Method, device and equipment for optimizing intravascular stent structure, medium and program product

Abstract

One or more embodiments of the present specification provide an intravascular stent structure optimization method, apparatus, device, medium and program product. The method comprises the following steps: selecting a plurality of samples for optimizing an intravascular stent structure, wherein the samples take various design parameters of the intravascular stent structure to be optimized as features; for each sample, determining a sample label of the sample under a plurality of performance dimensions by adopting a finite element analysis method; for each performance dimension, training a performance prediction model by using the sample and the sample label under the performance dimension to obtain a plurality of performance prediction models; determining a solution set of the plurality of performance prediction models by adopting a multi-objective optimization algorithm; obtaining a target solution from the solution set, and judging whether the target solution is converged or not; and under the condition that the target solution is converged, optimizing the structure of the to-be-optimized intravascular stent by adopting the design parameter value in the target solution.

Description

Technical Field

[0001] One or more embodiments of the present specification relate to the field of medical device technology, and in particular, to a method, device, equipment, medium and program product for optimizing a vascular stent structure. Background Art

[0002] A vascular stent is a medical device used to treat vascular stenosis. It is usually a circular tubular structure with a periodic microstructure. The vascular stent can be placed in a narrowed or occluded blood vessel segment to support the blood vessel, reduce vascular elastic retraction and reshaping, and maintain smooth blood flow in the lumen. It can be widely used in the treatment of vascular diseases such as coronary arteries, intracranial arteries, carotid arteries, renal arteries and femoral arteries.

[0003] How to optimize the structure of vascular stents is of great significance for improving treatment effects, reducing complications, and improving patients' quality of life. It is an important research direction in the field of cardiovascular disease treatment. Summary of the invention

[0004] In view of this, one or more embodiments of this specification provide the following technical solutions: According to a first aspect of one or more embodiments of this specification, a method for optimizing a vascular stent structure is proposed, the method comprising: Selecting a plurality of samples for optimizing the vascular stent structure, wherein the samples are characterized by a plurality of design parameters of the vascular stent structure to be optimized; For each sample, the finite element analysis method is used to determine the standard performance index value of the sample in multiple performance dimensions as the sample label of the sample in the corresponding performance dimension; For each performance dimension, a performance prediction model is trained using the sample and the sample label under the performance dimension to obtain multiple performance prediction models, wherein each performance prediction model corresponds to one performance dimension; Determine a solution set of the plurality of performance prediction models using a multi-objective optimization algorithm, wherein the solution set includes a plurality of groups of solutions, and each group of solutions includes design parameter values ​​of the plurality of design parameters; Obtaining a target solution from the solution set, and determining whether the target solution converges; When the target solution converges, the design parameter values ​​in the target solution are used to optimize the structure of the vascular stent to be optimized.

[0005] According to a second aspect of one or more embodiments of this specification, a device for optimizing a vascular stent structure is provided, the device comprising: A sample selection module selects a plurality of samples for optimizing the vascular stent structure, wherein the samples are characterized by a plurality of design parameters of the vascular stent structure to be optimized; The label determination module uses the finite element analysis method to determine the standard performance index value of each sample in multiple performance dimensions as the sample label of the sample in the corresponding performance dimension; A model training module, for each performance dimension, uses the sample and the sample label under the performance dimension to train a performance prediction model to obtain multiple performance prediction models, wherein each performance prediction model corresponds to one performance dimension; A solution set determination module, which uses a multi-objective optimization algorithm to determine a solution set of the plurality of performance prediction models, wherein the solution set includes a plurality of groups of solutions, and each group of solutions includes design parameter values ​​of the plurality of design parameters; A convergence judgment module is used to obtain a target solution from the solution set and judge whether the target solution converges; The structure optimization module optimizes the structure of the vascular stent to be optimized by using the design parameter values ​​in the target solution when the target solution converges.

[0006] According to a third aspect of one or more embodiments of this specification, an electronic device is proposed, comprising: a processor; a memory for storing processor-executable instructions; wherein the processor implements the steps of the aforementioned method by running the executable instructions.

[0007] According to a fourth aspect of one or more embodiments of the present specification, a computer-readable storage medium is provided, on which computer instructions are stored, and when the instructions are executed by a processor, the steps of the aforementioned method are implemented.

[0008] According to a fifth aspect of one or more embodiments of the present specification, a computer program product is proposed, comprising a computer program / instruction, which implements the steps of the aforementioned method when executed by a processor.

[0009] It can be seen from the above description that when optimizing the vascular stent, the present specification can select several samples, use the finite element analysis method to determine the sample labels of these samples, and then train multiple performance prediction models based on the samples and sample labels. Then, a multi-objective optimization algorithm is used to determine the solution set of these performance prediction models, and the target solution is obtained from the solution set. When the target solution converges, the design parameter values ​​in the target solution are used to optimize the vascular stent structure.

[0010] It can be seen that the present specification can train multiple vascular stent performance prediction models through a small number of samples, and then use a multi-objective optimization algorithm to determine the solution set of these vascular stent performance prediction models, and then determine the design parameter values ​​of the vascular stent structure from the solution set. The design parameter values ​​obtained by the multi-objective optimization algorithm can make the performance of all aspects of the optimized vascular stent reach a better state. At the same time, the whole process only requires a small number of samples to be implemented, without the need for a large amount of finite element analysis or experiments, which can greatly shorten the optimization time of the vascular stent structure and improve the efficiency of the vascular stent structure optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 It is a flow chart of a method for optimizing a vascular stent structure provided by an exemplary embodiment.

[0012] Figure 2 It is a flow chart of another method for optimizing a vascular stent structure provided by an exemplary embodiment.

[0013] Figure 3 It is a structural schematic diagram of a device provided by an exemplary embodiment.

[0014] Figure 4 It is a block diagram of a vascular stent structure optimization device provided by an exemplary embodiment. DETAILED DESCRIPTION

[0015] A vascular stent is a medical device used to treat vascular stenosis. It is usually a circular tubular structure with a periodic microstructure. The vascular stent can be placed in a narrowed or occluded blood vessel segment to support the blood vessel, reduce vascular elastic retraction and reshaping, and maintain smooth blood flow in the lumen. It can be widely used in the treatment of vascular diseases such as coronary arteries, intracranial arteries, carotid arteries, renal arteries and femoral arteries.

[0016] How to optimize the structure of vascular stents is of great significance for improving treatment effects, reducing complications, and improving patients' quality of life. It is an important research direction in the field of cardiovascular disease treatment.

[0017] At present, when optimizing the structure of vascular stents, engineers usually only optimize the original vascular stent structure in a single performance, which can only improve the performance of the vascular stent in one aspect, but cause the decline of other performances of the vascular stent. For example, increasing the rib width and rib thickness can improve the radial support force of the vascular stent, but it will reduce the flexibility of the vascular stent and the later endothelialization effect.

[0018] In the related art, in order to balance the various performances of the vascular stent so that the performances of each aspect reach a relatively optimal state, it is often necessary to determine the various design parameters of the vascular stent structure through a large amount of finite element analysis (FEA) or experiments, which will cost a lot of time.

[0019] This specification provides an optimization solution for a vascular stent structure. When determining the design parameters of the vascular stent structure, it is not necessary to conduct a large amount of finite element analysis or experiments, and the various performances of the vascular stent can reach a relatively optimal state, which greatly saves time and cost.

[0020] Figure 1 It is a flow chart of a method for optimizing a vascular stent structure provided by an exemplary embodiment.

[0021] The optimization method of the vascular stent structure provided in this embodiment can be applied to electronic devices, such as PCs (Personal Computers), tablet devices, PDAs (Personal Digital Assistants), laptop computers, etc. The optimization method of the vascular stent structure can also be applied to servers, which can specifically be a physical server including an independent host or a virtual server hosted by a host cluster. This specification does not impose any special restrictions on this.

[0022] Please refer to Figure 1 The method for optimizing the vascular stent structure may include the following steps: Step 102 : selecting a plurality of samples for optimizing the vascular stent structure, wherein the samples are characterized by a plurality of design parameters of the vascular stent structure to be optimized.

[0023] In this embodiment, before selecting samples for the vascular stent, the type of the vascular stent to be optimized may be determined first, and then the value ranges of various design parameters thereof may be determined based on the type of the vascular stent to be optimized.

[0024] Among them, the types of vascular stents can be divided based on the application scenarios of the vascular stents. For example, coronary vascular stents can be used to treat coronary artery disease, peripheral vascular stents can be used to treat peripheral arteriovenous disease, and chocolate balloon vascular stents can be used to constrain balloon expansion and reduce the occurrence of vascular dissection.

[0025] The vascular stent structure may include a variety of design parameters, such as rib width, rib thickness, radius of curvature of the arc, length of a single grid, etc. The value range of each design parameter is related to the type of vascular stent, and the value ranges of the design parameters of different types of vascular stents may not be exactly the same. Taking rib width and rib thickness as an example, the rib width of the chocolate balloon vascular stent may range from 0.07-0.10mm, and the rib thickness may also range from 0.07-0.10mm.

[0026] In actual implementation, the mapping relationship between the value range of each design parameter and the type of vascular stent can be pre-stored. When optimizing the structure of the vascular stent to be optimized, the mapping relationship can be searched according to the type of the vascular stent to be optimized to determine the value range of various design parameters.

[0027] In this embodiment, after determining the value ranges of various design parameters of the vascular stent to be optimized, samples can be selected within the value ranges, and the characteristics of the samples are the design parameters of the vascular stent structure to be optimized. For example, assuming that the vascular stent structure to be optimized includes 4 design parameters, which are respectively represented as to , then the feature vector of the sample can be expressed as ( , , , ), to The values ​​of are all within the range of the corresponding design parameters.

[0028] In this embodiment, the samples can be selected within the value range of each design parameter by using sampling methods such as grid sampling, modified grid sampling, Latin Hypercube Sampling (LHS), etc. The number of selected samples can be preset, which is often much less than the number of samples required by the finite element analysis or test method used in the related technology. For example, the number of selected samples can be 20.

[0029] Step 104: for each sample, a finite element analysis method is used to determine the standard performance index value of the sample in multiple performance dimensions as the sample label of the sample in the corresponding performance dimension.

[0030] In this embodiment, the performance of the vascular stent can be evaluated from multiple dimensions, such as radial support, resilience, compliance, etc. Based on the aforementioned step 102, after the sample is selected, for each sample, a finite element analysis method can be used to determine the performance index value of the sample in multiple performance dimensions (hereinafter referred to as the standard performance index value), and then multiple standard performance index values ​​are obtained, and each standard performance index value can be determined as a sample label under the corresponding performance dimension.

[0031] Taking the three performance dimensions of radial support, resilience, and compliance as examples, in this step, for each sample, the finite element analysis method can be used to obtain the three standard performance index values ​​of this sample, namely, the standard radial support index value, the standard resilience index value, and the standard flexibility index value. Among them, the standard radial support index value is the label of the sample under the radial support performance dimension, the standard resilience index value is the label of the sample under the resilience performance dimension, and the standard flexibility index value is the label of the sample under the flexibility performance dimension.

[0032] Specifically, when the finite element analysis method is used to calculate the standard performance index value, the three-dimensional design of the vascular stent can be performed based on the parameter values ​​of each design parameter in the sample, and then the designed vascular stent is imported into a finite element analysis tool, such as ABAQUS, ANSYS or COMSOL, and then a variety of standard performance index values ​​are obtained through the finite element analysis tool.

[0033] At this point, samples and sample labels for training the performance prediction model can be obtained, wherein each sample has multiple labels, and each label corresponds to a performance of the vascular stent.

[0034] Step 106: For each performance dimension, a performance prediction model is trained using the sample and the sample label under the performance dimension to obtain multiple performance prediction models, wherein each performance prediction model corresponds to one performance dimension.

[0035] In this embodiment, the selected samples and sample labels can be used to train a performance prediction model for predicting the performance of the vascular stent. Specifically, for each performance dimension of the vascular stent, the selected samples and the labels of the samples in the performance dimension can be used to train the corresponding performance prediction model.

[0036] Taking the performance dimension of resilience as an example, the selected samples and the labels of the samples under the resilience dimension (i.e., standard resilience index values) can be used to train a performance prediction model. The performance prediction model corresponds to the resilience performance and can be used to predict the resilience of the vascular stent.

[0037] In this embodiment, the performance prediction model can be a response surface model (RSM), a Kriging model (Kriging Model), an artificial neural network model (ANN), a radial basis function model (RBF), etc., and this specification does not impose any special restrictions on this.

[0038] Step 108: using a multi-objective optimization algorithm to determine a solution set of the plurality of performance prediction models, wherein the solution set includes a plurality of groups of solutions, and each group of solutions includes design parameter values ​​of the plurality of design parameters.

[0039] Based on the aforementioned step 106, after multiple performance prediction models are obtained through training, a multi-objective optimization algorithm can be used to determine the solution set of these performance prediction models. The solution set includes multiple groups of solutions, each of which includes parameter values ​​of each design parameter and a corresponding predicted performance index value, and the predicted performance index value has multiple values, which correspond to the performance dimensions one by one, and are obtained by predicting the performance of the vascular stent based on the parameter values ​​of the design parameters in the solution.

[0040] For example, assuming that there are 4 design parameters, they are to There are three performance dimensions, and the corresponding performance prediction results are recorded as to , then each set of solutions can be written as ( , , , , , , ).

[0041] In this embodiment, the multi-objective optimization algorithm may be a non-dominated sorting genetic algorithm-II (NSGA-II), a particle swarm optimization algorithm (MOPSO), an artificial hummingbird algorithm (MOAHA) and a grey wolf optimization algorithm (MOGWO), etc., and the solution set may be a Pareto solution set (ParetoSolution Set).

[0042] Step 110, obtaining a target solution from the solution set, and determining whether the target solution converges.

[0043] Based on the aforementioned step 108, after obtaining the solution sets of the multiple performance prediction models, a target solution can be obtained from the solution sets, and the target solution can be manually determined by an engineer according to an optimization target.

[0044] In this embodiment, after obtaining the target solution, it can be determined whether the target solution converges. Specifically, it can be determined whether the target solution converges by comparing the error of the predicted performance index value in the target solution with the standard performance index value. For example, it can be determined whether the error meets a preset condition. If so, it can be determined that the target solution converges. If not, it can be determined that the target solution does not converge.

[0045] Step 112: When the target solution converges, the structure of the vascular stent to be optimized is optimized using the design parameter values ​​in the target solution.

[0046] Based on the judgment result of the aforementioned step 110, when the target solution converges, it means that the error between the predicted performance index value and the standard performance index value of the design parameter in the target solution is small and relatively accurate. The design parameter values ​​in the target solution can be used to perform structural optimization on the vascular stent to be optimized, so that all aspects of the performance of the optimized vascular stent can reach a relatively good state.

[0047] It can be seen from the above description that when optimizing the vascular stent, the present specification can select several samples, use the finite element analysis method to determine the sample labels of these samples, and then train multiple performance prediction models based on the samples and sample labels. Then, a multi-objective optimization algorithm is used to determine the solution set of these performance prediction models, and the target solution is obtained from the solution set. When the target solution converges, the design parameter values ​​in the target solution are used to optimize the vascular stent structure.

[0048] It can be seen that the present specification can train multiple vascular stent performance prediction models through a small number of samples, and then use a multi-objective optimization algorithm to determine the solution set of these vascular stent performance prediction models, and then determine the design parameter values ​​of the vascular stent structure from the solution set. The design parameter values ​​obtained by the multi-objective optimization algorithm can make the performance of all aspects of the optimized vascular stent reach a better state. At the same time, the whole process only requires a small number of samples to be implemented, without the need for a large amount of finite element analysis or experiments, which can greatly shorten the optimization time of the vascular stent structure and improve the efficiency of the vascular stent structure optimization.

[0049] Figure 2 It is a flow chart of another method for optimizing a vascular stent structure provided by an exemplary embodiment.

[0050] Please refer to Figure 2 The method for optimizing the vascular stent structure may include the following steps: Step 202: Select multiple samples for optimizing the structure of the vascular stent.

[0051] In this embodiment, the implementation of step 202 can refer to the aforementioned Figure 1 Step 102 in the illustrated embodiment will not be described in detail here.

[0052] It should be noted that, in order to improve the efficiency of vascular stent structure optimization as much as possible, a small number of samples may be selected in this step, for example, 20 samples may be selected.

[0053] Step 204: for each sample, a finite element analysis method is used to determine the sample label of the sample under multiple performance dimensions.

[0054] In this embodiment, it is still assumed that the vascular stent structure has four design parameters, which are respectively represented as to , then the feature vector of each sample can be expressed as ( , , , ). Assuming that the performance of the vascular stent is evaluated from the three dimensions of radial support, resilience, and flexibility, in this step, the finite element analysis method can be used to determine the standard performance index values ​​of each sample in these three dimensions (denoted as , , ) as the sample label. In other words, the sample label of the sample under the radial support performance dimension is ; The sample labels under the resilience performance dimension are , the sample label under the flexibility performance dimension is .

[0055] Step 206: For each performance dimension, use the sample and the sample label under the performance dimension to train a corresponding performance prediction model.

[0056] Table 1

[0057] In this embodiment, please refer to the example in Table 1. For radial support, the selected samples and the sample labels of these samples can be used. To train the performance prediction model corresponding to radial support , , the performance prediction model Used to predict the radial support of vascular stents. Similarly, for rebound resilience, the selected samples and the sample labels of these samples can be used To train the performance prediction model corresponding to resilience , ; For flexibility, selected samples and sample labels of these samples can be used To train the performance prediction model corresponding to flexibility , , the performance prediction model Used to predict the flexibility of vascular stents.

[0058] In this embodiment, when training the performance prediction model, the number of iterations of the performance prediction model can be preset. When the number of iterations is reached, it can be determined whether the performance prediction model has converged. If it has converged, it can be determined that the performance prediction model has been trained and has good accuracy. When all performance prediction models have converged, the subsequent steps can be performed to use a multi-objective optimization algorithm to determine the solution set of the multiple performance prediction models. - For example, in - If all convergence occurs, proceed to the next step.

[0059] If there is a non-convergent performance prediction model, such as the performance prediction model - middle If it does not converge, step 202 may be re-executed, that is, multiple samples used to optimize the vascular stent structure may be updated to retrain the performance prediction model. Specifically, when step 202 is re-executed, the number of samples selected may be greater than the number of samples selected last time. This is because, in order to maximize the efficiency of vascular stent structure optimization, a small number of samples will be selected when step 202 is first executed. If the performance prediction model trained with these small number of samples does not converge, it may indicate that the number of samples is insufficient. When step 202 is re-executed for sample selection, more samples may be appropriately selected for performance prediction model training to increase the probability of convergence of the performance prediction model. Assuming that 20 samples are selected when step 202 is first executed, if the performance prediction model does not converge, 30 samples may be selected when step 202 is re-executed.

[0060] In this embodiment, when re-executing step 202 to select samples, all samples may be selected again, or additional samples may be selected based on the samples selected for the first time. Assuming that 20 samples are selected when step 202 is first executed, 10 additional samples may be selected when step 202 is re-executed, so that 30 samples are obtained, and the performance prediction model is trained using these 30 samples and the corresponding sample labels.

[0061] In this embodiment, the following formula can be used To determine whether each performance prediction model converges.

[0062]

[0063] in, represents the label of the i-th sample, yes The average value of is the prediction result of the performance prediction model, that is, the predicted value of the sample output by the performance prediction model, and m is the number of samples.

[0064] Specifically, if If the value is greater than or equal to the threshold (for example, 0.9), it means that the difference between the prediction result output by the performance prediction model and the label is small, and the performance prediction model converges. If it is less than the threshold, it means that the difference between the prediction result output by the performance prediction model and the label is large, and it can be determined that the performance prediction model has not converged.

[0065] Of course, in other examples, other formulas may be used to determine whether the performance prediction model converges, and this specification does not impose any special restrictions on this.

[0066] Step 208: using a multi-objective optimization algorithm to determine a solution set of multiple performance prediction models.

[0067] In this embodiment, the performance prediction model can be - When all convergence occurs, a multi-objective optimization algorithm is used to determine the performance prediction model. - Each set of solutions in the solution set includes the parameter values ​​of each design parameter and the corresponding predicted performance index value, and the predicted performance index value is obtained by predicting the performance of the vascular stent based on the parameter values ​​of the design parameters in the solution by each performance prediction model. Each set of solutions can be recorded as ( , , ).

[0068] Step 210: obtaining a target solution from the solution set, and using a finite element analysis method to determine standard performance index values ​​corresponding to parameter values ​​of multiple design parameters in the target solution.

[0069] In this embodiment, the engineer can select a target solution from the solution set according to the goal of optimizing the vascular stent result. Assume that the selected target solution is ( , , ), the finite element analysis method can be used to determine the design parameter values ​​( ) corresponding to the standard performance index values, the standard performance index values ​​corresponding to these design parameter values ​​can be recorded as ( , , ).

[0070] Step 212: Calculate the error between the standard performance index value and the predicted performance index value of the target solution, and determine whether the error meets a preset condition.

[0071] In this embodiment, since the target solution may not be in the selected sample, the error between the predicted performance index value corresponding to the design parameter value in the target solution and the standard performance index value can be calculated, and then the accuracy of the predicted performance index value can be verified by judging whether the error meets the preset conditions.

[0072] Among them, it is relatively simple to calculate the absolute value of the difference between the predicted performance index value and the standard performance index value as the error, and then determine whether the error is less than or equal to the difference threshold. If the error is less than or equal to the difference threshold, it can be determined that the error meets the preset condition. If the error is greater than the difference threshold, it can be determined that the error does not meet the preset condition. Of course, in other examples, other formulas can also be used to calculate the error between the predicted performance index value and the standard performance index value.

[0073] Step 214, when the error satisfies a preset condition, it is determined that the target solution converges, and the design parameter values ​​in the target solution are used to optimize the structure of the vascular stent to be optimized.

[0074] Based on the judgment result of the aforementioned step 210, when the error satisfies the preset conditions, it can be determined that the target solution has converged, and each design parameter value in the target solution can be determined as the final design parameter value, and then these design parameter values ​​can be used to optimize the vascular stent structure so that all aspects of the performance of the vascular stent structure can reach a better state.

[0075] Step 216, when the error does not satisfy the preset condition, it is determined that the target solution does not converge, and step 202 is re-executed to select multiple samples for optimizing the vascular stent structure.

[0076] Based on the judgment result of the aforementioned step 210, if the error does not meet the preset condition, it can be explained that the predicted performance index value in the target solution is significantly different from the standard performance index value, and it can be determined that the target solution does not converge. At this time, the process can return to step 202, that is, updating the multiple samples used to optimize the vascular stent structure, retraining the performance prediction model, and re-determining the target solution until the target solution converges.

[0077] Similar to the aforementioned step 206, when re-executing step 202, the number of samples selected may be greater than the number of samples selected last time. Still assuming that 20 samples were selected when step 202 was last executed, if the target solution does not converge, 30 samples may be selected when re-executing step 202.

[0078] It can be seen from the above description that the present specification can train multiple vascular stent performance prediction models through a small number of samples, and then use a multi-objective optimization algorithm to determine the solution set of these vascular stent performance prediction models, and then determine the design parameter values ​​of the vascular stent from the solution set. The design parameter values ​​obtained by the multi-objective optimization algorithm can make the performance of all aspects of the optimized vascular stent reach a better state. At the same time, the whole process only requires a small number of samples to be implemented, without the need for a large amount of finite element analysis or experiments, which can greatly shorten the optimization time of the vascular stent structure and improve the efficiency of the vascular stent structure optimization.

[0079] Figure 3 is a schematic structural diagram of a device provided by an exemplary embodiment. Figure 3 At the hardware level, the device includes a processor 302, an internal bus 304, a network interface 306, a memory 308, and a non-volatile memory 310, and may also include hardware required for other functions. One or more embodiments of this specification may be implemented based on software, such as the processor 302 reading the corresponding computer program from the non-volatile memory 310 into the memory 308 and then running it. Of course, in addition to the software implementation, one or more embodiments of this specification do not exclude other implementations, such as logic devices or a combination of software and hardware, etc., that is, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0080] Please refer to Figure 4 The vascular stent structure optimization device 400 can be applied to Figure 3 The device shown in the figure is used to implement the technical solution of this specification. The device 400 for optimizing the vascular stent structure may include: A sample selection module 402 selects a plurality of samples for optimizing the vascular stent structure, wherein the samples are characterized by a plurality of design parameters of the vascular stent structure to be optimized; The label determination module 404 determines, for each sample, the standard performance index value of the sample in multiple performance dimensions using a finite element analysis method as the sample label of the sample in the corresponding performance dimension; A model training module 406, for each performance dimension, uses the sample and the sample label under the performance dimension to train a performance prediction model to obtain multiple performance prediction models, wherein each performance prediction model corresponds to one performance dimension; A solution set determination module 408, which uses a multi-objective optimization algorithm to determine a solution set of the plurality of performance prediction models, wherein the solution set includes a plurality of groups of solutions, and each group of solutions includes design parameter values ​​of the plurality of design parameters; A convergence judgment module 410 is configured to obtain a target solution from the solution set and judge whether the target solution is converged; The structure optimization module 412 optimizes the structure of the vascular stent to be optimized by using the design parameter values ​​in the target solution when the target solution converges.

[0081] Optionally, when the target solution does not converge, the sample selection module 402 updates the multiple samples for optimizing the vascular stent structure, and continues to perform the step of determining the standard performance index value of each sample under multiple performance dimensions, so as to obtain the target solution again; Among them, the number of updated samples is greater than the number of samples selected in the previous sample selection.

[0082] Optionally, each set of solutions also includes a prediction performance index value corresponding to the design parameter value, and the convergence judgment module 410: Determine the standard performance index value corresponding to the design parameter value in the target solution by using the finite element analysis method; Calculating the error between the standard performance indicator value and the predicted performance indicator value; When the error satisfies a preset condition, it is determined that the target solution converges.

[0083] Optionally, the process of the sample selection module 402 selecting a plurality of samples for optimizing the vascular stent structure includes: Determining the type of the vascular stent to be optimized; Determining the value range of each design parameter of the vascular stent structure to be optimized according to the mapping relationship between the type and the value range of the design parameter; A plurality of samples for optimizing the structure of the vascular stent are selected within the value range.

[0084] Optionally, the model training module 406 determines whether the multiple performance prediction models are all converged; When the multiple performance prediction models have all converged, the solution set determination module 408 performs the step of using a multi-objective optimization algorithm to determine the solution sets of the multiple performance prediction models.

[0085] Optionally, in the case of a non-convergent performance prediction model, the sample selection module 402 updates a plurality of samples for optimizing the vascular stent structure and continues with subsequent steps; Among them, the number of updated samples is greater than the number of samples selected in the previous sample selection.

[0086] Based on the same concept as the above method, this specification also provides an electronic device, including: a processor; a memory for storing processor executable instructions; wherein the processor implements the steps of the method described in any of the above embodiments by running the executable instructions.

[0087] Based on the same concept as the above method, this specification also provides a computer-readable storage medium on which computer instructions are stored. When the instructions are executed by a processor, the steps of the method described in any of the above embodiments are implemented.

[0088] Based on the same concept as the above method, this specification also provides a computer program product, including a computer program / instruction, which implements the steps of the method described in any of the above embodiments when executed by a processor.

Claims

1. A method for optimizing a vascular stent structure, characterized in that: The method comprises: Selecting a plurality of samples for optimizing the vascular stent structure, wherein the samples are characterized by a plurality of design parameters of the vascular stent structure to be optimized; For each sample, the finite element analysis method is used to determine the standard performance index value of the sample in multiple performance dimensions as the sample label of the sample in the corresponding performance dimension; For each performance dimension, a performance prediction model is trained using the sample and the sample label under the performance dimension to obtain multiple performance prediction models, wherein each performance prediction model corresponds to one performance dimension; Determine a solution set of the plurality of performance prediction models using a multi-objective optimization algorithm, wherein the solution set includes a plurality of groups of solutions, and each group of solutions includes design parameter values ​​of the plurality of design parameters; Obtaining a target solution from the solution set, and determining whether the target solution converges; When the target solution converges, the design parameter values ​​in the target solution are used to optimize the structure of the vascular stent to be optimized.

2. The method according to claim 1, characterized in that The method further comprises: In the case where the target solution does not converge, the multiple samples for optimizing the vascular stent structure are updated, and the step of determining the standard performance index value of each sample under multiple performance dimensions is continued to be performed to re-acquire the target solution; Among them, the number of updated samples is greater than the number of samples selected in the previous sample selection.

3. The method according to claim 1, characterized in that Each set of solutions also includes a prediction performance index value corresponding to the design parameter value. The process of determining whether the target solution converges includes: Determine the standard performance index value corresponding to the design parameter value in the target solution by using the finite element analysis method; Calculating the error between the standard performance indicator value and the predicted performance indicator value; When the error satisfies a preset condition, it is determined that the target solution converges.

4. The method according to claim 1, characterized in that The process of selecting multiple samples for optimizing the structure of the vascular stent includes: Determining the type of the vascular stent to be optimized; Determining the value range of each design parameter of the vascular stent structure to be optimized according to the mapping relationship between the type and the value range of the design parameter; A plurality of samples for optimizing the structure of the vascular stent are selected within the value range.

5. The method according to claim 1, characterized in that The method further comprises: Before using the multi-objective optimization algorithm to determine the solution sets of the multiple performance prediction models, determining whether the multiple performance prediction models are all converged; When the multiple performance prediction models are all converged, a step of using a multi-objective optimization algorithm to determine a solution set of the multiple performance prediction models is performed.

6. The method according to claim 5, characterized in that The method further comprises: In the case of a non-convergent performance prediction model, multiple samples for optimizing the vascular stent structure are updated, and the step of determining a standard performance indicator value of each sample under multiple performance dimensions is continued to be performed to retrain the performance prediction model; Among them, the number of updated samples is greater than the number of samples selected in the previous sample selection.

7. A device for optimizing a vascular stent structure, characterized in that: The device comprises: A sample selection module selects a plurality of samples for optimizing the vascular stent structure, wherein the samples are characterized by a plurality of design parameters of the vascular stent structure to be optimized; The label determination module uses the finite element analysis method to determine the standard performance index value of each sample in multiple performance dimensions as the sample label of the sample in the corresponding performance dimension; A model training module, for each performance dimension, uses the sample and the sample label under the performance dimension to train a performance prediction model to obtain multiple performance prediction models, wherein each performance prediction model corresponds to one performance dimension; A solution set determination module, which uses a multi-objective optimization algorithm to determine a solution set of the plurality of performance prediction models, wherein the solution set includes a plurality of groups of solutions, and each group of solutions includes design parameter values ​​of the plurality of design parameters; A convergence judgment module, which obtains a target solution from the solution set and judges whether the target solution converges; The structure optimization module optimizes the structure of the vascular stent to be optimized by using the design parameter values ​​in the target solution when the target solution converges.

8. An electronic device, characterized in that: include: processor; A memory for storing processor-executable instructions; wherein the processor implements the steps of the method according to any one of claims 1 to 6 by executing the executable instructions.

9. A computer-readable storage medium, characterized in that: Computer instructions are stored thereon, and when the instructions are executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that The method comprises a computer program / instruction, which, when executed by a processor, implements the steps of the method according to any one of claims 1 to 6.

Images