Neural network modeling method for predicting and analyzing stent performance in biomedicine
By designing an artificial neural network model, using material composition and weight percentage to predict the physical and biological characteristics of biological scaffolds, the problem of material properties and biocompatibility control in traditional methods is solved, and efficient biological scaffold design and prediction is achieved, which significantly reduces the number of biological scaffolds. Research costs and time.
Patent Information
- Application Number
- CN202411031964.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-30
- Publication Date
- 2024-11-15
AI Technical Summary
Traditional biological scaffold design relies on a large number of physical experiments and empirical models, making it difficult to accurately control the relationship between material properties and biocompatibility, the research is expensive and time-consuming, it is difficult to deal with complex nonlinear problems, and its prediction ability when facing unknown material combinations. limited.
Artificial neural network models are designed, including input layer, hidden layer and output layer, trained by Levenberg-Marquardt algorithm, using material composition and weight percentage to predict the physical and biological characteristics of the scaffold, and generate linear regression graphs to analyze scaffold performance.
Improves the accuracy and efficiency of the design, reduces experimental costs and time, can predict key performance parameters of the scaffold without actual manufacturing and testing, optimizes material design, and significantly reduces material and time costs.
Smart Images

Figure CN118966435A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the fields of bone tissue engineering and rehabilitation medicine, and in particular to a neural network modeling method for predicting and analyzing scaffold performance in biomedicine. Background Art
[0002] In the field of bone tissue engineering and rehabilitation medicine, the development of composite scaffolds with mechanical strength and biodegradability has always been a research focus. The alternative applications of multiple polymers and different materials require not only biodegradability and biocompatibility, but also suitable physicochemical properties for tissue replacement and regeneration. Scaffolds need to have both mechanical strength and biodegradability to support the needs of tissue replacement and regeneration, and provide sufficient support and stability during the rehabilitation process. Current research attempts to use a variety of materials to achieve the ideal mechanical and physiological properties of these scaffolds, as well as suitable pore structures to promote cell growth and reproduction, and have appropriate water absorption capacity for drug delivery or other medical engineering and rehabilitation applications.
[0003] In terms of predicting the performance of bioscaffolds, existing research usually relies on empirical models and mechanical tests. Although these methods can provide the necessary initial data, they often require a large amount of sample data and repeated experiments, resulting in high research costs and time-consuming. Secondly, these methods often lack flexibility and accuracy when dealing with complex nonlinear problems, such as the effects of interactions between components in multi-component material systems. Finally, traditional models are difficult to update or optimize in real time because they are usually static and cannot adapt to new data or changes. These methods have limited performance prediction capabilities for unknown material combinations and are difficult to apply to the research and development of innovative materials and the optimal design of rehabilitation scaffolds. By improving the design and performance prediction methods of bioscaffolds, the rehabilitation effect can be significantly improved and more effective rehabilitation support can be provided to patients.
[0004] The patent named "A device and method for testing the mechanical properties of a heart stent" (CN 117110108 A) mainly focuses on the mechanical properties testing of heart stents, and uses complex devices and methods to analyze the mechanical properties of heart stents under different pressure conditions; Panoramic Hengsheng's patented method relies on acquiring and processing a large amount of OCT image and IVUS image data. These image sequences not only require high-precision medical equipment for acquisition, but also require complex image processing algorithms to generate various regional masks. This not only increases the complexity and cost of the prediction process, but also requires a long processing time.
[0005] Traditional bioscaffold design relies heavily on experiments and experience, and it is difficult to accurately control the relationship between material properties and biocompatibility. Without actual manufacturing and testing of scaffolds, traditional methods are difficult to predict key performance parameters of scaffolds, such as pore size, porosity, compression strength, and elastic modulus. Traditional methods require a large number of physical experiments and repeated design iterations, consuming a lot of material and time costs. Summary of the invention
[0006] In order to solve the problems existing in the traditional bio-scaffold design process, such as reliance on a large number of physical experiments and empirical models, difficulty in accurately controlling the relationship between material properties and biocompatibility, high research costs and long time consumption, insufficient flexibility and accuracy in dealing with complex nonlinear problems, and limited prediction ability when facing unknown material combinations, the present invention provides a neural network modeling method for predicting and analyzing scaffold performance in biomedicine, which mainly includes:
[0007] S1: normalize the acquired material data of the biological scaffold to obtain training samples;
[0008] S2: Design an artificial neural network, which includes an input layer, a hidden layer, and an output layer;
[0009] S3: According to the training samples, the artificial neural network is trained using the Levenberg-Marquardt algorithm to obtain the optimal artificial neural network;
[0010] S4: Input the actual bioscaffold material into the optimal artificial neural network, output the key performance indicators of the bioscaffold and generate a linear regression graph to analyze the performance of the scaffold.
[0011] A storage device stores instructions and data for implementing a neural network modeling method for predicting and analyzing stent performance in biomedicine.
[0012] A neural network modeling device for predicting and analyzing the performance of a biomedical stent comprises: a processor and a storage device; the processor loads and executes instructions and data in the storage device to implement a neural network modeling method for predicting and analyzing the performance of a biomedical stent.
[0013] The technical solution provided by the present invention has the following beneficial effects: the present invention can predict the physical and biological properties of the scaffold according to the material composition and weight percentage by designing an artificial neural network model, thereby improving the accuracy and efficiency of the design and reducing the experimental cost. In the absence of actual manufacturing and testing of the scaffold, the pore size, porosity, compression strength, elastic modulus and other key performance parameters can be predicted through the artificial neural network without the need for actual manufacturing and testing of the scaffold, providing a reliable theoretical basis and direction for research and development. The performance prediction by the artificial neural network used in the present invention can optimize the design before actual production. The artificial neural network and data processing method show obvious superiority in theoretical prediction and optimization of material design, avoiding cumbersome physical testing processes, which not only improves the accuracy and efficiency of the design, but also reduces the number and iteration cycle of physical experiments, reduces material and time costs, and has significant competitive advantages in the field of biological scaffold design. BRIEF DESCRIPTION OF THE DRAWINGS
[0014] The present invention will be further described below with reference to the accompanying drawings and embodiments, in which:
[0015] Figure 1 It is a flow chart of a neural network modeling method for predicting and analyzing stent performance in biomedicine in an embodiment of the present invention.
[0016] Figure 2 1 is a structural diagram of an artificial neural network in an embodiment of the present invention.
[0017] Figure 3 It is a schematic diagram of the operation of the hardware device in the embodiment of the present invention. DETAILED DESCRIPTION
[0018] In order to have a clearer understanding of the technical features, purposes and effects of the present invention, specific embodiments of the present invention are now described in detail with reference to the accompanying drawings.
[0019] Please refer to Figure 1 , Figure 1 : is a flowchart of a neural network modeling method for predicting and analyzing stent performance in biomedicine according to an embodiment of the present invention, which specifically includes:
[0020] The core objective of the present invention is to obtain a bioscaffold with excellent performance by optimizing the weight percentage of any two materials (assuming material A and material B) of the scaffold. Bioscaffolds are a material used in biomedical engineering, especially widely used in tissue engineering to support cell growth and tissue regeneration. Bioscaffolds are used to support blood vessels or hearts after surgery, and must have mechanical properties. After the blood vessels or heart recover, the stents must self-dissolve in the body, so they must have autolysis ability. These scaffolds provide a three-dimensional physical structure in which cells can grow and gradually form new tissues. As new tissues are formed, biodegradable scaffolds are gradually decomposed and absorbed by natural biochemical processes in the body, and eventually completely integrated into the surrounding tissues, eliminating the need for a second surgery to remove the scaffold. Scaffold performance evaluation is to measure, analyze and judge the physical, mechanical, chemical and biological properties of an object, system or material under specific conditions to determine whether it meets the expected functions and requirements. In the study of rehabilitation stents, performance evaluation involves measuring the pore size, porosity, compressive strength, elastic modulus and absorption capacity of the stent through scanning electron microscopy (SEM), Fourier transform infrared spectroscopy (FTIR) and X-ray diffraction (XRD), artificial intelligence and other technologies. This operation requires making a prediction for a stent once until the best stent is produced. In other words, to obtain the optimal stent, it is necessary to actually produce a lot of stents, which is expensive and time-consuming. The present invention only needs to use a computer to predict the optimal stent performance and actually produce it according to the predicted data. In other words, the present invention only needs to produce the optimal stent once. These key performance indicators of biological stents are nonlinear relationships. The prior art lacks flexibility and accuracy when dealing with nonlinear problems. Therefore, the present invention uses an artificial neural network for processing.
[0021] S2: Design an artificial neural network, which includes an input layer, a hidden layer, and an output layer. This embodiment uses a shallow artificial neural network, which can achieve a good balance between complexity and generalization ability of the artificial neural network under the condition of less training data and limited sample size. After determining the weight percentage range of the materials input to the artificial neural network, the weight percentage of material A and material B is used as the input parameter of the artificial neural network. The input layer of the artificial neural network consists of two neurons. The hidden layer contains five neurons and uses the Sigmoid activation function to process the nonlinear relationship of the input data. The Sigmoid activation function is a commonly used S-shaped function whose output value varies smoothly between 0 and 1, making it suitable for mapping any real number to a range of 0 to 1. This feature can not only introduce nonlinear relationships, but also prevent problems such as gradient explosion and gradient disappearance, which helps to improve the training effect and prediction accuracy of the neural network. The output layer is responsible for predicting the key performance indicators of the scaffold, including pore size, porosity, compressive strength, elastic modulus and absorption capacity. The structure of the artificial neural network is as follows Figure 2 As shown in Figure 1, the network includes an input layer, a hidden layer, and an output layer, and the two ends of the hidden layer are connected to the input layer and the output layer respectively. The pore size and porosity are crucial to the structure and function of the stent. A larger pore size and appropriate porosity are conducive to cell penetration and tissue growth, and also improve the biocompatibility and degradation performance of the stent. The compressive strength and elastic modulus reflect the mechanical properties of the stent under external forces. Higher compressive strength and moderate elastic modulus are used to ensure that the stent can withstand physiological loads after transplantation and provide necessary mechanical support. Absorption capacity indicates the water absorption and swelling of the stent in the physiological environment, which has an important influence on the degradation rate and drug release of the stent in the body. These indicators jointly affect the overall performance of the stent and ensure its effectiveness and stability in biomedical applications. Stent transplantation refers to the transplantation of the stent under the blood vessel or heart during surgery to play a supporting role.
[0022] The pore size and porosity of scaffolds at different locations and materials have different numerical ranges, but they all have a common feature, that is, larger pore size and appropriate porosity are conducive to cell penetration and tissue growth. For example, the porosity of porous titanium scaffolds ranges from 30-70%, the porosity of PVA / calcium salt / carbon nanotube composite scaffolds is 73.4%), and the porosity of human cancellous bone is 83%. Similarly, the numerical ranges of compressive strength and elastic modulus at different locations and materials are also different.
[0023] S3: In order to improve the accuracy of prediction and minimize the prediction error, the Levenberg-Marquardt algorithm is used for back propagation training. Before the input data enters the network, the data is first normalized so that the data is distributed in the range of 0 to 1 to improve the training efficiency and stability. Then the weight percentage data of materials A and B after normalization are input into the network as training samples; finally, the network is trained using the Levenberg-Marquardt algorithm to minimize the prediction error by repeatedly adjusting the weights. After the training is completed, the prediction results are converted back to non-normalized form for practical application and comparison. Through computer artificial neural network training and data processing, the bioscaffold can be optimized and designed before actual production, and the optimal data can be predicted in order to make the optimal bioscaffold. In order to verify the predictive ability of the model, a comparison chart of the predicted results and the actual values was generated using linear regression analysis to evaluate the accuracy of the artificial neural network and finally obtain the optimal artificial neural network.
[0024] S4: Input the actual bioscaffold material into the optimal artificial neural network, output the key performance indicators of the bioscaffold and generate a linear regression graph to analyze the performance of the scaffold
[0025] The X and Y axes of the linear regression graph represent the weight percentages of the two materials, respectively, and the Z axis represents a key performance indicator. By analyzing the linear regression graph, the weight percentages of the materials can be obtained, and the performance of the bioscaffold can be determined based on the weight percentages of the two materials.
[0026] See also Figure 3 , Figure 3 4 is a schematic diagram of the working of the hardware device of an embodiment of the present invention, wherein the hardware device specifically comprises: a neural network modeling device 401 for predicting and analyzing the performance of stents in biomedicine, a processor 402 and a storage device 403.
[0027] A neural network modeling device 401 for predicting and analyzing the performance of a biomedical stent: The neural network modeling device 401 for predicting and analyzing the performance of a biomedical stent implements the neural network modeling method for predicting and analyzing the performance of a biomedical stent.
[0028] Processor 402: The processor 402 loads and executes the instructions and data in the storage device 403 to implement the neural network modeling method for predicting and analyzing the performance of stents in biomedicine.
[0029] Storage device 403: The storage device 403 stores instructions and data; the storage device 403 is used to implement the neural network modeling method for predicting and analyzing the performance of stents in biomedicine.
[0030] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. A neural network modeling method for predicting and analyzing stent performance in biomedicine. Features: include: S1: normalize the acquired material data of the biological scaffold to obtain training samples; S2: Design an artificial neural network, which includes an input layer, a hidden layer, and an output layer; S3: According to the training samples, the artificial neural network is trained using the Levenberg-Marquardt algorithm to obtain the optimal artificial neural network; S4: Input the actual bioscaffold material into the optimal artificial neural network, output the key performance indicators of the bioscaffold and generate a linear regression graph to analyze the performance of the scaffold.
2. A neural network modeling method for predicting and analyzing stent performance in biomedicine as claimed in claim 1, Features: In S2, the input layer includes 2 neurons, and the inputs of these 2 neurons are the weight percentages of any two materials of the scaffold.
3. A neural network modeling method for predicting and analyzing stent performance in biomedicine as claimed in claim 1, Features: In S2, the hidden layer contains five neurons and uses the Sigmoid activation function to process the nonlinear relationship of the input data.
4. A neural network modeling method for predicting and analyzing stent performance in biomedicine as claimed in claim 1, Features: In S2, the output layer is used to output the key performance indicators of the predicted bioscaffold, including pore size, porosity, compressive strength, elastic modulus and absorption capacity. The pore size and porosity are conducive to cell penetration and tissue growth, and also improve the biocompatibility and degradation performance of the scaffold; the compressive strength and elastic modulus are used to reflect the mechanical properties of the scaffold under the action of external forces. The compressive strength and elastic modulus are used to ensure that the scaffold can withstand physiological loads after transplantation and provide necessary mechanical support; the absorption capacity is used to indicate the water absorption and expansion of the scaffold in a physiological environment.
5. A neural network modeling method for predicting and analyzing stent performance in biomedicine as claimed in claim 1, Features: In S3, during the training process, the output prediction results are converted back to non-normalized form and compared with the actual values to evaluate the accuracy of the artificial neural network in order to obtain the optimal artificial neural network.
6. A neural network modeling method for predicting and analyzing stent performance in biomedicine as claimed in claim 4, Features: In S4, the X and Y axes of the linear regression graph represent the weight percentages of the two materials, respectively, and the Z axis represents a key performance indicator. By analyzing the linear regression graph, the weight percentages of the materials can be obtained, and the performance of the bioscaffold can be determined based on the weight percentages of the two materials.
7. A storage device, Features: The storage device stores instructions and data for implementing the neural network modeling method for predicting and analyzing the performance of stents in biomedicine as described in any one of claims 1 to 6.
8. A neural network modeling device for predicting and analyzing stent performance in biomedicine. Features: include: A processor and a storage device; the processor loads and executes instructions and data in the storage device to implement the neural network modeling method for predicting and analyzing stent performance in biomedicine as described in any one of claims 1 to 6.
Citation Information
Patent Citations
Device and method for testing mechanical property of heart stent
CN117110108A
