Lung bull vesicle rupture risk intelligent assessment system based on big data analysis
Through big data analysis and finite element modeling combined with artificial intelligence, an intelligent evaluation system was developed, which solved the problem that traditional methods were difficult to accurately predict the risk of rupture of lung large vesiculars, and achieved accurate assessment and efficient diagnosis of rupture of lung large vesiculars.
Patent Information
- Application Number
- CN202510377748.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-27
- Publication Date
- 2025-06-24
AI Technical Summary
The prior art is difficult to accurately predict the risk of rupture of large vesicles. Traditional methods rely on doctor experience and imaging examinations, and are highly subjective and difficult to accurately predict the risk of rupture.
The intelligent assessment system for rupture risk of large vesicular rupture based on big data analysis is adopted, and through data collection, feature extraction, finite element modeling and risk assessment, combined with artificial intelligence and big data analysis methods, we provide clinical intelligent and quantifiable rupture risk prediction tools.
Quantitative and accurate assessment of the risk of rupture of large vesicular rupture is achieved, which improves the scientificity and consistency of the diagnosis, reduces subjective errors of traditional methods, and improves the reliability of prediction.
Smart Images
Figure CN120197151A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of data processing, and in particular, to an intelligent evaluation system for the rupture risk of pulmonary bullae based on big data analysis. Background Art
[0002] Pulmonary bullae refer to air-containing cystic lesions formed by local abnormal inflation of lung tissue, which is one of the important manifestations of various lung diseases. The rupture of pulmonary bullae can lead to serious complications such as pneumothorax, threatening the life safety of patients. At present, the clinical evaluation of the rupture risk of pulmonary bullae mainly relies on doctors' experience and imaging examinations, but this method is highly subjective and difficult to accurately predict the rupture risk.
[0003] In recent years, the application of big data analysis technology and artificial intelligence in the medical field has provided new solutions for precision medicine. Combining big data and finite element analysis technology, through high-precision mathematical models and biomechanical simulations, the stress distribution and rupture risk of pulmonary bullae can be quantified, providing a more objective evaluation basis. However, there is currently a lack of a systematic intelligent evaluation platform that can integrate imaging data, physiological parameters, and finite element analysis technology to efficiently and accurately evaluate the rupture risk of pulmonary bullae.
[0004] The present invention proposes an intelligent evaluation system for the rupture risk of pulmonary bullae based on big data analysis. Through data collection, feature extraction, finite element modeling, and risk assessment, combined with artificial intelligence and big data analysis methods, it provides an intelligent and quantifiable rupture risk prediction tool for clinical practice, filling the deficiencies of the existing technology. Summary of the Invention
[0005] The purpose of the present invention is to solve the deficiencies existing in the prior art and provide an intelligent evaluation system for the rupture risk of pulmonary bullae based on big data analysis.
[0006] To achieve the above purpose, the present invention provides the following technical solutions:
[0007] An intelligent evaluation system for the rupture risk of pulmonary bullae based on big data analysis includes a data collection module S1, a data preprocessing module S2, a feature extraction module S3, a finite element analysis module S4, a risk assessment module S5, and a feedback module S6. The system intelligently evaluates the rupture risk of pulmonary bullae by collecting patients' lung data and physiological parameters and combining finite element algorithms and big data analysis.
[0008] Preferably, the data collection module S1 includes an imaging data collection unit S11 and a physiological parameter collection unit S12. The imaging data collection unit S11 is used to obtain CT or MRI images of the patient's lungs, and the physiological parameter collection unit S12 is used to obtain physiological information such as age, smoking history, and lung function test results. The collected data serves as the input of the evaluation system.
[0009] Preferably, the data preprocessing module S2 includes an image segmentation unit S21 and a feature region recognition unit S22, which perform denoising, enhancement, and segmentation on CT images based on a convolutional neural network (CNN) model to separate the bulla region. The mathematical expression for image data feature extraction is: F i = f CNN (I i )
[0010] where F i represents the feature vector of the i-th image, I i represents the input CT image data, and f CNN represents the convolutional neural network extraction function.
[0011] Preferably, the feature extraction module S3 fuses the image features and physiological parameters, and uses a dimensionality reduction algorithm, such as principal component analysis PCA, to extract the key feature vector G, which is used for subsequent finite element calculations. The dimensionality reduction process is expressed as: G = PCA(F, P), where G is the key feature vector, F is the set of image features, and P is the set of physiological parameters.
[0012] Preferably, the finite element analysis module S4 constructs a finite element model of the bulla, and calculates the stress distribution and strain distribution based on the patient's lung tissue characteristics and external mechanical conditions through the finite element method. The calculation formula is as follows: σ = E·ε
[0013] where σ represents the stress tensor, E is the elastic modulus of the material, and ε represents the strain tensor; the stress distribution in the bulla region is obtained by finite element solution to identify high-risk regions.
[0014] Preferably, the finite element analysis module S4 further performs dynamic loading simulation, and uses the time increment method to analyze the stress changes of the bulla under different breathing states. The time stepping equation is defined as: Δσ = Ε·Δε.
[0015] where Δσ is the stress increment within the time step, Δε is the strain increment, and the analysis results are used to identify the possibility of rupture and critical conditions; the risk assessment module S5 predicts the risk of bulla rupture based on the stress-strain analysis results and the patient's individual characteristics using a logistic regression model. The logistic regression formula is as follows:
[0016] R = σ(ω·G + b)
[0017] where R is the rupture risk score, ω is the weight vector, b is the bias term, and σ is the Sigmoid activation function, which is used to output the risk probability; the risk assessment module S5 also includes a random forest (RF) classifier for further grading the risk of bulla rupture; the decision-making process of the random forest model is defined as: C = fRF (G)
[0018] where C is the risk level label, and f RF is the random forest classification function, which determines the final risk level through the voting results of multiple decision trees; the feedback module S6 transmits the rupture risk score of the pulmonary bulla and the corresponding analysis report to the doctor, including the risk level, the key stress area, the predicted rupture probability, and the main risk factors. The report identifies the main factors related to the rupture risk based on factor analysis, and the factor contribution rate is calculated as follows:
[0019] where, λ j is the eigenvalue of the i-th factor; the system also includes a model training module, which jointly optimizes the finite element model parameters and the logistic regression model using big data. The training process adopts cross-validation, and the loss function is defined as:
[0020] where, y i is the actual rupture label, and p i is the predicted rupture probability. The training process ensures the accuracy and stability of the evaluation model.
[0021] Compared with the prior art, an intelligent evaluation system for the rupture risk of pulmonary bulla provided by the present invention has the following beneficial effects:
[0022] (1) Accurate evaluation and improved diagnostic efficiency
[0023] By integrating big data analysis and finite element mechanics modeling technology, the quantitative and accurate evaluation of the rupture risk of pulmonary bulla is realized, making up for the subjectivity deficiency of traditional reliance on doctors' experience judgment, and significantly improving the scientificity and consistency of diagnosis;
[0024] (2) Multi-dimensional data fusion and full utilization of information
[0025] Combining the patient's imaging data with physiological parameters, and adopting feature extraction and dimensionality reduction algorithms, key information is effectively extracted, greatly improving the data utilization efficiency and avoiding information redundancy or omission;
[0026] (3) Dynamic simulation and improved evaluation reliability
[0027] Through finite element dynamic loading simulation, not only the stress distribution of the pulmonary bulla in the static state is analyzed, but also the dynamic risks under different breathing states are evaluated, providing a more comprehensive reference for risk assessment;
[0028] (4) Intelligent algorithm and more scientific risk prediction
[0029] Combined with artificial intelligence algorithms such as logistic regression and random forest, it can quantitatively score and grade the risk of pulmonary bulla rupture, greatly reducing the subjective error of traditional methods and improving the reliability of prediction;
[0030] (5) Assist in decision-making and improve clinical efficiency
[0031] The risk assessment report generated by the system clearly shows the rupture probability, key stress areas, and major risk factors, helping doctors quickly understand the patient's condition, thus making more scientific clinical decisions and saving diagnostic time;
[0032] (6) Provide dynamic feedback and optimize medical management
[0033] The feedback module of the system not only provides real-time assessment results but also conducts factor analysis on the risk factors of different patients, helping doctors formulate personalized treatment plans targeted and optimizing the medical management process;
[0034] (7) Driven by big data and support continuous optimization of the model
[0035] The system uses big data technology to continuously optimize the assessment model through the model training module, ensuring that the assessment results of the system are more accurate and stable as the data samples increase and having long-term use value;
[0036] (8) Reduce medical risks and improve patient safety
[0037] Accurately identifying high-risk patients helps doctors take intervention measures early, reducing the incidence of serious complications such as pulmonary bulla rupture and pneumothorax caused by it, thus improving patient safety and treatment effects;
[0038] (9) Improve the efficiency of resource utilization and reduce the medical burden
[0039] The intelligent assessment system can process a large number of cases in a short time, reducing the repetitive work of doctors and improving the diagnostic efficiency at the same time, helping to alleviate the problem of tight medical resources;
[0040] (10) Promote the development of precision medicine and have broad application prospects
[0041] This system combines big data analysis, artificial intelligence, and biomechanical modeling technology, conforms to the trend of modern medicine towards precision medicine, has the potential to be popularized and applied in different medical institutions, and has significant social benefits. Description of the Drawings
[0042] Figure 1 It is a schematic flow structure diagram of an intelligent assessment system for the risk of pulmonary bulla rupture based on big data analysis of the present invention. Detailed Implementation Modes
[0043] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the accompanying drawings required in the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other accompanying drawings can also be obtained based on these drawings.
[0044] To enable those skilled in the art to better understand the technical solutions of the present invention, the following will further introduce the present invention in detail in conjunction with the accompanying drawings.
[0045] As Figure 1 shown, an intelligent evaluation system for the risk of pulmonary bulla rupture based on big data analysis provided by an embodiment of the present invention includes a data acquisition module S1, a data preprocessing module S2, a feature extraction module S3, a finite element analysis module S4, a risk assessment module S5, and a feedback module S6. The system intelligently evaluates the risk of pulmonary bulla rupture by collecting patients' lung data and physiological parameters and combining finite element algorithms and big data analysis.
[0046] The data acquisition module S1 includes an imaging data acquisition unit S11 and a physiological parameter acquisition unit S12. The imaging data acquisition unit S11 is used to obtain CT or MRI images of the patient's lungs, and the physiological parameter acquisition unit S12 is used to obtain physiological information such as age, smoking history, and lung function test results. The collected data serves as the input of the evaluation system.
[0047] The data preprocessing module S2 includes an image segmentation unit S21 and a feature region recognition unit S22. Based on the convolutional neural network (CNN) model, the CT image is denoised, enhanced, and segmented to separate the pulmonary bulla region. The mathematical expression for the extraction of imaging data features is: F i = f CNN (I i )
[0048] where F i represents the feature vector of the i-th image, I i represents the input CT image data, and f CNN represents the convolutional neural network extraction function.
[0049] The feature extraction module S3 fuses the imaging features and physiological parameters, and uses a dimensionality reduction algorithm such as principal component analysis PCA to extract the key feature vector G. This vector is used for subsequent finite element calculations. The dimensionality reduction process is expressed as follows: G = PCA(F, P), where G is the key feature vector, F is the set of imaging features, and P is the set of physiological parameters.
[0050] The finite element analysis module S4 constructs a finite element model of the bulla, calculates the stress distribution and strain distribution by the finite element method based on the patient's lung tissue characteristics and external mechanical conditions, and the calculation formula is as follows: σ = E·ε
[0051] Wherein, σ represents the stress tensor, E is the elastic modulus of the material, and ε represents the strain tensor; the stress distribution in the bulla region is obtained by finite element solution to identify the high-risk region.
[0052] In an embodiment provided by the present invention, the finite element analysis module S4 further performs dynamic loading simulation, and uses the time increment method to analyze the stress changes of the bulla under different breathing states. The time stepping equation is defined as follows: Δσ = Ε·Δε.
[0053] Wherein, Δσ is the stress increment within the time step, Δε is the strain increment, and the analysis results are used to identify the possibility of rupture and critical conditions; the risk assessment module S5 predicts the rupture risk of the bulla based on the stress-strain analysis results and the patient's individual characteristics, and the logistic regression formula is as follows:
[0054] R = σ(ω·G + b)
[0055] Wherein, R is the rupture risk score, ω is the weight vector, b is the bias term, and σ is the Sigmoid activation function, which is used to output the risk probability; the risk assessment module S5 also includes a random forest (RF) classifier for further grading the rupture risk of the bulla; the decision-making process of the random forest model is defined as: C = f RF (G)
[0056] Where C is the risk level label, and f RF is the random forest classification function, and the final risk level is determined by the voting results of multiple decision trees; the feedback module S6 transmits the rupture risk score of the bulla and the corresponding analysis report to the doctor, including the risk level, the key stress area, the predicted rupture probability and the main risk factors. The report identifies the main factors related to the rupture risk based on factor analysis, and the factor contribution rate is calculated as follows:
[0057] Where, λ j is the eigenvalue of the i-th factor; the system also includes a model training module, which jointly optimizes the finite element model parameters and the logistic regression model using big data. The training process adopts cross-validation, and the loss function is defined as:
[0058] Where, y i is the actual rupture label, p i is the predicted rupture probability, and the training process ensures the accuracy and stability of the evaluation model.
[0059] In another embodiment provided by the present invention, the system first collects the patient's lung image data, such as CT, MRI, and physiological parameters, such as age, smoking history, and pulmonary function test results, through the data acquisition module. The image data acquisition unit obtains high-resolution medical images, and the physiological parameter acquisition unit records the key individual information of the patient. These data together constitute the basic input for evaluation;
[0060] The data preprocessing module performs denoising, enhancement, and segmentation processing on the collected image data: processes the lung CT or MRI images through a convolutional neural network (CNN) to separate the bulla area. The preprocessing results of image segmentation and feature area recognition can clearly locate the position and contour of the bulla, providing a basis for subsequent modeling;
[0061] The feature extraction module fuses the image features of the bulla and the patient's physiological parameters, and uses a dimensionality reduction algorithm, such as principal component analysis, PCA, to extract key feature vectors: image features, such as the size, shape, density, etc. of the bulla, and physiological parameters, such as age, smoking history, etc., jointly act on the dimensionality reduction model to remove redundant data. The extracted key feature vectors provide input variables for subsequent finite element analysis and risk prediction.
[0062] The finite element analysis module constructs a finite element model of the bulla based on the patient's lung tissue characteristics and external mechanical conditions, such as the dynamic stress of the lungs caused by the respiratory cycle: quantifies the force on the bulla in static and dynamic situations through stress and strain distribution calculation formulas, identifies high-risk areas of stress concentration, and further analyzes the stress changes of the bulla under different respiratory states through dynamic loading simulation to evaluate the possibility and critical conditions of rupture;
[0063] The risk assessment module combines the stress-strain results of the finite element analysis and the individual characteristics of the patient, and uses the following two models for risk prediction: Logistic regression model: predicts the rupture probability of the bulla by calculating the rupture risk probability value score, Random forest classifier: classifies the rupture risk as low, medium, and high through the voting mechanism of decision trees. This module comprehensively considers multiple factors to achieve accurate assessment and hierarchical prediction of risks;
[0064] The feedback module generates an intuitive analysis report based on the risk assessment results, including: the rupture risk score of the bulla, the distribution map of key stress areas, the analysis of main risk factors and their weights, the rupture probability, and grading suggestions. The report can be used as a reference for doctors to formulate personalized diagnosis and treatment plans or take necessary preventive measures.
[0065] The model training module built into the system continuously optimizes the parameters of the finite element model and the weights of the logistic regression model through big data: uses cross-validation and loss function minimization techniques to ensure the prediction accuracy and stability of the model. As the data samples increase, the evaluation ability of the system is continuously improved.
[0066] Only some exemplary embodiments of the present invention have been described above by way of illustration. Without doubt, for those of ordinary skill in the art, the described embodiments can be modified in various different ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.
Claims
1. An intelligent assessment system for the risk of bullous rupture based on big data analysis, characterized in that: The system includes a data acquisition module S1, a data preprocessing module S2, a feature extraction module S3, a finite element analysis module S4, a risk assessment module S5 and a feedback module S6. The system collects the patient's lung data and physiological parameters, combines finite element algorithms and big data analysis, and conducts an intelligent assessment of the risk of rupture of bullae.
2. According to claim 1, a big data analysis-based intelligent assessment system for the risk of bullous rupture, characterized in that: The data acquisition module S1 includes an image data acquisition unit S11 and a physiological parameter acquisition unit S12. The image data acquisition unit S11 is used to acquire CT or MRI images of the patient's lungs. The physiological parameter acquisition unit S12 is used to acquire physiological information such as age, smoking history, and lung function test results. The acquired data is used as input to the evaluation system.
3. According to claim 1, a big data analysis-based intelligent assessment system for the risk of bullous rupture, characterized in that: The data preprocessing module S2 includes an image segmentation unit S21 and a feature region recognition unit S22, which performs denoising, enhancement and segmentation on the CT image based on a convolutional neural network (CNN) model to separate the pulmonary bulla region. The mathematical expression of image data feature extraction is: i =f CNN (I i ) Among them, F i Represents the feature vector of the i-th image, I i represents the input CT image data, f CNN Represents the convolutional neural network extraction function.
4. According to the intelligent assessment system for the risk of bullous rupture based on big data analysis of claim 1, it is characterized in that: The feature extraction module S3 fuses the image features with the physiological parameters and uses a dimensionality reduction algorithm, such as principal component analysis PCA, to extract a key feature vector G, which is used for subsequent finite element calculations. The dimensionality reduction process is expressed as follows: G = PCA (F, P), where G is the key feature vector, F is the image feature set, and P is the physiological parameter set.
5. According to claim 1, the intelligent assessment system for the risk of bullous rupture based on big data analysis is characterized in that: The finite element analysis module S4 constructs a finite element model of the bulla, and calculates stress distribution and strain distribution by the finite element method based on the patient's lung tissue characteristics and external mechanical conditions. The calculation formula is as follows: σ=E·ε Among them, σ represents the stress tensor, E is the elastic modulus of the material, and ε represents the strain tensor. The stress distribution in the bulla area is obtained based on the finite element solution, and the high-risk area is identified.
6. According to claim 1, a big data analysis-based intelligent assessment system for the risk of bullous rupture, characterized in that: The finite element analysis module S4 performs dynamic loading simulation and uses the time increment method to analyze the stress changes of the bullae under different respiratory states. The time step equation is defined as follows: Δσ=E·Δε, where Δσ is the stress increment within the time step and Δε is the strain increment. The analysis results are used to identify the possibility and critical conditions of rupture.
7. According to claim 1, a big data analysis-based intelligent assessment system for the risk of bullous rupture, characterized in that: The risk assessment module S5 predicts the risk of bulla rupture using a logistic regression model based on the stress-strain analysis results and individual patient characteristics. The logistic regression formula is as follows: R=σ(ω·G+b) Among them, R is the rupture risk score, ω is the weight vector, b is the bias term, and σ is the Sigmoid activation function, which is used to output the risk probability.
8. The intelligent assessment system for the risk of bullous rupture based on big data analysis according to claim 7 is characterized in that: The risk assessment module S5 also includes a random forest (RF) classifier for further classifying the risk of bulla rupture; the decision process of the random forest model is defined as: C = f RF (G) Where C is the risk level label, f RF It is a random forest classification function, which determines the final risk level through the voting results of multiple decision trees.
9. The intelligent assessment system for the risk of bullous rupture based on big data analysis according to claim 1 is characterized in that: Feedback module S6 transmits the bulla rupture risk score and the corresponding analysis report to the doctor, including the risk level, key stress area, predicted rupture probability and main risk factors. The report identifies the main factors related to rupture risk based on factor analysis. The factor contribution rate is calculated as follows: Among them, λ j is the eigenvalue of the i-th factor.
10. The intelligent assessment system for the risk of bullous rupture based on big data analysis according to claim 1 is characterized in that: The system also includes a model training module, which uses big data to jointly optimize the finite element model parameters and the logistic regression model. The training process uses cross-validation, and the loss function is defined as: Among them, yi is the actual rupture label, pi is the predicted rupture probability, and the training process ensures the accuracy and stability of the evaluation model.