A method and system for predicting lung lesions in patients based on partial differential equations

By performing three-dimensional reconstruction of CT tomography and establishing a partial differential equation model, the problem of difficult to predict the development of lung diseases in patients in the prior art is solved, and accurate prediction of lung lesions and auxiliary medical judgments are achieved.

CN112258511BActive Publication Date: 2024-11-29GUANGDONG INST OF INTELLIGENT MFG
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202010937118.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-09-08
Publication Date
2024-11-29
Estimated Expiration
2040-09-08

AI Technical Summary

Technical Problem

The prior art is difficult to effectively predict the development trend of patients' lung diseases through CT tomography information.

Method used

Using a method based on partial differential equation, a partial differential equation model is established to describe the concentration of lesions by three-dimensional reconstruction of CT two-dimensional images, and the finite element method and backward Euler method are used to solve it, and combined with parallel simulation annealing method to optimize the simula Together, patient parameters are obtained to assist in predicting lesions.

Benefits of technology

Accurate prediction of patients' lung lesions is achieved, providing a basis for assisting doctors to judge the development of the disease, and improving the accuracy and practicality of medical predictions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112258511B_ABST
    Figure CN112258511B_ABST
Patent Text Reader

Abstract

The present invention discloses a method and system for predicting lung lesions of patients based on partial differential equations, and the method includes: reconstructing each collected CT two-dimensional image to generate a three-dimensional image, and segmenting and obtaining the grayscale of each original lesion site from the three-dimensional image; combining the Euler equation and the cell action mechanism, establishing a partial differential equation model for describing the lesion concentration of each simulated lesion site; solving the partial differential equation model based on the finite element method and the backward Euler method to obtain the lesion concentration of each simulated lesion site; combining the original lesion sites and the simulated lesion sites, optimizing and fitting the partial differential equation model based on the parallel simulated annealing method, and obtaining the patient parameters contained in the partial differential equation model to assist in predicting the development of the patient's lesion. The embodiment of the present invention can assist doctors in predicting the development of the patient's disease course and meet the growing medical needs of the people.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of medical technology, and in particular to a method and system for predicting lung lesions of patients based on partial differential equations. Background Art

[0002] With the development of modern science and technology, a large amount of data information about some patients with lung diseases during the treatment process has been saved, including CT tomographic image data and chest X-ray image data. These image information can not only intuitively reflect the patient's lung symptoms, but also extract the specific lung disease information of the patient. Prior to this, due to technical limitations, these image information was more used to assist doctors in judging the patient's disease information. In recent years, technicians have focused on studying how to use the cross-application of disciplines such as statistics, mathematics, and computers, and combine CT tomographic image information to enable doctors to effectively predict the direction of the patient's disease, which is also the problem solved by the present invention. Summary of the invention

[0003] The purpose of the present invention is to overcome the shortcomings of the prior art. The present invention provides a method and system for predicting patient lung lesions based on partial differential equations, which can assist doctors in predicting the course of the patient's disease and meet the growing medical needs of the people.

[0004] In order to solve the above problems, the present invention proposes a method for predicting lung lesions of patients based on partial differential equations, the method comprising:

[0005] Performing three-dimensional reconstruction on each collected CT two-dimensional image to generate a three-dimensional image, and segmenting and acquiring the grayscale of each original lesion site from the three-dimensional image;

[0006] Combining the Euler equation and the cell action mechanism, a partial differential equation model was established to describe the lesion concentration at each simulated lesion site;

[0007] Solving the partial differential equation model based on the finite element method and the backward Euler method to obtain the lesion concentration of each simulated lesion site;

[0008] The partial differential equation model is optimized and fitted based on the parallel simulated annealing method in combination with the original lesion sites and the simulated lesion sites to obtain the patient parameters contained in the partial differential equation model to assist in the prediction of the development of the patient's lesions.

[0009] Optionally, the three-dimensional reconstruction of each collected CT two-dimensional image to generate a three-dimensional image includes:

[0010] Arrange the CT two-dimensional images on the same horizontal line according to their respective predetermined orders and longitudinal positions to generate an initial lung three-dimensional image;

[0011] Extracting the grayscale data of the bitmap pixels of each layer of the initial lung three-dimensional image, and constructing a three-dimensional finite element mesh based on the grayscale data of the bitmap pixels of each layer of the image;

[0012] The three-dimensional finite element grid is interpolated based on the cubic spline interpolation method to obtain a final three-dimensional lung image.

[0013] Optionally, the step of segmenting and acquiring the grayscale of each original lesion site from the three-dimensional image includes:

[0014] Segmenting each original lesion site from the final three-dimensional lung image, and extracting grayscale data of each original lesion site;

[0015] The grayscale data of each original lesion site is normalized to obtain the specific value corresponding to each original lesion site.

[0016] Optionally, the partial differential equation model is:

[0017]

[0018]

[0019] Where n is the time at time t and The lesion concentration of the simulated lesion site collected at the position, D is the diffusion coefficient, is the gradient operator, f is the resultant force on the simulated lesion site represented by n, g(n,v) is the complex growth function of the simulated lesion site represented by n, and v is the parameter affecting the external factors.

[0020] Optionally, solving the partial differential equation model based on the finite element method and the backward Euler method to obtain the lesion concentration of each simulated lesion site includes:

[0021] Solving the linear part of the partial differential equation model under the condition of half a time step to obtain a temporary lesion concentration of the current simulated lesion site;

[0022] Taking a time step as a condition, solving the nonlinear part of the partial differential equation model in combination with the temporary lesion concentration of the current simulated lesion site, and obtaining the temporary lesion concentration of the next simulated lesion site;

[0023] The linear part of the partial differential equation model is solved with half a time step as a condition and combined with the temporary lesion concentration of the next simulated lesion site to obtain the lesion concentration of the next simulated lesion site.

[0024] In addition, an embodiment of the present invention further provides a patient lung lesion prediction system based on partial differential equations, the system comprising:

[0025] A segmentation module is used to perform three-dimensional reconstruction on each collected CT two-dimensional image to generate a three-dimensional image, and to segment and obtain the grayscale of each original lesion site from the three-dimensional image;

[0026] Establishing a module for combining the Euler equation and the cell action mechanism to establish a partial differential equation model for describing the lesion concentration at each simulated lesion site;

[0027] A solution module, used for solving the partial differential equation model based on the finite element method and the backward Euler method to obtain the lesion concentration of each simulated lesion site;

[0028] The optimization module is used to combine the original lesion sites and the simulated lesion sites, optimize the fitting of the partial differential equation model based on the parallel simulated annealing method, and obtain the patient parameters contained in the partial differential equation model to assist in the prediction of the development of the patient's lesions.

[0029] Optionally, the segmentation module includes a three-dimensional reconstruction unit;

[0030] The three-dimensional reconstruction unit is used to arrange the various CT two-dimensional images on the same horizontal line according to their respective established orders and longitudinal positions to generate an initial lung three-dimensional image; extract the bitmap pixel grayscale data of each layer of the initial lung three-dimensional image, and construct a three-dimensional finite element grid based on the bitmap pixel grayscale data of each layer of the image; interpolate the three-dimensional finite element grid based on the cubic spline interpolation method to obtain a final lung three-dimensional image.

[0031] Optionally, the segmentation module further includes a lesion segmentation unit;

[0032] The lesion segmentation unit is used to segment each original lesion site from the final three-dimensional lung image and extract the grayscale data of each original lesion site; normalize the grayscale data of each original lesion site to obtain the specific numerical value corresponding to each original lesion site.

[0033] Optionally, the partial differential equation model is:

[0034]

[0035]

[0036] Where n is the time at time t and The lesion concentration of the simulated lesion site collected at the position, D is the diffusion coefficient, is the gradient operator, f is the resultant force on the simulated lesion site represented by n, g(n,v) is the complex growth function of the simulated lesion site represented by n, and v is the parameter affecting the external factors.

[0037] Optionally, the solution module is used to solve the linear part of the partial differential equation model with half a time step as a condition to obtain the temporary lesion concentration of the current simulated lesion site; solve the nonlinear part of the partial differential equation model with one time step as a condition to obtain the temporary lesion concentration of the next simulated lesion site in combination with the temporary lesion concentration of the current simulated lesion site; solve the linear part of the partial differential equation model with half a time step as a condition to obtain the lesion concentration of the next simulated lesion site in combination with the temporary lesion concentration of the next simulated lesion site.

[0038] In an embodiment of the present invention, a three-dimensional imaging image synthesized from multiple existing CT tomographic images is used as a fitting reference point, and each solution value of a partial differential equation model describing the growth process of a lung lesion site is used as a fitting initial point. The partial differential equation model is optimized based on the fitting reference point and the fitting initial point to form the patient's true parameters, which can assist doctors in predicting the patient's condition and has strong practicality. It also helps subsequent doctors to make reasonable medication decisions for the optimal lung lesion sites at different stages. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0040] Figure 1 It is a flow chart of a method for predicting lung lesions of patients based on partial differential equations disclosed in an embodiment of the present invention;

[0041] Figure 2 It is a schematic diagram of the structural composition of a patient lung lesion prediction system based on partial differential equations disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] Figure 1 A schematic flow chart of a method for predicting lung lesions of a patient based on a partial differential equation in an embodiment of the present invention is shown, and the method includes the following steps:

[0044] S101, performing three-dimensional reconstruction on each collected CT two-dimensional image to generate a three-dimensional image, and segmenting and acquiring the grayscale of each original lesion site from the three-dimensional image;

[0045] The specific implementation process includes:

[0046] (1) arranging the CT two-dimensional images on the same horizontal line according to their respective predetermined orders and longitudinal positions to generate an initial lung three-dimensional image;

[0047] In an embodiment of the present invention, each CT two-dimensional image is generated by a spiral CT scanning device, which is composed of an X-ray tomography device, a microcomputer and an image display storage system. The X-ray tomography device performs a circular motion around the human body to collect and acquire scanning data, and the microcomputer and the image display storage system form a CT tomography image of the scanning data, and the CT tomography image has a calibrated arrangement number, wherein the arrangement number includes the acquisition time and the number of slice layers, that is, each of the CT two-dimensional images acquired here has a unique number. In the implementation process, the CT two-dimensional images are first classified according to the acquisition time, and then the CT two-dimensional images contained in each category are longitudinally arranged according to the number of slice layers, and then the initial lung three-dimensional image in space is generated, which includes the three-dimensional images of the lungs at different sampling times.

[0048] (2) extracting the grayscale data of the bitmap pixels of each layer of the initial lung three-dimensional image, and constructing a three-dimensional finite element mesh based on the grayscale data of the bitmap pixels of each layer of the image;

[0049] In an embodiment of the present invention, after extracting the grayscale data of the bitmap pixel points of each layer of the initial lung three-dimensional image, a cross-connection relationship is established between two adjacent layers of the image, that is, each pixel point on the first layer of the image is connected to two pixel points on the second layer of the image according to the principle of shorter distance, and so on until all layers of the image are interconnected, thereby constructing the three-dimensional finite element mesh.

[0050] (3) interpolating the three-dimensional finite element grid based on the cubic spline interpolation method to obtain a final three-dimensional lung image;

[0051] In the embodiment of the present invention, interpolation processing of the three-dimensional finite element mesh is actually equivalent to eliminating noise and distortion of the three-dimensional finite element mesh, that is, fitting the boundary defects of the three-dimensional finite element mesh to form the final three-dimensional lung image with higher pixel density.

[0052] (4) segmenting each original lesion site from the final lung three-dimensional image, and extracting grayscale data of each original lesion site;

[0053] In the embodiment of the present invention, according to step (1), it can be known that the final lung three-dimensional image also includes various final lung three-dimensional images at different sampling times. Here, the final lung three-dimensional image is segmented using existing image segmentation tools to obtain the original lesion sites corresponding to the various final lung three-dimensional images, and the grayscale values ​​of the various original lesion sites are extracted.

[0054] (5) Normalizing the grayscale data of each original lesion site to obtain specific values ​​corresponding to each original lesion site.

[0055] In the embodiment of the present invention, one of the original lesion sites is taken as an example for explanation: first, the multiple grayscale data contained in the original lesion site are normalized respectively to obtain the spatial position of the original lesion site. At the same time, combined with the sampling time t of the original lesion site, the specific value corresponding to the original lesion site is obtained as follows:

[0056] S102, combining the Euler equation and the cell action mechanism, establishing a partial differential equation model for describing the lesion concentration of each simulated lesion site;

[0057] Specifically, according to the equation describing the growth of biological tissue and the law of conservation of mass, the partial differential equation model is established as follows:

[0058]

[0059]

[0060] Where n is the time at time t and The lesion concentration (such as the concentration of cancer cells) of the simulated lesion site collected at the position, D is the diffusion coefficient, is the gradient operator, f is the resultant force on the simulated lesion site represented by n, g(n,v) is the complex growth function of the simulated lesion site represented by n, and v is the parameter affecting the external factors.

[0061] In the embodiment of the present invention, the Euler equation is an equation describing the deformation of a continuous medium in classical continuous medium mechanics. If processes with increasing quantities such as mitosis are not considered, the partial differential equation model can be directly changed to:

[0062]

[0063] For lung blood vessels and lung tissue structures in a mitotic state, the force condition f of cells in the tissue should be analyzed based on Newton's third law. In addition, the reaction term g(n,v) of the partial differential equation model is a complex term affected by many factors, including the mitotic process of cell growth itself, pathological terms and hyperplasia terms caused by bacterial or viral invasion, etc. The modeling idea of ​​the compartment model is adopted here to classify the possible factors affecting the reaction term g(n,v), and obtain the specific value of the reaction term g(n,v) by obtaining the logical relationship between the possible factors through consulting relevant information.

[0064] S103, solving the partial differential equation based on the finite element method and the backward Euler method to obtain the lesion concentration of each simulated lesion site;

[0065] In the embodiment of the present invention, the purpose of using the finite element method is to use a discrete approximation strategy to completely calculate how the lesion site grows according to the partial differential equation model for a deformable body with any complex geometric shape, such as a lung lesion site. Specifically, according to the finite element formula and the solution idea of ​​the backward Euler method, the implementation process includes the following:

[0066] (1) With half a time step dt / 2 as the condition, the linear part of the partial differential equation model is Solve for:

[0067]

[0068] The classical finite element method is used here to simulate the lesion concentration at the lesion site. Obtain the temporary lesion concentration of the current simulated lesion site

[0069] (2) Taking a time step dt as a condition, combined with the temporary lesion concentration of the current simulated lesion site For the nonlinear part of the partial differential equation model Solve for:

[0070]

[0071] The backward Euler method is used here to obtain the temporary lesion concentration of the next simulated lesion site. Of course, the more accurate Runge-Kutta method can also be used to solve the problem here;

[0072] (3) Taking half the time step dt / 2 as the condition, combined with the temporary lesion concentration of the next simulated lesion site For the linear part of the partial differential equation model Solve and obtain the lesion concentration of the next simulated lesion site And its solution formula is consistent with the formula proposed in the above step (1).

[0073] It should be noted that, through the cyclic calculation of the above steps (1) to (3), the lesion concentration of each simulated lesion site at the grid point of the three-dimensional finite element grid and the corresponding ndt time step can be obtained.

[0074] S104. Combining the original lesion sites and the simulated lesion sites, optimizing and fitting the partial differential equation model based on the parallel simulated annealing method, and obtaining the patient parameters included in the partial differential equation model to assist in predicting the development of the patient's lesions.

[0075] In the embodiment of the present invention, the data fitting problem of each original lesion site and each simulated lesion site is converted into an infinite-dimensional optimization problem:

[0076]

[0077] in, is the distance function between the simulated lesion site and the original lesion site in space corresponding to the same sampling time, which can be solved according to the least square distance or Bayesian formula, n0 is the initial lesion concentration of the lesion site, and θ is the set of other parameters in the equation. In the implementation process, the simulated lesion site and the original lesion site at the same sampling time are input into the partial differential equation model for solution. By solving the obtained θ value and n0 value, the development of the current simulated lesion site at any time in the future can be predicted, which can be used as the basis for the doctor to treat the patient's lesion.

[0078] In addition, for drugs 1, 2, 3, etc. developed by researchers for the treatment of lung tumors, the lesion concentration of the simulated lung lesion site at a certain moment in the future can be obtained according to step S104, and treatment experiments can be carried out using drugs 1, 2 and 3 respectively, with the same dose as the test condition, and the elimination effects of drugs 1, 2 and 3 on the simulated lung lesion site can be observed respectively through the existing drug diffusion equation, so as to select the best therapeutic drug for the simulated lung lesion site.

[0079] Figure 2 A schematic diagram of the structure of a patient lung lesion prediction system based on partial differential equations in an embodiment of the present invention is shown, and the system includes:

[0080] The segmentation module 201 is used to perform three-dimensional reconstruction on each collected CT two-dimensional image to generate a three-dimensional image, and to segment and obtain the grayscale of each original lesion site from the three-dimensional image;

[0081] Specifically, the segmentation module 201 includes a three-dimensional reconstruction unit and a lesion segmentation unit; wherein the three-dimensional reconstruction unit is used to arrange the various CT two-dimensional images on the same horizontal line according to their respective established orders and longitudinal positions to generate an initial lung three-dimensional image; extract the bitmap pixel grayscale data of each layer of the initial lung three-dimensional image, and construct a three-dimensional finite element grid based on the bitmap pixel grayscale data of each layer of the image; interpolate the three-dimensional finite element grid based on the cubic spline interpolation method to obtain a final lung three-dimensional image; and the lesion segmentation unit is used to segment each original lesion site from the final lung three-dimensional image, and extract the grayscale data of each original lesion site; normalize the grayscale data of each original lesion site to obtain the specific numerical value corresponding to each original lesion site.

[0082] The module 202 is established to combine the Euler equation and the cell action mechanism to establish a partial differential equation model for describing the lesion concentration of each simulated lesion site:

[0083]

[0084]

[0085] Where n is the time at time t and The lesion concentration (such as the concentration of cancer cells) of the simulated lesion site collected at the position, D is the diffusion coefficient, is the gradient, f is the resultant force on the simulated lesion site represented by n, g(n,v) is the complex growth function of the simulated lesion site represented by n, and v is the parameter affecting the external factors.

[0086] A solution module 203 is used to solve the partial differential equation model based on a finite element method and a backward Euler method to obtain the lesion concentration of each simulated lesion site;

[0087] Specifically, the solution module 203 is used to solve the linear part in the partial differential equation model with half a time step as a condition, and obtain the temporary lesion concentration of the current simulated lesion site; solve the nonlinear part in the partial differential equation model with one time step as a condition, combined with the temporary lesion concentration of the current simulated lesion site, and obtain the temporary lesion concentration of the next simulated lesion site; solve the linear part in the partial differential equation model with half a time step as a condition, combined with the temporary lesion concentration of the next simulated lesion site, and obtain the lesion concentration of the next simulated lesion site.

[0088] The optimization module 204 is used to combine the various original lesion sites and the various simulated lesion sites, optimize the fitting of the partial differential equation model based on the parallel simulated annealing method, and obtain the patient parameters contained in the partial differential equation model to assist in the prediction of the patient's lesion development.

[0089] The system is configured to execute the above-mentioned method for predicting lung lesions in patients based on partial differential equations. For the specific implementation of each module in the system, please refer to Figure 1 The method flow chart and specific implementation content shown will not be repeated here.

[0090] In an embodiment of the present invention, a three-dimensional imaging image synthesized from multiple existing CT tomographic images is used as a fitting reference point, and each solution value of a partial differential equation model describing the growth process of a lung lesion site is used as a fitting initial point. The partial differential equation model is optimized based on the fitting reference point and the fitting initial point to form the patient's true parameters, which can assist doctors in predicting the patient's condition and has strong practicality. It also helps subsequent doctors to make reasonable medication decisions for the optimal lung lesion sites at different stages.

[0091] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, and the storage medium may include: a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0092] The above is a detailed introduction to a method and system for predicting lung lesions in patients based on partial differential equations provided in an embodiment of the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for a person skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for predicting lung lesions in patients based on partial differential equations, characterized in that: The method comprises: Performing three-dimensional reconstruction on each collected CT two-dimensional image to generate a three-dimensional image, and segmenting and acquiring the grayscale of each original lesion site from the three-dimensional image; Combining the Euler equation and the cell action mechanism, a partial differential equation model was established to describe the lesion concentration at each simulated lesion site; Solving the partial differential equation model based on the finite element method and the backward Euler method to obtain the lesion concentration of each simulated lesion site; Combining the original lesion sites and the simulated lesion sites, optimizing and fitting the partial differential equation model based on the parallel simulated annealing method, and obtaining the patient parameters contained in the partial differential equation model to assist in predicting the development of the patient's lesions; The partial differential equation model is: Where n is the time at time t and The lesion concentration of the simulated lesion site collected at the position, D is the diffusion coefficient, is the gradient operator, f is the resultant force on the simulated lesion site represented by n, g(n,v) is the complex growth function of the simulated lesion site represented by n, and v is the parameter affecting the external factors; The method of solving the partial differential equation model based on the finite element method and the backward Euler method to obtain the lesion concentration of each simulated lesion site includes: Solving the linear part of the partial differential equation model under the condition of half a time step to obtain a temporary lesion concentration of the current simulated lesion site; Taking a time step as a condition, solving the nonlinear part of the partial differential equation model in combination with the temporary lesion concentration of the current simulated lesion site, and obtaining the temporary lesion concentration of the next simulated lesion site; The linear part of the partial differential equation model is solved with half a time step as a condition and combined with the temporary lesion concentration of the next simulated lesion site to obtain the lesion concentration of the next simulated lesion site.

2. The method for predicting lung lesions of a patient according to claim 1, characterized in that: The three-dimensional reconstruction of each collected CT two-dimensional image to generate a three-dimensional image comprises: Arrange the CT two-dimensional images on the same horizontal line according to their respective predetermined orders and longitudinal positions to generate an initial lung three-dimensional image; Extracting the grayscale data of the bitmap pixels of each layer of the initial lung three-dimensional image, and constructing a three-dimensional finite element mesh based on the grayscale data of the bitmap pixels of each layer of the image; The three-dimensional finite element grid is interpolated based on the cubic spline interpolation method to obtain a final three-dimensional lung image.

3. The method for predicting lung lesions of a patient according to claim 2, characterized in that: The step of segmenting and acquiring the grayscale of each original lesion site from the three-dimensional image comprises: Segmenting each original lesion site from the final three-dimensional lung image, and extracting grayscale data of each original lesion site; The grayscale data of each original lesion site is normalized to obtain the specific value corresponding to each original lesion site.

4. A patient lung lesion prediction system based on partial differential equations, characterized in that: The system comprises: A segmentation module is used to perform three-dimensional reconstruction on each collected CT two-dimensional image to generate a three-dimensional image, and to segment and obtain the grayscale of each original lesion site from the three-dimensional image; Establishing a module for combining the Euler equation and the cell action mechanism to establish a partial differential equation model for describing the lesion concentration at each simulated lesion site; A solution module, used for solving the partial differential equation model based on the finite element method and the backward Euler method to obtain the lesion concentration of each simulated lesion site; An optimization module, for combining the original lesion sites and the simulated lesion sites, optimizing and fitting the partial differential equation model based on a parallel simulated annealing method, and obtaining patient parameters included in the partial differential equation model to assist in predicting the development of the patient's lesions; The partial differential equation model is: Where n is the time at time t and The lesion concentration of the simulated lesion site collected at the position, D is the diffusion coefficient, is the gradient operator, f is the resultant force on the simulated lesion site represented by n, g(n,v) is the complex growth function of the simulated lesion site represented by n, and v is the parameter affecting the external factors; The solution module is used to solve the linear part in the partial differential equation model with half a time step as a condition to obtain the temporary lesion concentration of the current simulated lesion site; solve the nonlinear part in the partial differential equation model with one time step as a condition to obtain the temporary lesion concentration of the next simulated lesion site in combination with the temporary lesion concentration of the current simulated lesion site; solve the linear part in the partial differential equation model with half a time step as a condition to obtain the lesion concentration of the next simulated lesion site in combination with the temporary lesion concentration of the next simulated lesion site.

5. The patient lung lesion prediction system according to claim 4, characterized in that: The segmentation module includes a three-dimensional reconstruction unit; The three-dimensional reconstruction unit is used to arrange the various CT two-dimensional images on the same horizontal line according to their respective established orders and longitudinal positions to generate an initial lung three-dimensional image; extract the bitmap pixel grayscale data of each layer of the initial lung three-dimensional image, and construct a three-dimensional finite element grid based on the bitmap pixel grayscale data of each layer of the image; interpolate the three-dimensional finite element grid based on the cubic spline interpolation method to obtain a final lung three-dimensional image.

6. The patient lung lesion prediction system according to claim 4, characterized in that: The segmentation module also includes a lesion segmentation unit; The lesion segmentation unit is used to segment each original lesion site from the final lung three-dimensional image and extract the grayscale data of each original lesion site; The grayscale data of each original lesion site is normalized to obtain the specific value corresponding to each original lesion site.

Citation Information

Patent Citations

  • Method and system for denoising image

    CN101739669A

  • Traditional Chinese medicinal composition for treating turbid pathogen blood-stasis type rhinitis as well as preparation method and application thereof

    CN108159136A