Operation plan planning method based on AI algorithm

Abstract

The invention discloses an operation plan planning method based on an AI algorithm. The operation plan planning method comprises the step of scanning organs, portal veins, hepatic arteries, hepatic veins and tumors of a patient before an operation. The operation date of a patient is determined according to the disease condition of the patient, a doctor in a first operation group possibly performs other operations at the operation date of the patient, in order to avoid the situation, the invention further provides a second operation group, and the doctor which cannot perform the operation on the day is removed from the first operation group. Due to the fact that different patients have different surgical difficulties, the number of doctors in the surgical group needing to be predicted is different, if the number of doctors in the second surgical group cannot meet the surgical requirement after removal, the invention further provides a third surgical group, and based on the first surgical group, the second surgical group and the third surgical group, the system can well meet the surgical requirement of any difficulty, and the surgical efficiency is improved. And a team consisting of top doctors is used for operating patients, so that flexible deployment is realized, and the requirements of hospitals are met.

Description

Technical Field

[0001] The present invention relates to the technical field of surgical plan planning, and in particular to a surgical plan planning method based on an AI algorithm. Background Art

[0002] Organs perform a variety of vital functions in the human body. However, they are also prone to tumor growth, with malignant tumors being more common and benign tumors being less common. Malignant tumors can cause significant damage to the human body. According to a survey, my country accounts for nearly 50% of liver cancer cases worldwide. Currently, the preferred treatment for liver cancer is surgical resection, typified by hepatocellular carcinoma resection.

[0003] Patients need to take two-dimensional medical images before surgery. Doctors can only determine the severity of the patient's condition and analyze the appropriate surgical resection plan through repeated study of the two-dimensional medical images. This method relies more on the doctor's experience and is difficult to meet the needs of the hospital. Summary of the Invention

[0004] In order to solve the above technical problems, a surgical plan planning method based on AI algorithm is provided. This technical solution solves the problem proposed in the above background technology that patients need to take two-dimensional medical images before surgery. Doctors can only determine the severity of the patient's condition and analyze the corresponding surgical resection plan through repeated study of the two-dimensional medical images. This method is more dependent on the doctor's experience.

[0005] In order to achieve the above objects, the technical solution adopted by the present invention is:

[0006] A surgical plan planning method based on an AI algorithm, comprising:

[0007] Scanning the patient's preoperative organs, portal vein, hepatic artery, hepatic vein, and tumor to obtain the patient's two-dimensional medical imaging data, and storing the patient's preoperative two-dimensional medical imaging data in a medical database;

[0008] Performing a denoising operation on the patient's preoperative two-dimensional medical imaging data to obtain the denoised patient's preoperative two-dimensional medical imaging data;

[0009] Extract features from the denoised preoperative 2D medical imaging data of patients and construct a 3D digital twin organ model through mapping relationships;

[0010] Rendering the three-dimensional model of the digital twin organ to obtain a three-dimensional enhanced model of the digital twin organ;

[0011] Analyze the preoperative two-dimensional medical imaging data of historical patients and the corresponding surgical plans to build a surgical plan model;

[0012] Input the real-time preoperative patient 2D medical imaging data into the constructed surgical plan model and output the surgical plan;

[0013] Retrieve the information of all doctors who have performed the surgery from the database and number them. Each numbered doctor will perform surgical simulations in the 3D enhanced model of the digital twin organ in turn.

[0014] Obtain the simulated operation results of the surgical plans of each numbered doctor, score the results, and obtain the first surgical group;

[0015] The second surgical group is obtained based on the comparison of the first surgical group information with the patient's surgery date.

[0016] Preferably, performing a denoising operation on the patient's preoperative two-dimensional medical image data to obtain the denoised patient's preoperative two-dimensional medical image data specifically comprises the following steps:

[0017] Decompose high-frequency and low-frequency information of noisy images;

[0018] The low-frequency pixel points are processed into local pixel groups, and an unbiased estimation of the error is used to approximate the similarity between the local pixel block and the target pixel block, and a sample set of similar local pixel blocks is obtained;

[0019] Traverse each sample set obtained and use the principal component analysis algorithm to denoise it in turn. By calculating the covariance matrix, the orthogonal transformation matrix is ​​obtained. Combined with the eigenvalue matrix, the dimensions containing a small amount of information in the sample set are removed to obtain the reconstructed low-frequency components.

[0020] The high-frequency information is decomposed into overlapping blocks of the same size. The Euclidean distance is calculated to group similar blocks. The adaptive learning dictionary of each group is learned using singular value decomposition. The sparse coding is calculated using the split Bregman iterative algorithm combined with a convex optimization algorithm. The high-frequency components are reconstructed using the sparse coding and the adaptive learning dictionary.

[0021] The inverse wavelet transform aggregates the high-frequency components and the low-frequency components to obtain the denoised image.

[0022] Preferably, the feature extraction of the denoised preoperative two-dimensional medical imaging data of the patient and the construction of a three-dimensional digital twin organ model through a mapping relationship specifically include the following steps:

[0023] Extracting the foreground image of the patient's preoperative 2D medical image after denoising;

[0024] Perform color normalization on the foreground image, specifically including: normalizing the pixel color value to the sum of the brightness of the RGB three channels;

[0025] The normalized image is divided into multiple sub-regions using a segmentation algorithm, and the segmented sub-regions are fitted into relatively regular geometric shapes using computational geometry methods, where each sub-region corresponds to RGB color and spatial coordinate information;

[0026] Extract the color and position features corresponding to each sub-region;

[0027] Through the preset mapping relationship, the position features of each sub-region are mapped in the three-dimensional space to obtain the three-dimensional model of the digital twin organ.

[0028] Preferably, the rendering operation on the digital twin organ three-dimensional model to obtain the digital twin organ three-dimensional enhanced model specifically includes the following steps:

[0029] Obtaining color information and position information of each sub-region, wherein the color information and position information correspond one to one;

[0030] Modeling software is used to perform color rendering at each sub-region position in turn to obtain a three-dimensional enhanced model of the digital twin organ.

[0031] Preferably, the analysis of the historical patient's preoperative two-dimensional medical imaging data and the corresponding surgical plan to construct a surgical plan model specifically includes the following steps:

[0032] Constructing a surgical plan model based on a feedforward neural network, wherein the input data of the surgical plan model includes acquired two-dimensional medical imaging data information;

[0033] Based on machine learning, multiple sample organs are labeled and divided to obtain training sets, validation sets, and test sets;

[0034] The training set, validation set and test set are used to perform supervised training, validation and testing on the surgical plan model to obtain the surgical plan model whose accuracy meets the preset accuracy requirements.

[0035] Preferably, the step of inputting the real-time preoperative two-dimensional medical imaging data of the patient into the constructed surgical plan model and outputting the surgical plan specifically comprises the following steps:

[0036] Input the real-time preoperative patient two-dimensional medical imaging data into the constructed surgical plan model;

[0037] The surgical plan model traverses the organ, portal vein, artery, vein, and tumor information in the real-time 2D medical image through the entire historical database to obtain the historical 2D medical image with the highest matching degree;

[0038] The surgical plan of the historical patient corresponding to the historical two-dimensional medical image with the greatest matching degree is retrieved and the surgical plan is output.

[0039] Preferably, the process of retrieving the information of all doctors who have performed the surgery from the database and numbering the doctors, and having each numbered doctor sequentially perform surgical plan simulation operations in the three-dimensional enhanced model of the digital twin organ, specifically includes the following steps:

[0040] The database stores information about doctors who have experience with the surgery and those who have;

[0041] Doctors with experience in this procedure are tagged with specific information;

[0042] Traverse the doctor information in the entire database with the specific information tag, retrieve the doctors with the specific information tag, and number the doctors;

[0043] Doctors with different numbers take turns to simulate surgical plans in the three-dimensional enhanced model of the digital twin organ.

[0044] Preferably, obtaining the simulated operation results of the surgical plans of the numbered doctors, scoring the results, and obtaining the first surgical group specifically includes the following steps:

[0045] Obtain the simulated operation results of each numbered doctor's surgical plan and score the results;

[0046] Comparing the scoring results of each numbered doctor with a first preset threshold, which is automatically adjusted according to different surgical difficulties;

[0047] If the scoring result is greater than or equal to a first preset threshold, the information of each numbered doctor is output and a first surgical group is formed;

[0048] If the scoring result is less than the first preset threshold, it will not be output.

[0049] Preferably, the step of obtaining the second surgical group based on the comparison of the first surgical group information with the patient's surgery date specifically includes the following steps:

[0050] Obtain the patient's surgery date;

[0051] Doctors who were busy on the patient's surgery date were eliminated from the first surgery group, and the remaining doctors in the first surgery group formed the second surgery group;

[0052] comparing the number of patients in the second surgical group with a second preset threshold, the second preset threshold being automatically adjusted according to different surgical difficulties;

[0053] If the number of people in the second surgical group is greater than or equal to the second preset threshold, the surgical requirement is met;

[0054] If the number of people in the second surgical group is less than the second preset threshold, the surgical needs cannot be met and the third surgical group planning is carried out.

[0055] Preferably, the planning of the third surgical team specifically includes the following steps:

[0056] Share digital twin organ 3D enhanced models and surgical plans within the region;

[0057] Based on the patient's surgery date, doctors with experience in the area are selected to simulate the surgery and score the simulation results.

[0058] Numbering and counting the doctors whose scoring results are greater than a first preset threshold;

[0059] According to the second preset threshold, doctors with high scores are arranged into the second surgical group to obtain the third surgical group.

[0060] Compared with the existing technology, the present invention provides a surgical plan planning method based on AI algorithm, which has the following beneficial effects:

[0061] 1. The present invention utilizes digital twin technology to denoise the patient's preoperative two-dimensional medical imaging data based on the patient's preoperative two-dimensional medical imaging data. On the one hand, it improves the accuracy of the data, and on the other hand, it reduces the size of the image data, speeds up the subsequent modeling. The data is then feature extracted, and a digital twin organ three-dimensional model is established through mapping relationships. Rendering operations are performed to make the organ three-dimensional model more realistic.

[0062] 2. The present invention uses a neural network model to compare historical two-dimensional medical imaging data in the database with real-time two-dimensional medical imaging data, matches a highly compatible historical patient, and retrieves the surgical plan for the historical patient; doctors in the hospital with experience in this operation perform operations on the digital twin organ three-dimensional enhanced model according to the surgical plan and perform scoring. In real life, organ tumor resection requires the cooperation of a team of doctors and surgeons, so the present invention proposes a first surgical team.

[0063] 3. After the patient is examined, the surgery date is determined according to the patient's condition. Therefore, it is possible that the doctors in the first surgery group are performing other surgeries on the patient's surgery date. In order to avoid this situation, the present invention proposes a second surgery group, which eliminates the doctors who cannot perform surgery on that day from the first surgery group. Since the difficulty of surgery varies for different patients, the number of doctors in the surgery group required is also different. If the number of doctors in the second surgery group cannot meet the surgery requirements after elimination, the present invention also proposes a third surgery group. Based on the first surgery group, the second surgery group and the third surgery group, the system can well meet the surgery requirements of any difficulty, and the team is composed of top doctors to perform surgery on patients, which can be flexibly deployed, meets the needs of the hospital, and is conducive to promotion and use. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 A flowchart of a method for the present application;

[0065] Figure 2 A flowchart of a method for denoising preoperative two-dimensional medical image data of a patient in the present application;

[0066] Figure 3 A flowchart of a method for constructing a digital twin organ three-dimensional model in the present application;

[0067] Figure 4 A flowchart of a method for constructing a surgical plan model in the present application;

[0068] Figure 5 A flowchart of a method for outputting a surgical plan in the present application;

[0069] Figure 6 A flowchart of a method for simulating a surgical plan in a digital twin organ three-dimensional enhanced model in the present application;

[0070] Figure 7 A flowchart of a method for obtaining a first surgical team in the present application;

[0071] Figure 8 A flowchart of a method for obtaining a second surgical team in the present application;

[0072] Figure 9 A flowchart of a method for planning a third surgical team in the present application. DETAILED DESCRIPTION

[0073] The following description is used to disclose the present application so that those skilled in the art can implement the present application. The preferred embodiments in the following description are only as examples, and other obvious modifications can be made by those skilled in the art.

[0074] Please refer to Figures 1-9 The surgical plan planning method based on AI algorithm includes:

[0075] S100, scanning the organ, portal vein, hepatic artery, hepatic vein and tumor of the patient before operation, obtaining the two-dimensional medical image data of the patient, and storing the two-dimensional medical image data of the patient before operation in the medical database;

[0076] S200, denoising the preoperative two-dimensional medical image data of the patient, obtaining the denoised preoperative two-dimensional medical image data of the patient;

[0077] S300, extracting features from the denoised preoperative two-dimensional medical image data of the patient, constructing a digital twin organ three-dimensional model through a mapping relationship;

[0078] S400: Rendering the three-dimensional model of the digital twin organ to obtain a three-dimensional enhanced model of the digital twin organ;

[0079] S500: Analyze the preoperative two-dimensional medical imaging data of historical patients and the corresponding surgical plans to construct a surgical plan model;

[0080] S600: Inputting the real-time preoperative two-dimensional medical imaging data of the patient into the constructed surgical plan model and outputting the surgical plan;

[0081] S700: Retrieve the information of all doctors who have performed the surgery in the database and number the doctors. Each numbered doctor will perform a surgical simulation in the 3D enhanced model of the digital twin organ in turn.

[0082] S800: Obtain the simulated operation results of the surgical plans of the numbered doctors, score the results, and obtain the first surgical group;

[0083] S900 : Obtain a second surgical group based on the comparison of the first surgical group information and the surgical date of the patient.

[0084] It can be understood by those skilled in the art that organ tumor resection is a highly specialized type of surgery. The present invention utilizes digital twin technology to denoise the patient's preoperative two-dimensional medical imaging data based on the patient's preoperative two-dimensional medical imaging data. On the one hand, this improves the accuracy of the data, and on the other hand, it reduces the size of the image data, speeds up the subsequent modeling, and then extracts features from the data. Through mapping relationships, a digital twin organ three-dimensional model is established, and rendering operations are performed to make the organ three-dimensional model more realistic. In addition, the present invention uses a neural network model to compare historical two-dimensional medical imaging data in the database with real-time two-dimensional medical imaging data, matches a highly compatible historical patient, and retrieves the surgical plan for the historical patient. Doctors in the hospital who have experience in the operation perform operations on the digital twin organ three-dimensional enhanced model in turn according to the surgical plan, and score it in reality. In real life, organ tumor resection requires the cooperation of a surgical team of doctors. Therefore, the present invention proposes a first surgical team. After examining the patient, the patient's surgery date is determined according to the patient's condition. Therefore, there may be doctors in the first surgical team who are performing other operations on the patient's surgery date. In order to avoid this situation, the present invention proposes a second surgical team. Doctors who are unable to perform surgery on that day are eliminated from the first surgical team. Since the surgical difficulty of different patients is different, the expected number of doctors in the surgical team is also different. If the number of doctors in the second surgical team cannot meet the surgical requirements after elimination, the present invention also proposes a third surgical team. Based on the first, second and third surgical teams, the system can well meet the surgical needs of any difficulty, and the team is composed of top doctors to perform surgery on patients, which is flexible and meets the needs of the hospital, which is conducive to promotion and use.

[0085] Performing a denoising operation on the patient's preoperative two-dimensional medical image data to obtain the denoised patient's preoperative two-dimensional medical image data specifically includes the following steps:

[0086] S201, decomposing high-frequency information and low-frequency information of a noisy image;

[0087] S202, performing local pixel grouping processing on the pixels of the low-frequency information, using an unbiased estimation of the error to approximately represent the similarity between the local pixel block and the target pixel block, and obtaining a sample set of similar local pixel blocks;

[0088] S203, traversing each acquired sample set, performing denoising in turn using the principal component analysis algorithm, calculating the covariance matrix to obtain an orthogonal transformation matrix, and combining it with the eigenvalue matrix to remove dimensions containing a small amount of information in the sample set, thereby obtaining a reconstructed low-frequency component;

[0089] S204, decompose the high-frequency information into overlapping blocks of the same size, calculate the Euclidean distance, construct similar blocks into groups, learn an adaptive learning dictionary for each group using singular value decomposition, calculate sparse coding by splitting Bregman iteration algorithm combined with convex optimization algorithm, and reconstruct the high-frequency component using sparse coding and adaptive learning dictionary;

[0090] S205, wavelet inverse transform is carried out on the high-frequency component and the low-frequency component to obtain a denoised image.

[0091] Those skilled in the art can understand that through wavelet transform, the high-frequency information and the low-frequency information of the image are processed respectively, the LPG-PCA method is used to restore the low-frequency component, in the process of restoring the high-frequency component, the natural image is sparsely represented in the group domain, and the inherent local sparsity and non-local self-similarity of the natural image are simultaneously and effectively represented in a unified manner, and the high-frequency part is extracted from the image without the influence of noise. While effectively removing noise and preserving image details, high sparsity and high recovery quality are achieved, and the visual performance, peak signal-to-noise ratio and structural similarity of the image are improved.

[0092] Feature extraction is performed on the denoised preoperative two-dimensional medical image data of the patient, and a digital twin organ three-dimensional model is constructed through a mapping relationship, specifically including the following steps:

[0093] S301, extracting the foreground image of the denoised preoperative two-dimensional medical image of the patient;

[0094] S302, performing color normalization processing on the foreground image, specifically including: normalizing the pixel color value to the sum of the RGB three-channel brightness;

[0095] S303, using a segmentation algorithm to segment the normalized image into multiple sub-regions, and using computational geometry method to fit the segmented sub-regions into regular geometric shapes, wherein each sub-region corresponds to RGB color and spatial coordinate information;

[0096] S304, extracting the color and position features corresponding to each sub-region;

[0097] S305, mapping the position features of each sub-region in the three-dimensional space through a preset mapping relationship to obtain a digital twin organ three-dimensional model.

[0098] Rendering operation is performed on the digital twin organ three-dimensional model to obtain a digital twin organ three-dimensional enhanced model, specifically including the following steps:

[0099] S401, obtaining color information and position information of each sub-region, and the color information and the position information correspond one by one;

[0100] S402. Color rendering is performed on each sub-region position in sequence by using modeling software to obtain a digital twin organ three-dimensional enhanced model.

[0101] Based on the preoperative two-dimensional medical image data of historical patients and the corresponding surgical plan, a surgical plan model is constructed, which specifically includes the following steps:

[0102] S501. Based on a feedforward neural network, a surgical plan model is constructed, and the input data of the surgical plan model includes the obtained two-dimensional medical image data information;

[0103] S502. Based on machine learning, a plurality of sample organs are labeled and divided to obtain a training set, a validation set, and a test set;

[0104] S503. The training set, the validation set, and the test set are used to supervise the training, validation, and testing of the surgical plan model to obtain a surgical plan model with an accuracy rate meeting the preset accuracy requirement.

[0105] The real-time preoperative two-dimensional medical image data of the patient is input into the constructed surgical plan model, and the output surgical plan specifically includes the following steps:

[0106] S601. The real-time preoperative two-dimensional medical image data of the patient is input into the constructed surgical plan model;

[0107] S602. The surgical plan model traverses the organ, portal vein, hepatic artery, hepatic vein, and tumor information in the real-time two-dimensional medical image through the entire historical database to obtain a historical two-dimensional medical image with the maximum matching degree;

[0108] S603. The surgical plan of the historical patient corresponding to the historical two-dimensional medical image with the maximum matching degree is called and output.

[0109] Those skilled in the art can understand that the feedforward neural network is one of artificial neural networks, which adopts a one-way multi-layer structure, wherein the 0th layer is called the input layer, the last layer is called the output layer, and the other intermediate layers are called hidden layers. The hidden layers can be one or more, and each hidden layer includes a plurality of neurons. Each neuron can receive signals from the previous layer of neurons and generate output to the next layer. There is no feedback in the entire network, and signals propagate in one direction from the input layer to the output layer. A plurality of sample organs, portal veins, hepatic arteries, hepatic veins and tumor information feature markers are used as construction data. The construction data is randomly divided into a training set, a validation set and a test set according to a certain proportion, and the divided training set, validation set and test set are labeled. Exemplarily, the proportion preferred by the present application is 8:1:1. The training set is used to train the model, i.e., to determine the learning parameters such as the weights and biases of the model. The validation set is used to verify each model after training a plurality of models from the training set, record the model accuracy, and then select the best model and its corresponding parameters. The test set is used only once, i.e., when evaluating the final model after training is completed. It does not participate in the process of learning parameters or the process of selecting hyperparameters, but only uses the model evaluation. Based on machine learning, the model is supervised trained through the training set, the obtained model is verified through the validation set, the trained model is evaluated through the test set, the hyperparameters are continuously adjusted until the accuracy meets the preset requirements, and the surgical plan model is output.

[0110] All doctors who have operated the surgery in the database are retrieved, and the doctors are numbered. Each numbered doctor performs surgical plan simulation operation in the digital twin organ three-dimensional enhanced model, which specifically includes the following steps:

[0111] S701, the database stores doctor information with and without experience of the surgery;

[0112] S702, the doctors with experience of the surgery are marked with specific information;

[0113] S703, the specific information is marked in the entire database of doctors, the doctors carrying the specific information are retrieved, and the doctors are numbered;

[0114] S704, each numbered doctor performs surgical plan simulation operation in the digital twin organ three-dimensional enhanced model.

[0115] Obtaining the surgical plan simulation operation results of each numbered doctor and scoring the results, and obtaining the first surgical group specifically includes the following steps:

[0116] S801, obtaining the surgical plan simulation operation results of each numbered doctor and scoring the results;

[0117] S802: Compare the scoring results of each numbered doctor with a first preset threshold, and automatically adjust the first preset threshold according to different surgical difficulties;

[0118] S803: If the scoring result is greater than or equal to a first preset threshold, output the information of each numbered doctor and form a first surgical group;

[0119] S804: If the scoring result is less than the first preset threshold, no output is performed.

[0120] Based on the comparison of the first surgical group information with the patient's surgery date, obtaining the second surgical group specifically includes the following steps:

[0121] S901. Obtain the patient's surgery date;

[0122] S902: Eliminate doctors who are busy on the patient's surgery date from the first surgery group, and the remaining doctors in the first surgery group form a second surgery group;

[0123] S903, comparing the number of people in the second surgical group with a second preset threshold, where the second preset threshold is automatically adjusted according to different surgical difficulties;

[0124] S904: If the number of people in the second surgical group is greater than or equal to the second preset threshold, the surgical requirement is met;

[0125] S905: If the number of people in the second surgical group is less than the second preset threshold, the surgical requirements cannot be met, and a third surgical group is planned.

[0126] The planning of the third surgical team includes the following steps:

[0127] S9051. Share digital twin organ 3D enhanced models and surgical plans within the region;

[0128] S9052. Based on the patient's surgery date, select doctors in the region who have experience with the surgery, simulate the surgery, and score the simulation results;

[0129] S9053. Number and count the doctors whose scoring results are greater than a first preset threshold;

[0130] S9054. According to the second preset threshold, the doctors with high scores are arranged into the second surgical group to obtain the third surgical group.

[0131] People in this technical field can understand that since the difficulty of surgery varies for different patients, the expected number of doctors in the surgical team is also different. The number of people in the second surgical group is compared with the second preset threshold. If it is less than the second threshold, the doctors in the hospital are under load and cannot meet the surgical needs. Therefore, the system also has a regional network, which shares the digital twin organ three-dimensional enhanced model and surgical plan within the regional network. Regional doctors perform simulated operations and select doctors with higher scores to supplement the second surgical group to form a third surgical group.

[0132] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions merely illustrate the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A surgical plan planning method based on AI algorithm, characterized in that: include: Scanning the patient's preoperative organs, portal vein, hepatic artery, hepatic vein, and tumor to obtain the patient's two-dimensional medical imaging data, and storing the patient's preoperative two-dimensional medical imaging data in a medical database; Performing a denoising operation on the patient's preoperative two-dimensional medical imaging data to obtain the denoised patient's preoperative two-dimensional medical imaging data; Extract features from the denoised preoperative 2D medical imaging data of patients and construct a 3D digital twin organ model through mapping relationships; Rendering the three-dimensional model of the digital twin organ to obtain a three-dimensional enhanced model of the digital twin organ; Analyze the preoperative two-dimensional medical imaging data of historical patients and the corresponding surgical plans to build a surgical plan model; Input the real-time preoperative patient 2D medical imaging data into the constructed surgical plan model and output the surgical plan; Retrieve the information of all doctors who have performed the surgery from the database and number them. Each numbered doctor will perform surgical simulations in the 3D enhanced model of the digital twin organ in turn. Obtain the simulated operation results of the surgical plans of each numbered doctor, score the results, and obtain the first surgical group; The second surgical group is obtained based on the comparison of the first surgical group information with the patient's surgery date.

2. The surgical plan planning method based on AI algorithm according to claim 1, characterized in that: The denoising operation on the patient's preoperative two-dimensional medical image data to obtain the denoised patient's preoperative two-dimensional medical image data specifically includes the following steps: Decompose high-frequency and low-frequency information of noisy images; The low-frequency pixel points are processed into local pixel groups, and an unbiased estimation of the error is used to approximate the similarity between the local pixel block and the target pixel block, and a sample set of similar local pixel blocks is obtained; Traverse each sample set obtained and use the principal component analysis algorithm to denoise it in turn. By calculating the covariance matrix, the orthogonal transformation matrix is ​​obtained. Combined with the eigenvalue matrix, the dimensions containing a small amount of information in the sample set are removed to obtain the reconstructed low-frequency components. The high-frequency information is decomposed into overlapping blocks of the same size. The Euclidean distance is calculated to group similar blocks. The adaptive learning dictionary of each group is learned using singular value decomposition. The sparse coding is calculated using the split Bregman iterative algorithm combined with a convex optimization algorithm. The high-frequency components are reconstructed using the sparse coding and the adaptive learning dictionary. The inverse wavelet transform aggregates the high-frequency components and the low-frequency components to obtain the denoised image.

3. The surgical plan planning method based on AI algorithm according to claim 1, characterized in that: The process of extracting features from the denoised preoperative two-dimensional medical imaging data of the patient and constructing a three-dimensional digital twin organ model through a mapping relationship specifically includes the following steps: Extracting the foreground image of the patient's preoperative 2D medical image after denoising; Perform color normalization on the foreground image, specifically including: normalizing the pixel color value to the sum of the brightness of the RGB three channels; The normalized image is divided into multiple sub-regions using a segmentation algorithm, and the segmented sub-regions are fitted into relatively regular geometric shapes using computational geometry methods, where each sub-region corresponds to RGB color and spatial coordinate information; Extract the color and position features corresponding to each sub-region; Through the preset mapping relationship, the position features of each sub-region are mapped in the three-dimensional space to obtain the three-dimensional model of the digital twin organ.

4. The surgical plan planning method based on AI algorithm according to claim 1, characterized in that: The rendering operation on the digital twin organ three-dimensional model to obtain the digital twin organ three-dimensional enhanced model specifically includes the following steps: Obtaining color information and position information of each sub-region, wherein the color information and position information correspond one to one; Modeling software is used to perform color rendering at each sub-region position in turn to obtain a three-dimensional enhanced model of the digital twin organ.

5. The surgical plan planning method based on AI algorithm according to claim 1, characterized in that: The analysis of the preoperative two-dimensional medical imaging data of historical patients and the corresponding surgical plans to construct a surgical plan model specifically includes the following steps: Constructing a surgical plan model based on a feedforward neural network, wherein the input data of the surgical plan model includes acquired two-dimensional medical imaging data information; Based on machine learning, multiple sample organs are labeled and divided to obtain training sets, validation sets, and test sets; The training set, validation set and test set are used to perform supervised training, validation and testing on the surgical plan model to obtain the surgical plan model whose accuracy meets the preset accuracy requirements.

6. The surgical plan planning method based on AI algorithm according to claim 1, characterized in that: The step of inputting the real-time preoperative two-dimensional medical imaging data of the patient into the constructed surgical plan model and outputting the surgical plan specifically includes the following steps: Input the real-time preoperative patient two-dimensional medical imaging data into the constructed surgical plan model; The surgical plan model traverses the organ information in the real-time 2D medical image through the entire historical database to obtain the historical 2D medical image with the highest matching degree; The surgical plan of the historical patient corresponding to the historical two-dimensional medical image with the greatest matching degree is retrieved and the surgical plan is output.

7. The surgical plan planning method based on AI algorithm according to claim 1, characterized in that: The process of retrieving the information of all doctors who have performed the surgery in the database and numbering the doctors, and then having each numbered doctor perform surgical simulation operations in the three-dimensional enhanced model of the digital twin organ in turn, specifically includes the following steps: The database stores information about doctors who have experience with the surgery and those who have; Doctors with experience in this procedure are tagged with specific information; Traverse the doctor information in the entire database with the specific information tag, retrieve the doctors with the specific information tag, and number the doctors; Doctors with different numbers take turns to simulate surgical plans in the three-dimensional enhanced model of the digital twin organ.

8. The surgical plan planning method based on AI algorithm according to claim 1, characterized in that: The steps of obtaining the simulated operation results of the surgical plans of the numbered doctors and scoring the results to obtain the first surgical group specifically include the following steps: Obtain the simulated operation results of each numbered doctor's surgical plan and score the results; Comparing the scoring results of each numbered doctor with a first preset threshold, which is automatically adjusted according to different surgical difficulties; If the scoring result is greater than or equal to a first preset threshold, the information of each numbered doctor is output and a first surgical group is formed; If the scoring result is less than the first preset threshold, it will not be output.

9. The surgical plan planning method based on AI algorithm according to claim 1, characterized in that: The step of obtaining the second surgical group based on the comparison of the first surgical group information with the patient's surgical date specifically includes the following steps: Obtain the patient's surgery date; Doctors who were busy on the patient's surgery date were eliminated from the first surgery group, and the remaining doctors in the first surgery group formed the second surgery group; comparing the number of patients in the second surgical group with a second preset threshold, the second preset threshold being automatically adjusted according to different surgical difficulties; If the number of people in the second surgical group is greater than or equal to the second preset threshold, the surgical requirement is met; If the number of people in the second surgical group is less than the second preset threshold, the surgical needs cannot be met and the third surgical group planning is carried out.

10. The surgical plan planning method based on AI algorithm according to claim 1, characterized in that: The third surgical team planning specifically includes the following steps: Share digital twin organ 3D enhanced models and surgical plans within the region; Based on the patient's surgery date, doctors with experience in the area are selected to simulate the surgery and score the simulation results. Numbering and counting the doctors whose scoring results are greater than a first preset threshold; According to the second preset threshold, doctors with high scores are arranged into the second surgical group to obtain the third surgical group.

Images