The invention discloses a human red blood cell health status monitoring method based on artificial intelligence and optical measurement, which comprises the following steps of: enabling single-spectrum light to enter a human red blood cell culture solution, and constructing an optical radiation transmission model in the human red blood cell culture solution based on a T-Matrix model to obtain emergent optical radiation characteristics of the cell culture solution; the method comprises the following steps: analyzing an influence rule of a human body cell health state on cell emergent light radiation characteristics, determining light radiation characteristic parameters sensitive to cell health state changes, and constructing a cell morphological characteristic and cell culture fluid emergent light radiation characteristic association database based on an orthogonal design test method; establishing an association agent model between the human body cell health state and the cell emergent light radiation characteristic parameters based on an RBF neural network method; based on the proxy model, the radiation characteristic condition of the emergent light after the single-spectrum light enters the human red blood cell culture solution is measured through photoelectric measurement equipment, the health state of the human red blood cells is judged, and clinicians are assisted to master the health state of patients more intelligently and more accurately.

Description

Technical Field

[0001] The present invention belongs to the interdisciplinary field of biomedicine, optical engineering, artificial intelligence, etc., and relates to a method for monitoring the health status of human red blood cells based on artificial intelligence and optical measurement technology. Background Art

[0002] The composition and physical and chemical properties of blood are closely related to multiple human systems, including the nervous system, digestive system, and endocrine system, etc. Clinically, blood tests are an important means for diagnosing diseases such as cardiovascular diseases, digestive system diseases, urinary system diseases, endocrine system diseases, and metabolic disorders. Among these blood components, red blood cells are one of the key cells for maintaining life, and their main function is to carry oxygen. The normal performance of this function depends on the special shape of red blood cells (biconcave disc shape), the deformability of their membranes, and the stability of the hemoglobin structure inside the cells. Therefore, maintaining the health status of red blood cells is crucial for ensuring sufficient oxygen supply and maintaining human health.

[0003] Currently, the examination of the health status of human red blood cells mostly adopts sampling and combines microscopic observation to judge whether the cells are healthy. Inevitably, the cell state will be affected during the measurement process. Summary of the Invention

[0004] In order to overcome the problems existing in the prior art, the purpose of the present invention is to provide a method for monitoring the health status of human red blood cells based on artificial intelligence and optical measurement technology. By measuring the outgoing light radiation characteristics of human red blood cell culture solutions in different health states, and then combining the T-Matrix model and artificial intelligence technology to construct an association proxy model between the health status of human red blood cells and the outgoing light radiation characteristics of the culture solution. Finally, based on this proxy model, the outgoing light radiation characteristic signal of the cells can be obtained through cell sample test measurement, and the health status of red blood cells can be judged.

[0005] The purpose of the present invention is achieved through the following technical solutions:

[0006] A method for monitoring the health status of human red blood cells based on artificial intelligence and optical measurement technology. First, single-spectrum light is incident on the culture solution of human red blood cells, and a light radiation transmission model of the culture solution of human red blood cells based on the T-Matrix model is constructed to obtain the characteristics of the light radiation emerging from the cell culture solution (including the light scattering phase function, light extinction factor, light absorption factor, light scattering albedo, etc.). Then, analyze the influence law of the health status of human cells on the characteristics of the light radiation emerging from the cells, determine the light radiation characteristic parameters sensitive to the change of cell health status, establish a database based on the orthogonal design experiment method, and construct an association proxy model between the health status of human cells and the light radiation characteristic parameters of the light radiation emerging from the cells based on the RBF neural network method. Finally, based on the constructed proxy model, by measuring the light radiation characteristic parameter situation after the single-spectrum light is incident on the culture solution of human red blood cells through an optoelectronic measurement device, the health status of human red blood cells can be judged.

[0007] The specific steps of the present invention are as follows:

[0008] Step 1, according to the clinical hospital inspection experience, determine the association characteristics between the morphological parameters of human red blood cells (such as cell diameter, shape, etc.) and the cell health status. For example, the morphology of mature healthy red blood cells is biconcave or single concave disc-shaped, the edge thickness is about 2μm, the central thickness is about 1μm, the diameter is about 6-9μm, and the pathological red blood cells usually exceed the normal range (such as diameter ≥ 10μm), and the morphology is abnormal, showing spherical, oval, sickle-shaped, etc.

[0009] Step 2, according to the measurement requirements of the light radiation characteristics of human red blood cells, suspend the human red blood cells to be measured in a purified human serum culture medium, and continuously stir through a magnetic stirrer to make the cells in a uniformly dispersed and suspended state.

[0010] Step 3, construct a light radiation transmission model of the culture solution of human red blood cells based on the T matrix model (T-Matrix), and measure and obtain the light radiation characteristic signals emerging from the culture solution of human red blood cells through an optical measurement instrument, such as the light scattering phase function, light extinction factor, light absorption factor, light scattering albedo, etc.

[0011] Step 4, conduct a sensitivity analysis of the light radiation characteristic signals emerging from the culture solution of human red blood cells, that is, analyze the influence law of the morphological parameters of human red blood cells on the light radiation characteristic signals emerging from the cell culture solution, and determine the light radiation characteristic signals sensitive to the cell morphological changes.

[0012] Step 5, construct a database for the association between cell morphology and the light radiation characteristic signals emerging from the cell culture solution based on the orthogonal design experiment method, and construct an association proxy model between cell morphology and the light radiation characteristic signals emerging from the cell culture solution based on the RBF neural network technology.

[0013] Step 6: Build a test platform for monitoring the health status of human red blood cells. Measure the signal of the outgoing light radiation characteristics after a single-spectrum light is incident on the human red blood cell culture solution through an optoelectronic measurement device, and combine the constructed surrogate model to judge the health status of human red blood cells.

[0014] The present invention combines optical measurement technology and artificial intelligence technology to propose a non-contact method for monitoring the health status of human red blood cells. First, a single-spectrum light is incident on the human red blood cell culture solution, and a light radiation transmission model in the human red blood cell culture solution based on the T-Matrix model is constructed to calculate and obtain the light radiation characteristics of the cell culture solution (including the light scattering phase function, light extinction factor, light absorption factor, light scattering albedo, etc.). Then, analyze the influence law of the health status of human cells on the light radiation characteristics of cell outgoing light, determine the light radiation characteristic parameters sensitive to the change of cell health status, construct a correlation database between cell morphology and the signal of the light radiation characteristics of cell culture solution outgoing light based on the orthogonal design test method, and establish a correlation surrogate model between the health status of human cells and the light radiation characteristic parameters of cell outgoing light based on the RBF neural network method. Finally, based on the constructed surrogate model, measure the light radiation characteristic parameter situation after a single-spectrum light is incident on the human red blood cell culture solution through an optoelectronic measurement device, and then the health status of human red blood cells can be judged.

[0015] The beneficial effects of the present invention are:

[0016] The T-Matrix theory is used to construct a light scattering characteristic model of human red blood cells. This theory can construct light scattering models of cells with different shapes and obtain the scattering characteristics of cells in the 4π spherical space under monochromatic light incidence. At the same time, the present invention uses the RBF neural network technology to construct a correlation surrogate model between the health status of human cells and the light radiation characteristic parameters of cell outgoing light. This neural network technology can construct a multi-layer neural network structure according to the complexity of the problem, and then can adapt to the construction of surrogate models of cell light radiation characteristics with different complexities and ensure the accuracy of the surrogate model.

[0017] The present invention fully utilizes optical measurement technology and artificial intelligence technology to empower the development of human red blood cell health detection technology in the field of biomedicine, and helps clinicians to more intelligently and accurately master the health status of patients. Description of the Drawings

[0018] Figure 1 It is a schematic flowchart of the present invention.

[0019] Figure 2 It is a structure diagram of the RBF neural network.

[0020] Figure 3 It is a diagram of the human red blood cell model.

[0021] Figure 4 It is a graph showing the influence law of different cell diameters on the light scattering phase function of the culture medium.

[0022] Figure 5 It is a graph showing the influence law of different cell diameters on the light extinction factor of the culture medium.

[0023] Figure 6 It is a graph showing the influence law of different cell diameters on the light absorption factor of the culture medium. Specific implementation manners

[0024] As Figure 1 shown, a method for monitoring the health status of human red blood cells based on artificial intelligence and optical measurement technology. First, single-spectrum light is incident on the culture medium of human red blood cells, and a light radiation transmission model in the culture medium of human red blood cells based on the T-Matrix model is constructed to obtain the light radiation characteristics of the outgoing light from the cell culture medium (including light scattering phase function, light extinction factor, light absorption factor, light scattering albedo, etc.). Then, analyze the influence law of the health status of human cells on the light radiation characteristics of the outgoing light from the cells, determine the light radiation characteristic parameters sensitive to the change of cell health status, establish a database based on the orthogonal design test method, and construct an association proxy model between the health status of human cells and the light radiation characteristic parameters of the outgoing light from the cells based on the RBF neural network method. Finally, based on the constructed proxy model, by measuring the light radiation characteristic parameter situation of the outgoing light after single-spectrum light is incident on the culture medium of human red blood cells through an optoelectronic measurement device, the health status of human red blood cells can be judged.

[0025] A method for monitoring the health status of human red blood cells based on artificial intelligence and optical measurement technology. The specific operation steps of this method are as follows:

[0026] Step 1, according to the clinical hospital inspection experience, determine the association characteristics between the morphological parameters of human red blood cells (such as cell diameter, shape, etc.) and the cell health status. For example, the morphology of mature healthy red blood cells is biconcave or single-concave disc-shaped, the edge thickness is about 2μm, the central thickness is about 1μm, the diameter is about 6 - 9μm. The diameter of pathological red blood cells usually exceeds the normal range (such as diameter ≥ 10μm), and the morphology is abnormal, showing spherical, elliptical, sickle-shaped, etc. Figure 3 It is a diagram of the human red blood cell model.

[0027] Step 2, according to the measurement requirements of the light scattering characteristics of human red blood cells, suspend the human red blood cells to be measured in the purified human serum medium, and continuously stir through a magnetic stirrer to make the cells in a uniformly dispersed and suspended state.

[0028] Step 3: Construct a light radiation transmission model in human red blood cell culture medium based on the T-Matrix model, and measure the light radiation characteristic signals emitted from the human red blood cell culture medium through an optical measurement instrument, such as the light scattering phase function, light extinction factor, light absorption factor, light scattering albedo, etc.

[0029] The light radiation transmission model in human red blood cell culture medium constructed by using the T-Matrix model is a calculation method for electromagnetic scattering characteristics of a single, uniform, arbitrarily shaped particle based on Huygens' principle. The expansion coefficients of the scattered vector spherical harmonics in the matrix are expressed by the expansion coefficients of the incident field vector spherical harmonics. The incident field E inc and the scattered field E sca are expanded with vector spherical harmonics:

[0030]

[0031] where m and n are the series of the vector spherical harmonic expansion; ∞ represents positive infinity; r0 is the radius of the circumscribed sphere of the cell; k0 = 2π / λ, which is the wave number in the environmental medium, and λ is the wavelength of the incident light; r is the radius vector where a single cell is located, and its center is the origin of the global coordinate system of the entire system; RgM mn and RgN mn are regular vector wave functions based on Bessel functions; M mn and N mn are vector wave functions expanded from the minimum circumscribed sphere of the scatterer; a mn and b mn represent the expansion coefficients of the incident field; p mn and q mn represent the expansion coefficients of the scattered field. By solving formulas (1) and (2), the light extinction factor Q ext 、light absorption factor Q abs and the light scattering phase function Φ p of a randomly oriented single cell are obtained.

[0032] Figure 4 is the influence law diagram of different cell diameters on the light scattering phase function of the culture medium. Figure 5 is the influence law diagram of different cell diameters on the light extinction factor of the culture medium. Figure 6 is the influence law diagram of different cell diameters on the light absorption factor of the culture medium.

[0033] Step 4: Conduct a sensitivity analysis of the light radiation characteristic signals emitted from the human red blood cell culture medium, that is, analyze the influence law of human red blood cell morphological parameters on the light radiation characteristic signals emitted from the cell culture medium, and determine the light radiation characteristic signals sensitive to cell morphological changes.

[0034] Step 5: Construct a signal correlation database between cell morphology and the emitted light radiation characteristics of cell culture medium based on the orthogonal design experiment method, and construct a signal correlation proxy model between cell morphology and the emitted light radiation characteristics of cell culture medium based on the RBF neural network technology.

[0035] Step 6: Build a monitoring test platform for the health status of human red blood cells. Measure the emitted light radiation characteristic signals after single-spectrum light is incident on the human red blood cell culture medium through optoelectronic measurement equipment, and combine the constructed proxy model to judge the health status of human red blood cells.

[0036] In this embodiment, a physical model of light radiation transmission of human red blood cell culture medium is first designed, and then a mathematical model and a solution method based on the T-Matrix theory are established. The emitted light radiation characteristic signals of the culture medium are obtained through measurement, and a correlation database and a proxy model between the health status of human red blood cells and the emitted light radiation characteristics of cell culture medium are constructed. Finally, the health status of cells can be judged according to the emitted light radiation characteristics of the culture medium measured from cell samples.

[0037] Furthermore, the light radiation transmission process of the cell culture medium described in Step 3 is realized by using the T-Matrix theory.

[0038] Furthermore, the correlation proxy model between the human red blood cell morphology parameters and the emitted light radiation characteristic signals of the cell culture medium in Step 6 is realized by using the radial basis function neural network (RBF) neural network technology. As Figure 2 shown, in the RBF neural network model of the present invention, the input layer is the human red blood cell morphology parameters X = [x1, x2, x3...], and the output layer is the emitted light radiation characteristics of the cell culture medium Y = [y1, y2, y3...].

[0039] The input layer to the hidden layer of the RBF network is a non-linear transformation layer, and the output of the i-th hidden unit h i is:

[0040]

[0041] In the formula, j(·) is the transformation function of the hidden layer, which is a non-linear function with local distribution and radially symmetric attenuation around the center point. In this paper, it is taken as the Gaussian function, i = 1, 2, 3; b i and c i are respectively the base width vector and the center vector of the i-th hidden layer node. The output of the network is realized by the following weighted function:

[0042] Y = ∑w i h i (4)

[0043] In the formula, w i is the connection weight value between the i-th hidden layer and the output layer; Y is the output value of the neural network.

Claims

1. A method for monitoring the health status of human red blood cells based on artificial intelligence and optical measurement, characterized in that, First, a single-spectrum light is incident on a human red blood cell culture solution, and a light radiation transmission model in the human red blood cell culture solution based on the T-Matrix model is constructed to obtain the light radiation characteristics of the cell culture solution exiting; Then, analyze the influence law of the human cell health state on the light radiation characteristics of the cell exiting, determine the light radiation characteristic parameters sensitive to the change of the cell health state, establish a database based on the orthogonal design test method, and construct an association proxy model between the human cell health state and the light radiation characteristic parameters of the cell exiting. Finally, based on the constructed proxy model, measure the light radiation characteristic parameter situation of the cell culture solution exiting after the single-spectrum light is incident on the human red blood cell culture solution through an optoelectronic measurement device to judge the health state of the human red blood cells.

2. The method for monitoring the health status of human red blood cells based on artificial intelligence and optical measurement according to claim 1, wherein, The specific steps are as follows: Step 1, according to clinical tests, determine the association characteristics between the morphological parameters of human red blood cells and the cell health state; Step 2, according to the measurement requirements of the light scattering characteristics of human red blood cells, suspend the human red blood cells to be measured in a purified human serum culture medium, and continuously stir through a magnetic stirrer to make the cells in a uniformly dispersed and suspended state; Step 3, construct a light radiation transmission model in the human red blood cell culture solution based on the T-matrix model, and measure the light radiation characteristic signal exiting from the human red blood cell culture solution through an optical measurement instrument; Step 4, conduct a sensitivity analysis of the light radiation characteristic signal exiting from the human red blood cell culture solution, that is, analyze the influence law of the morphological parameters of human red blood cells on the light radiation characteristic signal of the cell culture solution exiting, and determine the light radiation characteristic signal sensitive to the cell morphological change; Step 5, construct an association database between the cell morphology and the light radiation characteristic signal of the cell culture solution exiting based on the orthogonal design test method, and construct an association proxy model between the cell morphology and the light radiation characteristic signal of the cell culture solution exiting based on the RBF neural network technology; Step 6, build a monitoring test platform for the health state of human red blood cells, measure the light radiation characteristic signal situation of the cell culture solution exiting after the single-spectrum light is incident on the human red blood cell culture solution through an optoelectronic measurement device, and combine the constructed proxy model to judge the health state of the human red blood cells.

3. The method for monitoring the health status of human red blood cells based on artificial intelligence and optical measurement according to claim 2, characterized in that, In Step 1, the morphology of mature healthy red blood cells is biconcave or single-concave disc-shaped, with an edge thickness of 2μm, a central thickness of 1μm, and a diameter of 6-9μm. Pathological red blood cells exceed the normal range, with a diameter ≥10μm and abnormal morphology, presenting spherical, oval, or sickle-shaped.

4. The method for monitoring the health status of human red blood cells based on artificial intelligence and optical measurement according to claim 2, wherein In Step 3, the light radiation characteristic signals include: light scattering phase function, light extinction factor, light absorption factor, and light scattering albedo.

5. The method for monitoring the health status of human red blood cells based on artificial intelligence and optical measurement according to claim 2, wherein, In Step 3, the light radiation transmission model in the human red blood cell culture solution constructed using the T-matrix model is a calculation method for the electromagnetic scattering characteristics of a single, uniform, arbitrarily shaped particle based on the Huygens principle. The expansion coefficients of the spherical harmonics of the scattering vector in the matrix are expressed in terms of the expansion coefficients of the spherical harmonics of the incident field vector. The incident field E inc and the scattered field E sca are expanded using vector spherical harmonics: where m and n are the series of the vector spherical harmonic expansion; ∞ represents positive infinity; r0 is the radius of the circumscribed sphere of the cell; k0 = 2π / λ, which is the wave number in the environmental medium, and λ is the wavelength of the incident light; r is the radius vector where a single cell is located, and its center is the origin of the global coordinate system of the entire system; RgM mn and RgN mn are regular vector wave functions based on the Bessel function; M mn and N mn are vector wave functions expanded by the circumscribed sphere of the scatterer; a mn and b mn represent the expansion coefficients of the incident field; p mn and q mn represent the expansion coefficients of the scattered field. Solving formulas (1) and (2), the light extinction factor Q ext 、light absorption factor Q abs and the light scattering phase function Φ p of a single randomly oriented cell are obtained.

6. The method for monitoring the health status of human red blood cells based on artificial intelligence and optical measurement according to claim 2, wherein In Step 5, a radial basis function neural network is used to construct an association proxy model between the morphological parameters of human red blood cells and the light radiation characteristic signals of the cell culture solution exiting; The RBF neural network model is a three-layer feedforward neural network with radial basis functions as the activation functions of the hidden layer neurons, and it has a topological structure composed of an input layer, a hidden layer, and an output layer; the input layer is the morphological parameters of human red blood cells X = [x1, x2, x3...], and the output layer is the outgoing light radiation characteristics of the cell culture medium Y = [y1, y2, y3...]. The layer from the input layer to the hidden layer of the RBF network is a non-linear transformation layer, and the output of the i-th hidden unit h i is: where \(j(\cdot)\) is the transformation function of the hidden layer, which is a non-linear function with local distribution and radially symmetric attenuation around the center point, and is taken as a Gaussian function, \(i = 1, 2, 3\); \(b\) i and \(c\) i are the base width vector and the center vector of the \(i\)-th hidden layer node respectively. The output of the network is realized by the following weighted function: Y = ∑w i h i (4) where, w i is the connection weight between the i-th hidden layer and the output layer; Y is the output value of the neural network.

Images