Cardiovascular interventional operation teaching model construction method and system

Abstract

The invention belongs to the technical field of operation teaching, and discloses a cardiovascular interventional operation teaching model construction method and system. The method comprises the following steps that a real-time user portrait of a doctor is generated, and real-time operation action parameters under different operation schemes are obtained; constructing a real-time cardiovascular three-dimensional simulation model, and extracting real-time cardiovascular structure features; based on the real-time cardiovascular three-dimensional simulation model, a CFD simulation tool is used for simulation; constructing an order reduction model and a data analysis model by using a deep learning algorithm; and setting a teaching interaction interface, combining a real-time cardiovascular three-dimensional simulation model, a reduced-order model and a data analysis model, and constructing a cardiovascular interventional operation teaching model by using a digital twinborn technology. According to the invention, the problems of high cost investment, low teaching efficiency, poor learning effect, low experience degree and lack of customization in the prior art are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of surgical teaching, and in particular relates to a method and system for constructing a cardiovascular interventional surgery teaching model. Background Art

[0002] Cardiovascular interventional surgery is a non-open surgery used to treat heart disease and related vascular diseases. These surgeries are usually performed through vascular access, avoiding the need for traditional open-chest surgery and reducing patient trauma and recovery time. Therefore, cardiovascular interventional surgery has become widely popular in clinical medicine. Cardiovascular interventional surgery is a high-precision, high-risk medical operation that requires extremely high technical skills from doctors. With the increasing demand for cardiovascular interventional surgery, how to accurately and efficiently teach cardiovascular interventional surgery has become an important development direction in the field of clinical medical education.

[0003] The existing teaching of cardiovascular interventional surgery has the following defects: 1) Existing teaching methods for cardiovascular interventional surgery rely on practical experience. However, in clinical practice, there are certain risks in performing surgery directly on patients. It often requires hands-on teaching from mentors and chief doctors, and requires long-term on-site learning of cardiovascular interventional surgery. The cost investment is high, the teaching efficiency is low, and it cannot meet the urgent market demand. 2) The existing cardiovascular interventional surgery learning is still based on the traditional method of audio-visual teaching materials. Students lack practical operation experience, the learning effect is poor, the experience is low, and there is a lack of customization. It cannot meet the professional preferences and surgical habits of different doctors, resulting in the learning results being out of touch with actual operations. Summary of the invention

[0004] In order to solve the problems of high cost investment, low teaching efficiency, poor learning effect, low experience and lack of customization in the prior art, the purpose of the present invention is to provide a method and system for constructing a cardiovascular interventional surgery teaching model.

[0005] The technical solution adopted by the present invention is: A method for constructing a cardiovascular interventional surgery teaching model comprises the following steps: Generate a user profile based on the doctor's real-time basic information data, and generate surgical action parameters based on the real-time user profile to obtain real-time surgical action parameters under different surgical plans; According to the patient's real-time cardiovascular medical imaging data, a real-time cardiovascular three-dimensional simulation model is constructed using a three-dimensional reconstruction simulation algorithm, and the real-time cardiovascular structural features of the real-time cardiovascular three-dimensional simulation model are extracted; Based on the real-time cardiovascular three-dimensional simulation model, the CFD simulation tool is used to simulate and obtain several real-time cardiovascular dynamic data of the real-time cardiovascular three-dimensional simulation model under different real-time surgical action parameters; Based on the patient's real-time cardiovascular structural characteristics and several real-time cardiovascular dynamic data under different real-time surgical action parameters, a deep learning algorithm is used to build a reduced-order model and a data analysis model; A teaching interactive interface is set up, and a cardiovascular interventional surgery teaching model is constructed using digital twin technology in combination with real-time cardiovascular three-dimensional simulation models, reduced-order models, and data analysis models.

[0006] Furthermore, a user portrait is generated based on the real-time basic information data of the doctor, and surgical action parameters are generated based on the obtained real-time user portrait to obtain real-time surgical action parameters under different surgical plans, including the following steps: Use deep learning algorithms to build a user portrait generation model, and use enhanced reinforcement learning algorithms to build a surgical action parameter generation model; Based on the real-time basic information data of the doctor, the user portrait generation model is used to generate the user portrait and obtain the real-time user portrait of the doctor; According to the real-time user portrait, the surgical action parameter generation model is used to generate surgical action parameters to obtain real-time surgical action parameters under different surgical plans.

[0007] Furthermore, the user portrait generation model is built based on the LSTM algorithm; The surgical action parameter generation model is constructed based on the MPO-DQN algorithm.

[0008] Furthermore, according to the patient's real-time cardiovascular medical imaging data, a real-time cardiovascular three-dimensional simulation model is constructed using a three-dimensional reconstruction simulation algorithm, and real-time cardiovascular structural features of the real-time cardiovascular three-dimensional simulation model are extracted, including the following steps: Use deep learning algorithms to build image segmentation models and cardiovascular structure feature extraction models, and use deep learning-based 3D reconstruction simulation algorithms to build 3D reconstruction models; According to the patient's real-time cardiovascular medical imaging data, an image segmentation model is used to perform image segmentation to obtain the patient's real-time cardiovascular structure image; Based on real-time cardiovascular medical imaging data, an initial real-time cardiovascular three-dimensional simulation model is constructed using a three-dimensional reconstruction model; According to the real-time cardiovascular structure image, the cardiovascular structure in the initial real-time cardiovascular three-dimensional simulation model is finely adjusted to obtain the final real-time cardiovascular three-dimensional simulation model; The cardiovascular structure feature extraction model is used to extract the real-time cardiovascular structure features of the cardiovascular structure in the final real-time cardiovascular three-dimensional simulation model.

[0009] Furthermore, the image segmentation model is built based on the ResNet-FCN algorithm; The cardiovascular structure feature extraction model is built based on the DBN algorithm; The 3D reconstruction model is built based on the GANs algorithm.

[0010] Furthermore, based on the real-time cardiovascular three-dimensional simulation model, a CFD simulation tool is used to perform simulation to obtain a number of real-time cardiovascular dynamic data of the real-time cardiovascular three-dimensional simulation model under different real-time surgical action parameters, including the following steps: Import the real-time cardiovascular 3D simulation model into the CFD simulation tool, and use the real-time cardiovascular structural characteristics and real-time surgical action parameters as the adjustment parameters of the CFD simulation tool; Adjust the real-time surgical action parameters and use CFD simulation tools to simulate, and obtain a number of real-time cardiovascular dynamic data of the real-time cardiovascular three-dimensional simulation model under different real-time surgical action parameters; The real-time cardiovascular structural features are adjusted, and the previous step is repeated to obtain a number of real-time cardiovascular dynamic data with different real-time cardiovascular structural features under different real-time surgical action parameters.

[0011] Furthermore, based on the patient's real-time cardiovascular structural characteristics and a number of real-time cardiovascular dynamic data under different real-time surgical action parameters, a deep learning algorithm is used to construct a reduced-order model and a data analysis model, including the following steps: Using real-time cardiovascular structural features and real-time surgical action parameters as the first input, using real-time cardiovascular dynamic data as the true label, and using predicted cardiovascular dynamic data as the first output; According to different real-time cardiovascular structural characteristics and several real-time cardiovascular dynamic data under different real-time surgical action parameters, the CNN-Attention-MLP algorithm was used to construct a reduced-order model; Using real-time cardiovascular structural features and predicted cardiovascular dynamic data as a first input, and using real-time data analysis predicted labels as a second output; The CNN-Attention-MLP algorithm was used to construct a data analysis model using different real-time cardiovascular structural features and several predicted cardiovascular dynamic data under different real-time surgical action parameters.

[0012] Furthermore, a teaching interactive interface is set up, and a cardiovascular interventional surgery teaching model is constructed using digital twin technology by combining a real-time cardiovascular three-dimensional simulation model, a reduced-order model, and a data analysis model, including the following steps: Setting up a teaching interaction interface, and setting up a basic information data input interface, a surgical plan adjustment interface, a cardiovascular medical imaging data input interface, a cardiovascular dynamic data display interface, and a data analysis result display interface in the teaching interaction interface; Connect the basic information data input interface to the user portrait generation model, connect the surgical plan adjustment interface to the surgical action parameter generation model, connect the cardiovascular medical image data input interface to the image segmentation model, connect the data analysis result display interface to the data analysis model, and connect the cardiovascular dynamic data display interface to the reduced-order model; The image segmentation model, the cardiovascular structure feature extraction model, the three-dimensional reconstruction model and the real-time cardiovascular three-dimensional simulation model are sequentially connected, the predicted cardiovascular dynamic data output by the reduced-order model is converted into a data stream, and the data stream is connected to the real-time cardiovascular three-dimensional simulation model; The teaching interactive interface, user portrait generation model, surgical action parameter generation model, image segmentation model, cardiovascular structure feature extraction model, three-dimensional reconstruction model, real-time cardiovascular three-dimensional simulation model, reduced-order model and data analysis model are integrated to obtain a cardiovascular interventional surgery teaching model.

[0013] Furthermore, the method for using the cardiovascular interventional surgery teaching model includes the following steps: Based on the cardiovascular interventional surgery teaching model, the target surgical plan is input into the teaching interactive interface, and the target predicted cardiovascular dynamic data is output using the reduced-order model according to the obtained target surgical action parameters; According to target predicted cardiovascular dynamic data, the real-time cardiovascular three-dimensional simulation model is adjusted to obtain an adjusted real-time cardiovascular three-dimensional simulation model; Using a cardiovascular structure feature extraction model, extracting target real-time cardiovascular structure features of the adjusted real-time cardiovascular three-dimensional simulation model; According to the target real-time cardiovascular structural characteristics and the target predicted cardiovascular dynamic data, the data analysis model is used to perform data analysis to obtain the target real-time data analysis results; Use the teaching interactive interface to visualize the target predicted cardiovascular dynamic data and target real-time data analysis results.

[0014] A cardiovascular interventional surgery teaching model construction system is used to implement a cardiovascular interventional surgery teaching model construction method. The system includes a surgical action parameter generation unit, a three-dimensional reconstruction simulation unit, a CFD simulation unit, a reduction and data analysis model construction unit, and a digital twin model construction unit that are connected in sequence.

[0015] The beneficial effects of the present invention are: The present invention discloses a method and system for constructing a cardiovascular interventional surgery teaching model, constructs a cardiovascular interventional surgery teaching model based on digital twin technology, generates highly personalized surgical action parameters and cardiovascular three-dimensional simulation models by combining the real-time basic information data of doctors and the real-time cardiovascular medical imaging data of patients, meets the professional preferences, surgical habits, etc. of different doctors, greatly improves the authenticity and pertinence of simulation training, reduces the cost of teaching, and improves teaching efficiency; introduces interactive teaching and instant feedback mechanism, enables doctors to receive real-time dynamic data feedback during simulated surgery, thereby better training emergency response capabilities and surgical strategy formulation, and improves learning effects and experience; adds reduced-order models and data analysis models to reduce the amount of data calculation, and provides timely data analysis functions, greatly enhancing the intelligence of the cardiovascular interventional surgery teaching model; designs a teaching interactive interface that is easy to operate and maintain, lowers the technical threshold, and enables the teaching model to be widely used in medical institutions at different levels.

[0016] Other beneficial effects of the present invention will be further described in the specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flowchart of the method for constructing a teaching model for cardiovascular intervention surgery in the present invention.

[0018] Figure 2 It is a structural block diagram of the cardiovascular intervention surgery teaching model construction system of the present invention. DETAILED DESCRIPTION

[0019] The present invention will be further explained below in conjunction with the accompanying drawings and specific embodiments.

[0020] Embodiment 1: like Figure 1 As shown, this embodiment provides a method for constructing a cardiovascular intervention surgery teaching model, comprising the following steps: S1: Generate a user profile based on the real-time basic information data of the doctor, and generate surgical action parameters based on the obtained real-time user profile to obtain real-time surgical action parameters under different surgical plans, including the following steps: S1-1: Use deep learning algorithms to build a user portrait generation model, and use enhanced reinforcement learning algorithms to build a surgical action parameter generation model, including the following steps: S1-1-1: Collect historical basic information data of different doctors and perform preprocessing to obtain some preprocessed historical basic information data; S1-1-2: Use the LSTM algorithm to build an initial user portrait generation model; The user portrait generation model is built based on the Long Short-Term Memory (LSTM) algorithm. The LSTM network can mine the deep features of doctors' surgical characteristics, plan characteristics, etc. in the basic information data, and generate user portraits of doctors based on these features. S1-1-3: Based on some pre-processed historical basic information data, the initial user portrait generation model is trained and optimized to obtain the final user portrait generation model, and several historical user portraits are generated; S1-1-4: The surgical action parameter generation problem is used as the simulation environment of the DQN algorithm, and the network parameter generation problem of the DQN algorithm is divided into several subtasks of the MPO algorithm according to different surgical plans; S1-1-5: Use the MPO algorithm to build the initial meta-strategy optimization module, input several historical user portraits, train under several subtasks, obtain the final meta-strategy optimization module, and generate several historical meta-strategy optimization experiences; S1-1-6: Use the final meta-strategy optimization module to output the historical initial network parameters of the DQN algorithm, and use the DQN algorithm to build the initial reinforcement learning module based on the historical initial network parameters; S1-1-7: Define the state space based on historical user portraits, define the action space based on preset surgical action parameters, and set the reward function; The state space includes user profile information (such as surgical habits, plan preferences, department information, etc.); the action space includes different preset surgical action parameters (such as knife position, knife force, knife angle, etc.); the reward function is used to evaluate the degree of adaptation of the preset surgical action parameters to the user profile information; S1-1-8: Based on the state space, action space and reward function, several historical user portraits and several historical resource data features are input to train the agent and deep Q network of the initial reinforcement learning module to obtain the final reinforcement learning module and generate several historical reinforcement learning experiences; S1-1-9: storing several historical meta-strategy optimization experiences and several historical reinforcement learning experiences in the experience replay pool of the final reinforcement learning module, and integrating the final meta-strategy optimization module and the final reinforcement learning module to obtain a surgical action parameter generation model; The surgical action parameter generation model is constructed based on the Meta-Policy Optimization (MPO)-Deep Q Network (DQN) algorithm, and the surgical action parameter generation model includes a meta-policy optimization module constructed based on the MPO algorithm and a reinforcement learning module constructed based on the DQN algorithm. The reinforcement learning module is provided with an intelligent agent, a deep Q network and an experience replay pool. The intelligent agent is connected to the deep Q network, the experience replay pool and the meta-policy optimization module respectively, and the meta-policy optimization module is connected to the experience replay pool. The meta-strategy optimization module is used to optimize the deep Q network parameters in the reinforcement learning module so that these parameters can quickly adapt to new and unseen tasks, improve the generalization ability of the model, and update the deep Q network based on previous learning experience even under unseen data; through meta-strategy optimization, the DQN algorithm can converge faster, reduce the amount of data and time required for training, and significantly improve learning efficiency; the reinforcement learning module outputs personalized surgical action parameters based on the characteristics of the user portrait; S1-2: Based on the real-time basic information data of the doctor, a user portrait generation model is used to generate a user portrait to obtain a real-time user portrait of the doctor; S1-3: Based on the real-time user portrait, the surgical action parameter generation model is used to generate surgical action parameters to obtain real-time surgical action parameters under different surgical plans, including the following steps: S1-3-1: According to the real-time user portrait, a number of historical meta-strategy optimization experiences stored in the experience playback pool of the surgical action parameter generation model are matched to obtain the real-time meta-strategy optimization experience with successful matching; S1-3-2: Based on the real-time meta-strategy optimization experience, use the meta-strategy optimization module to generate real-time initialization network parameters, and adjust the reinforcement learning module according to the initialization network parameters to obtain the adjusted reinforcement learning module; S1-3-3: According to the real-time user portrait, a number of historical reinforcement learning experiences stored in the experience replay pool of the surgical action parameter generation model are matched to obtain the real-time reinforcement learning experience with successful matching; S1-3-4: updating and adjusting the action space of the reinforcement learning module according to the real-time reinforcement learning experience to obtain an updated action space; S1-3-5: Update and adjust the state space of the reinforcement learning module according to the real-time user portrait to obtain an updated state space; S1-3-6: Generate surgical action parameters using the adjusted reinforcement learning module according to the updated state space and the updated action space, and obtain real-time surgical action parameters under different surgical plans; S2: Based on the patient's real-time cardiovascular medical imaging data, a real-time cardiovascular three-dimensional simulation model is constructed using a three-dimensional reconstruction simulation algorithm, and real-time cardiovascular structural features of the real-time cardiovascular three-dimensional simulation model are extracted, including the following steps: S2-1: Use deep learning algorithms to build image segmentation models and cardiovascular structure feature extraction models, and use deep learning-based 3D reconstruction simulation algorithms to build 3D reconstruction models; The image segmentation model is built based on the residual network (Residual Network, ResNet)-full convolutional network (Fully Convolutional Networks, FCN) algorithm; the image segmentation model is based on the pre-trained ResNet model as the basic network, and the last few fully connected layers of the ResNet network are modified to convolutional layers to obtain the FCN network, which is used to separate the cardiovascular structure from the background and other tissue structures, and output the cardiovascular structure image; The cardiovascular structure feature extraction model is built based on the Deep Belief Network (DBN) algorithm. Due to the lack of labeled training data as training samples for deep learning in the existing field, the unsupervised learning DBN network is used to build the cardiovascular structure feature extraction model, which reduces the demand for labeled samples and improves the training efficiency of the cardiovascular structure feature extraction model. It is used to extract the cardiovascular structure features of the cardiovascular structure in the cardiovascular three-dimensional simulation model. The 3D reconstruction model is built based on the Generative Adversarial Networks (GANs) algorithm; it consists of two parts: the generator and the discriminator. The generator attempts to generate data that is realistic enough to deceive the discriminator, while the discriminator attempts to distinguish between real data and fake data generated by the generator. In the field of 3D reconstruction, GANs can be used to generate 3D models from 2D images or other forms of input, and to build cardiovascular 3D simulation models based on cardiovascular medical imaging data, which improves the efficiency and accuracy of building 3D simulation models. S2-2: Based on the patient’s real-time cardiovascular medical imaging data, use the image segmentation model to perform image segmentation to obtain the patient’s real-time cardiovascular structural image; cardiovascular medical imaging data includes CT scans, MRI scans, etc., and cardiovascular medical imaging data should include the patient’s cardiovascular structure and pathological characteristics; S2-3: Based on real-time cardiovascular medical imaging data, use the three-dimensional reconstruction model to build an initial real-time cardiovascular three-dimensional simulation model; S2-4: Based on the real-time cardiovascular structure image, fine-tune the cardiovascular structure in the initial real-time cardiovascular 3D simulation model, including the adjustment of the bends, branches and lesions of the blood vessels, as well as the setting of the model's material properties, boundary conditions, etc., to obtain the final real-time cardiovascular 3D simulation model; S2-5: Using the cardiovascular structure feature extraction model, extract the real-time cardiovascular structure features of the cardiovascular structure in the final real-time cardiovascular three-dimensional simulation model; the real-time cardiovascular structure features include the degree of cardiovascular lesions, blood vessel diameter, size and shape of the lesion area, thickness of the blood vessel wall, etc.; S3: Based on the real-time cardiovascular three-dimensional simulation model, a computational fluid dynamics (CFD) simulation tool is used to perform simulation to obtain a number of real-time cardiovascular dynamic data of the real-time cardiovascular three-dimensional simulation model under different real-time surgical action parameters, including the following steps: S3-1: Import the real-time cardiovascular 3D simulation model into the CFD simulation tool, and use the real-time cardiovascular structural characteristics and real-time surgical action parameters as the adjustment parameters of the CFD simulation tool; S3-2: Adjust the real-time surgical action parameters and use CFD simulation tools to simulate, and obtain a number of real-time cardiovascular dynamic data of the real-time cardiovascular 3D simulation model under different real-time surgical action parameters. The real-time cardiovascular dynamic data reflects the actual situation of blood flow, such as flow pattern, eddy current, turbulence, etc.; S3-3: adjusting the real-time cardiovascular structural features, and repeating the previous step to obtain a number of real-time cardiovascular dynamic data with different real-time cardiovascular structural features under different real-time surgical action parameters; S4: Based on the patient's real-time cardiovascular structural characteristics and several real-time cardiovascular dynamic data under different real-time surgical action parameters, a deep learning algorithm is used to construct a reduced-order model and a data analysis model, including the following steps: S4-1: using the real-time cardiovascular structural features and the real-time surgical action parameters as the first input, using the real-time cardiovascular dynamic data as the true label, and using the predicted cardiovascular dynamic data as the first output; S4-2: Based on different real-time cardiovascular structural characteristics and several real-time cardiovascular dynamic data under different real-time surgical action parameters, a reduced-order model was constructed using the Convolutional Neural Network (CNN)-Attention-Multilayer Perceptron (MLP) algorithm; The reduced-order model includes a first input layer, a first data feature extraction module based on CNN, a first attention weighting module based on the Attention mechanism, and a cardiovascular dynamic prediction module based on the MLP algorithm; The first input layer is used to receive the first input quantity, the first data feature extraction module extracts the first data feature of the real-time surgical action parameter, and according to the first preset attention weight of the first attention weighting module, the real-time cardiovascular structure feature and the first data feature are weighted and fused, the cardiovascular dynamic prediction module performs prediction according to the first fusion feature, generates corresponding predicted cardiovascular dynamic data, and obtains the accuracy of the model by statistically comparing the predicted cardiovascular dynamic data with the real label, and outputs the final reduced-order model if the accuracy is higher than the threshold; S4-3: using the real-time cardiovascular structural features and the predicted cardiovascular dynamic data as the first input, and using the real-time data analysis predicted label as the second output; S4-4: Using the CNN-Attention-MLP algorithm, we construct a data analysis model using different real-time cardiovascular structural features and several predicted cardiovascular dynamic data under different real-time surgical action parameters; The data analysis model includes a second input layer, a second data feature extraction module based on CNN, a second attention weighting module based on the Attention mechanism, and a data analysis prediction module based on the MLP algorithm; The second input layer is used to receive the second input quantity, the second data feature extraction module extracts the second data feature for predicting the surgical action parameter, and according to the second preset attention weight of the second attention weighting module, the real-time cardiovascular structure feature and the second data feature are weightedly fused, and the data analysis prediction module performs prediction according to the second fusion feature to generate a corresponding data analysis prediction label; S5: Set up a teaching interactive interface, and use digital twin technology to build a cardiovascular interventional surgery teaching model by combining a real-time cardiovascular 3D simulation model, a reduced-order model, and a data analysis model, including the following steps: S5-1: Setting up a teaching interaction interface, and setting up a basic information data input interface, a surgical plan adjustment interface, a cardiovascular medical imaging data input interface, a cardiovascular dynamic data display interface, and a data analysis result display interface in the teaching interaction interface; S5-2: connect the basic information data input interface to the user portrait generation model, connect the surgical plan adjustment interface to the surgical action parameter generation model, connect the cardiovascular medical imaging data input interface to the image segmentation model, connect the data analysis result display interface to the data analysis model, and connect the cardiovascular dynamic data display interface to the reduced-order model; S5-3: sequentially connecting the image segmentation model, the cardiovascular structure feature extraction model, the three-dimensional reconstruction model, and the real-time cardiovascular three-dimensional simulation model, converting the predicted cardiovascular dynamic data output by the reduced-order model into a data stream, and connecting the data stream to the real-time cardiovascular three-dimensional simulation model; S5-4: Integrate the teaching interaction interface, user portrait generation model, surgical action parameter generation model, image segmentation model, cardiovascular structure feature extraction model, 3D reconstruction model, real-time cardiovascular 3D simulation model, reduced-order model and data analysis model to obtain a cardiovascular interventional surgery teaching model; The method of using the cardiovascular interventional surgery teaching model includes the following steps: A-1: Based on the cardiovascular interventional surgery teaching model, the target surgical plan is input into the teaching interactive interface, and the target predicted cardiovascular dynamic data is output using the reduced-order model according to the obtained target surgical action parameters; A-2: According to the target predicted cardiovascular dynamic data, the real-time cardiovascular three-dimensional simulation model is adjusted to obtain the adjusted real-time cardiovascular three-dimensional simulation model; A-3: Using the cardiovascular structure feature extraction model, extracting the target real-time cardiovascular structure features of the adjusted real-time cardiovascular 3D simulation model; A-4: Based on the target real-time cardiovascular structural characteristics and target predicted cardiovascular dynamic data, use the data analysis model to perform data analysis and obtain the target real-time data analysis results; A-5: Use the teaching interactive interface to visualize the target predicted cardiovascular dynamic data and target real-time data analysis results.

[0021] Embodiment 2: like Figure 2 As shown, this embodiment provides a cardiovascular intervention surgery teaching model construction system, which is used to implement a cardiovascular intervention surgery teaching model construction method. The system includes a surgical action parameter generation unit, a three-dimensional reconstruction simulation unit, a CFD simulation unit, a reduction and data analysis model construction unit, and a digital twin model construction unit connected in sequence; A surgical action parameter generating unit is used to generate a user portrait based on the real-time basic information data of the doctor, and to generate surgical action parameters based on the obtained real-time user portrait, so as to obtain real-time surgical action parameters under different surgical plans; A three-dimensional reconstruction simulation unit is used to construct a real-time cardiovascular three-dimensional simulation model based on the patient's real-time cardiovascular medical imaging data using a three-dimensional reconstruction simulation algorithm, and to extract real-time cardiovascular structural features of the real-time cardiovascular three-dimensional simulation model; A CFD simulation unit is used to simulate the real-time cardiovascular three-dimensional simulation model using a CFD simulation tool to obtain a number of real-time cardiovascular dynamic data of the real-time cardiovascular three-dimensional simulation model under different real-time surgical action parameters; A reduction and data analysis model building unit, used to build a reduction model and a data analysis model using a deep learning algorithm according to the patient's real-time cardiovascular structural characteristics and a number of real-time cardiovascular dynamic data under different real-time surgical action parameters; The digital twin model construction unit is used to set up the teaching interactive interface, and combines the real-time cardiovascular three-dimensional simulation model, the reduced-order model and the data analysis model to build a cardiovascular interventional surgery teaching model using digital twin technology.

[0022] The present invention discloses a method and system for constructing a cardiovascular interventional surgery teaching model, constructs a cardiovascular interventional surgery teaching model based on digital twin technology, generates highly personalized surgical action parameters and cardiovascular three-dimensional simulation models by combining the real-time basic information data of doctors and the real-time cardiovascular medical imaging data of patients, meets the professional preferences, surgical habits, etc. of different doctors, greatly improves the authenticity and pertinence of simulation training, reduces the cost of teaching, and improves teaching efficiency; introduces interactive teaching and instant feedback mechanism, enables doctors to receive real-time dynamic data feedback during simulated surgery, thereby better training emergency response capabilities and surgical strategy formulation, and improves learning effects and experience; adds reduced-order models and data analysis models to reduce the amount of data calculation, and provides timely data analysis functions, greatly enhancing the intelligence of the cardiovascular interventional surgery teaching model; designs a teaching interactive interface that is easy to operate and maintain, lowers the technical threshold, and enables the teaching model to be widely used in medical institutions at different levels.

[0023] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for constructing a cardiovascular intervention surgery teaching model, characterized in that: The steps include: Generate a user profile based on the doctor's real-time basic information data, and generate surgical action parameters based on the real-time user profile to obtain real-time surgical action parameters under different surgical plans; According to the patient's real-time cardiovascular medical imaging data, a real-time cardiovascular three-dimensional simulation model is constructed using a three-dimensional reconstruction simulation algorithm, and the real-time cardiovascular structural features of the real-time cardiovascular three-dimensional simulation model are extracted; Based on the real-time cardiovascular three-dimensional simulation model, the CFD simulation tool is used to simulate and obtain several real-time cardiovascular dynamic data of the real-time cardiovascular three-dimensional simulation model under different real-time surgical action parameters; Based on the patient's real-time cardiovascular structural characteristics and several real-time cardiovascular dynamic data under different real-time surgical action parameters, a deep learning algorithm is used to build a reduced-order model and a data analysis model; A teaching interactive interface is set up, and a cardiovascular interventional surgery teaching model is constructed using digital twin technology in combination with real-time cardiovascular three-dimensional simulation models, reduced-order models, and data analysis models.

2. A method for constructing a teaching model for cardiovascular intervention surgery according to claim 1, characterized in that: According to the real-time basic information data of the doctor, a user portrait is generated, and according to the obtained real-time user portrait, surgical action parameters are generated to obtain real-time surgical action parameters under different surgical plans, including the following steps: Use deep learning algorithms to build a user portrait generation model, and use enhanced reinforcement learning algorithms to build a surgical action parameter generation model; Based on the real-time basic information data of the doctor, the user portrait generation model is used to generate the user portrait and obtain the real-time user portrait of the doctor; According to the real-time user portrait, the surgical action parameter generation model is used to generate surgical action parameters to obtain real-time surgical action parameters under different surgical plans.

3. A method for constructing a teaching model for cardiovascular intervention surgery according to claim 2, characterized in that: The user portrait generation model is constructed based on the LSTM algorithm; The surgical action parameter generation model is constructed based on the MPO-DQN algorithm.

4. A method for constructing a cardiovascular interventional surgery teaching model according to claim 2, characterized in that: According to the patient's real-time cardiovascular medical imaging data, a real-time cardiovascular three-dimensional simulation model is constructed using a three-dimensional reconstruction simulation algorithm, and the real-time cardiovascular structural features of the real-time cardiovascular three-dimensional simulation model are extracted, including the following steps: Use deep learning algorithms to build image segmentation models and cardiovascular structure feature extraction models, and use deep learning-based 3D reconstruction simulation algorithms to build 3D reconstruction models; According to the patient's real-time cardiovascular medical imaging data, an image segmentation model is used to perform image segmentation to obtain the patient's real-time cardiovascular structure image; Based on real-time cardiovascular medical imaging data, an initial real-time cardiovascular three-dimensional simulation model is constructed using a three-dimensional reconstruction model; According to the real-time cardiovascular structure image, the cardiovascular structure in the initial real-time cardiovascular three-dimensional simulation model is finely adjusted to obtain the final real-time cardiovascular three-dimensional simulation model; The cardiovascular structure feature extraction model is used to extract the real-time cardiovascular structure features of the cardiovascular structure in the final real-time cardiovascular three-dimensional simulation model.

5. A method for constructing a teaching model for cardiovascular intervention surgery according to claim 4, characterized in that: The image segmentation model is built based on the ResNet-FCN algorithm; The cardiovascular structure feature extraction model is constructed based on the DBN algorithm; The three-dimensional reconstruction model is constructed based on the GANs algorithm.

6. A method for constructing a teaching model for cardiovascular intervention surgery according to claim 4, characterized in that: Based on the real-time cardiovascular three-dimensional simulation model, a CFD simulation tool is used to simulate and obtain a number of real-time cardiovascular dynamic data of the real-time cardiovascular three-dimensional simulation model under different real-time surgical action parameters, including the following steps: Import the real-time cardiovascular 3D simulation model into the CFD simulation tool, and use the real-time cardiovascular structural characteristics and real-time surgical action parameters as the adjustment parameters of the CFD simulation tool; Adjust the real-time surgical action parameters and use CFD simulation tools to simulate, and obtain a number of real-time cardiovascular dynamic data of the real-time cardiovascular three-dimensional simulation model under different real-time surgical action parameters; The real-time cardiovascular structural features are adjusted, and the previous step is repeated to obtain a number of real-time cardiovascular dynamic data with different real-time cardiovascular structural features under different real-time surgical action parameters.

7. A method for constructing a cardiovascular interventional surgery teaching model according to claim 6, characterized in that: According to the patient's real-time cardiovascular structural characteristics and several real-time cardiovascular dynamic data under different real-time surgical action parameters, a deep learning algorithm is used to build a reduced-order model and a data analysis model, including the following steps: Using real-time cardiovascular structural features and real-time surgical action parameters as the first input, using real-time cardiovascular dynamic data as the true label, and using predicted cardiovascular dynamic data as the first output; According to different real-time cardiovascular structural characteristics and several real-time cardiovascular dynamic data under different real-time surgical action parameters, the CNN-Attention-MLP algorithm was used to construct a reduced-order model; Using real-time cardiovascular structural features and predicted cardiovascular dynamic data as a first input, and using real-time data analysis predicted labels as a second output; The CNN-Attention-MLP algorithm was used to construct a data analysis model using different real-time cardiovascular structural features and several predicted cardiovascular dynamic data under different real-time surgical action parameters.

8. A method for constructing a teaching model for cardiovascular intervention surgery according to claim 7, characterized in that: Setting up a teaching interactive interface, combining a real-time cardiovascular 3D simulation model, a reduced-order model, and a data analysis model, and using digital twin technology to build a cardiovascular interventional surgery teaching model includes the following steps: Setting up a teaching interaction interface, and setting up a basic information data input interface, a surgical plan adjustment interface, a cardiovascular medical imaging data input interface, a cardiovascular dynamic data display interface, and a data analysis result display interface in the teaching interaction interface; Connect the basic information data input interface to the user portrait generation model, connect the surgical plan adjustment interface to the surgical action parameter generation model, connect the cardiovascular medical image data input interface to the image segmentation model, connect the data analysis result display interface to the data analysis model, and connect the cardiovascular dynamic data display interface to the reduced-order model; The image segmentation model, the cardiovascular structure feature extraction model, the three-dimensional reconstruction model and the real-time cardiovascular three-dimensional simulation model are sequentially connected, the predicted cardiovascular dynamic data output by the reduced-order model is converted into a data stream, and the data stream is connected to the real-time cardiovascular three-dimensional simulation model; The teaching interactive interface, user portrait generation model, surgical action parameter generation model, image segmentation model, cardiovascular structure feature extraction model, three-dimensional reconstruction model, real-time cardiovascular three-dimensional simulation model, reduced-order model and data analysis model are integrated to obtain a cardiovascular interventional surgery teaching model.

9. A method for constructing a teaching model for cardiovascular intervention surgery according to claim 8, characterized in that: The method for using the cardiovascular interventional surgery teaching model comprises the following steps: Based on the cardiovascular interventional surgery teaching model, the target surgical plan is input into the teaching interactive interface, and the target predicted cardiovascular dynamic data is output using the reduced-order model according to the obtained target surgical action parameters; According to target predicted cardiovascular dynamic data, the real-time cardiovascular three-dimensional simulation model is adjusted to obtain an adjusted real-time cardiovascular three-dimensional simulation model; Using a cardiovascular structure feature extraction model, extracting target real-time cardiovascular structure features of the adjusted real-time cardiovascular three-dimensional simulation model; According to the target real-time cardiovascular structural characteristics and the target predicted cardiovascular dynamic data, the data analysis model is used to perform data analysis to obtain the target real-time data analysis results; Use the teaching interactive interface to visualize the target predicted cardiovascular dynamic data and target real-time data analysis results.

10. A cardiovascular intervention surgery teaching model construction system, used to implement the cardiovascular intervention surgery teaching model construction method according to any one of claims 1 to 9, characterized in that: The system includes a surgical action parameter generation unit, a three-dimensional reconstruction simulation unit, a CFD simulation unit, an order reduction and data analysis model construction unit, and a digital twin model construction unit, which are connected in sequence.

Images