Digital phantom and library construction method for four-dimensional EIT detection of the lungs at the whole organ system level in the thoracic cavity

Abstract

This invention belongs to the field of medical imaging and digital twin technology, and discloses a method for constructing a four-dimensional EIT (Electro-Induced Thrombectomy) digital phantom and database for lung imaging at the whole-organ system level. The method includes: S1) segmenting multiple organs such as the lungs, heart, sternum, ribs, thoracic vertebrae, main bronchus, and esophagus based on individual CT images to construct a three-dimensional geometric model integrating an EIT electrode array; S2) simulating dynamic respiratory deformation using a solid mechanics model; S3) establishing the coupling relationship between deformation and conductivity to generate a dynamic three-dimensional conductivity distribution; and S4) solving the electromagnetic field forward problem to generate a virtual EIT voltage database synchronized with respiration. This invention also covers pathological model integration and data augmentation strategies. This invention can generate high-fidelity, large-scale, and diverse virtual EIT datasets, providing core data support for training and validating AI intelligent reconstruction algorithms, and laying a model foundation for developing a dynamic digital twin system for lung function.

Description

Technical Field

[0001] This invention relates to the fields of medical imaging, digital twins, and biomedical engineering technology. Specifically, it relates to a method for constructing a digital phantom for pulmonary electrical impedance imaging (EIT) based on a personalized whole-organ system model of the thoracic cavity, and a method for generating a high-fidelity virtual EIT sensor database using the phantom. Background Technology

[0002] Electrical impedance imaging (EIT), as an emerging, non-invasive, radiation-free, and real-time monitoring imaging technology, has shown great potential in clinical lung function assessment. EIT technology reconstructs the conductivity distribution in the body by placing multi-channel electrodes on the body surface, injecting weak alternating current into the body and measuring the boundary voltage of the body surface, thereby achieving dynamic visualization of the ventilation and perfusion status of the lungs.

[0003] However, traditional EIT technology has several drawbacks in clinical applications:

[0004] 1) Dimensional limitations: Traditional EIT mainly relies on two-dimensional cross-sections for reconstruction, which makes it difficult to capture the three-dimensional dynamic ventilation characteristics of the whole lung, resulting in limited information.

[0005] 2) Lack of individualization: General reconstruction algorithms often ignore the anatomical differences of thoracic organs between individuals, such as the complex effects of cardiac compression and diaphragmatic movement on lung deformation, resulting in a mismatch between the reconstruction model and the patient's actual physiological state.

[0006] 3) Data scarcity: The acquisition of large-scale, diverse, and standardized EIT experimental data in clinical practice is constrained by factors such as ethical restrictions, operational risks, and individual variability, making it difficult to meet the needs of data-driven intelligent reconstruction algorithms for a large number of training samples, thus limiting the robustness and generalization ability of the algorithms.

[0007] With the rise of the digital twin concept, the deep integration of virtual simulation and real-world monitoring has become a hot topic in interdisciplinary research. In the medical field, digital twins can integrate individualized anatomical models with multimodal data such as CT, MRI, and EIT to construct a virtual copy that reflects the patient's real-time physiological state, providing a new approach to achieving precision medicine and personalized treatment.

[0008] However, current applications of lung EIT lack a complete digital twin framework that can cover the entire thoracic cavity, encompass multi-organ linkage, couple multi-physical field characteristics, and link with real-time data. Existing research focuses on modeling single organs, while system-level digital twins for EIT need to address three major challenges: 1) the complexity of multi-organ linkage modeling; 2) cross-scale electrophysiological property mapping of tissue dielectric parameters as they dynamically change with the respiratory phase; and 3) closed-loop calibration between virtual simulation data and real measurement data.

[0009] Therefore, there is an urgent need for a digital twin phantom that can combine realistic 3D thoracic cavity structure modeling, lung motion simulation, dynamic mapping of electrical parameters, and EIT measurement data feedback to achieve high-fidelity visualization, predictable simulation, and personalized analysis of lung status, providing a solid foundation for the development and verification of next-generation EIT intelligent algorithms. Summary of the Invention

[0010] The purpose of this invention is to overcome the shortcomings of the prior art and provide a digital phantom and library construction method for four-dimensional EIT detection of the lungs at the whole organ system level of the thoracic cavity, aiming to solve the challenges of three-dimensional imaging, individualized modeling and data scarcity in EIT technology.

[0011] The core technical solution of this invention is to construct a digital phantom of the entire thoracic cavity for four-dimensional EIT detection of the lungs, and to use this phantom to achieve the following core functions:

[0012] High-quality virtual database generation: Through multi-organ coupling simulation, a standardized EIT data pool covering different anatomical variations, pathological states, and respiratory patterns is generated, providing a sufficient and high-quality training and validation foundation for AI-driven intelligent reconstruction algorithms.

[0013] Digital twin system development support: Provides a sophisticated dynamic model framework for the thoracic cavity, supporting rapid migration and personalized adaptation from virtual simulation to real clinical monitoring systems, and providing a core model foundation for the development of dynamic digital twin systems.

[0014] To achieve the above objectives, the present invention provides the following technical solution:

[0015] A method for constructing a database of digital phantoms for four-dimensional EIT detection of the lungs at the whole-organ system level includes the following steps:

[0016] Step S1: Construction of individualized three-dimensional geometric model: Based on chest cavity medical imaging data such as CT images obtained from individuals, multiple major organs in the chest cavity, such as the lungs, heart, axial skeleton, main bronchus and esophagus, are segmented, and three-dimensional geometric models of each organ are generated. These models are then integrated into a complete and individualized three-dimensional geometric model of the chest cavity. At the same time, an EIT electrode array model is integrated into the model at a preset position on the body surface.

[0017] Step S2, Multiphysics Dynamic Model Construction: Assign electrical parameters from public databases to each organ and tissue in the thoracic cavity model. Specifically, electrical parameters are usually conductivity and dielectric constant. Using a solid mechanics model, simulate the dynamic deformation process of the lungs and related organs during the respiratory cycle by applying dynamic loads and boundary conditions to simulate the respiratory process, thereby obtaining organ geometric morphology data that changes over time.

[0018] Step S3, Multiphysics Coupling and Dynamic Conductivity Mapping: Establish a coupling relationship model between organ dynamic deformation and electrical properties; map the dynamic geometric data such as lung volume changes obtained in step S2 into a spatial three-dimensional conductivity distribution model that dynamically changes over time.

[0019] Step S4: Virtual EIT Sensing Data Generation: Based on the dynamic three-dimensional conductivity distribution model and EIT electrode array model, the electromagnetic field forward problem is solved by formulas such as the Laplace equation to simulate the current excitation and voltage measurement process of EIT, and the voltage sequence generated on the boundary electrodes is calculated. These voltage data are synchronized with the dynamic deformation process of the organ and together constitute the virtual EIT sensing database.

[0020] Preferably, in step S1, the thoracic medical imaging data is CT imaging data; the segmented multiple organs include at least the two lungs, heart, axial skeleton, main bronchus and esophagus.

[0021] Preferably, in step S1, when integrating the EIT electrode array model, the electrode-skin contact impedance characteristics are simulated, and the electrode array model is arranged circumferentially along the chest model at a height between the fourth and fifth ribs.

[0022] Preferably, in step S2, the solid mechanics model simulates the action of respiratory pressure by applying a dynamically distributed force to the surface of the lung, applying a fixed constraint in the apex region of the lung, and applying a multiaxial dynamic force load in the base region of the lung to simulate the deformation driven by diaphragmatic movement.

[0023] Preferably, the method further includes a state model integration step: introducing at least one lung state model into the three-dimensional geometric model of the lung, the lung state model including a healthy model and a pathological model, and assigning specific geometric constraints and electrical parameters to the corresponding regions of the state model.

[0024] Preferably, when the lung condition model is a pathological model, the pathological model is selected from at least one of a pneumonia model, a pneumothorax model, and a lung tumor model, wherein:

[0025] When the pathological model is a pneumonia model, the defined pneumonia area is assigned a higher conductivity and dielectric constant than normal lung tissue, and a fixed constraint is applied to limit its strain.

[0026] When the pathological model is a pneumothorax model, an invisible air cavity is set in the lung apex region, and electrical parameters close to those of air are assigned to the air cavity;

[0027] When the pathological model is a lung tumor model, the segmented tumor geometric model is imported into the normal lung model, and electrical parameters higher than those of normal lung tissue are assigned to it, and fixed constraints are applied to limit its strain.

[0028] Preferably, in step S3, the electrical conductivity of the lung tissue changes with lung deformation according to the following relationship:

[0029]

[0030] in, for Time and Space Point conductivity at that point The conductivity at the end of the intake state. Maximum lung volume, for Lung volume at any given time This represents the change in electrical conductivity.

[0031] Preferably, in step S4, the EIT electrode array model includes a single-layer 16-electrode system or a dual-layer 32-electrode system; for a dual-layer 32-electrode system, the virtual EIT sensing data includes independent measurement data of the upper electrode, independent measurement data of the lower electrode, and cross-layer joint measurement data.

[0032] Preferably, the method further includes data augmentation and annotation steps:

[0033] By adjusting the breathing parameters in the solid mechanics model, breathing patterns under different tidal volumes, respiratory rates, or inspiratory-expiratory ratios are simulated to expand the virtual EIT sensor database.

[0034] Inject Gaussian white noise or contact impedance fluctuation noise into the generated virtual EIT sensing data;

[0035] The generated boundary voltage measurements are time-stamped with the corresponding breathing phase and three-dimensional conductivity distribution to construct a triplet dataset containing breathing phase, conductivity distribution, and boundary voltage.

[0036] A method for training an artificial intelligence model, based on the aforementioned thoracic whole-organ system-level four-dimensional EIT detection digital phantom and database generation method for the lungs, is characterized by comprising:

[0037] The boundary voltage measurement value in the virtual EIT sensing database is used as input, and the corresponding three-dimensional conductivity distribution is used as output to train the artificial intelligence model, so as to establish a nonlinear mapping relationship from the EIT boundary voltage to the internal three-dimensional conductivity distribution.

[0038] During the training process, the anatomical structural parameters in the individualized three-dimensional geometric model of the thoracic cavity are also incorporated as physical constraints into the artificial intelligence model.

[0039] Compared with the prior art, the present invention has the following advantages:

[0040] 1. This invention overcomes the limitations of traditional two-dimensional and general EIT models, and constructs a patient-specific three-dimensional thoracic cavity model that includes multi-organ linkage, which significantly improves the physiological realism of the model.

[0041] 2. This invention simulates lung deformation during respiration and its dynamic impact on conductivity distribution by coupling solid mechanics and electromagnetic fields, making the generated virtual data closer to the real physiological process.

[0042] 3. This invention can systematically and cost-effectively generate EIT datasets covering normal, multiple pathological states, and different respiratory modes, solving the problem of difficult clinical data collection and providing the possibility for training and validating high-performance AI reconstruction algorithms.

[0043] 4. The digital phantom and library construction method constructed in this invention provide core technologies and model frameworks for developing dynamic digital twin systems for lung function in clinical applications, and promote the development of EIT technology towards precision and intelligence. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments will be briefly introduced below.

[0045] Figure 1 This is a schematic diagram of a patient's chest CT image used for organ segmentation in this invention;

[0046] Figures 2-6 These are, respectively, three-dimensional geometric models of the thoracic cavity, axial skeleton, lungs, heart, main bronchus, and esophagus obtained from CT images after segmentation and processing in this invention;

[0047] Figure 7 This is a schematic diagram of a single-layer 16-electrode EIT sensor array model in this invention;

[0048] Figure 8 This is a schematic diagram of the double-layer 32-electrode EIT sensor array model in this invention;

[0049] Figure 9 This is a schematic diagram of the geometric model of the thoracic cavity digital phantom that integrates organ models with single-layer and double-layer EIT sensors in this invention;

[0050] Figure 10 This is a virtual measurement voltage data diagram of the single-layer 16-electrode EIT system generated in this invention;

[0051] Figure 11 This is a virtual measurement voltage data diagram of the double-layer 32-electrode EIT system generated in this invention. Detailed Implementation

[0052] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] Example 1

[0054] This embodiment describes the digital phantom construction and database generation process, detailing the complete process of a four-dimensional EIT detection digital phantom and database construction method for the whole organ system of the thoracic cavity.

[0055] Step S1: Construction of an individualized thoracic whole-organ system-level geometric model

[0056] First, organ segmentation is performed. The purpose of this step is to extract the precise three-dimensional structure of the organ from the original medical image, such as... Figure 1 As shown, using patient chest CT images as input data, medical image processing software such as Mimics and 3D Slicer are used to accurately segment the main organs in the thoracic cavity using semi-automatic or fully automatic segmentation algorithms. These organs include, but are not limited to, the lungs, heart, axial skeleton including the sternum, ribs, and thoracic vertebrae, main bronchus, and esophagus. For the segmented 3D models of each organ, denoising and smoothing processes are performed. Specifically, denoising can be achieved through median filtering, and smoothing can be achieved through Laplacian smoothing. The aim is to remove image noise and segmentation defects without affecting the key geometric features of the organs, ensuring the correct topological structure and smooth surface of the model, thus providing a high-quality geometric foundation for subsequent finite element mesh generation and physical simulation. Figures 2 to 6 It displays a processed three-dimensional geometric model of the thoracic cavity, axial skeleton, lungs, heart, and main bronchus and esophagus.

[0057] Secondly, the organ models and EIT sensors are integrated. The multiple three-dimensional organ models processed in step S1 are imported into a multiphysics simulation software platform, such as COMSOL Multiphysics. In the software, these independent organ models are precisely assembled to construct a complete, anatomically accurate, individualized geometric model of the entire thoracic cavity.

[0058] Subsequently, the integration of the EIT sensor is performed, such as... Figure 7 and Figure 8 As shown, this invention can construct EIT electrode array models with different configurations; one implementation involves constructing a model such as... Figure 7The single-layer 16-electrode ring array shown; another preferred embodiment is to construct a... Figure 8 The double-layer 32-electrode array is shown; the selected electrode array model is embedded into the surface of the thoracic cavity model, such as... Figure 9 As shown in the figure, the final form of integrating the organ model with single-layer (left) and double-layer (right) electrode arrays respectively is illustrated. The electrode arrangement follows clinical practice, for example, surrounding the patient's torso at a height between the fourth and fifth ribs of the model. The resulting whole-organ system digital phantom of the thoracic cavity provides a precise anatomical basis for all subsequent simulations.

[0059] Step S2: Construction of an individualized physical model of the entire thoracic organ system

[0060] First, electrical parameter assignment is performed. Assigning accurate electrical parameters to each tissue in the model is crucial for ensuring simulation accuracy. In this embodiment, we query and extract the electrical parameters of each biological tissue at specific frequencies from publicly available biomedical databases, such as the ITIS Database for Thermal and Electromagnetic Parameters of Biological Tissues. Specifically, the electrical parameter is conductivity. and relative permittivity For example, at the commonly used EIT frequency of 125kHz, different parts of the model are assigned values. Example parameters are shown in the table below:

[0061]

[0062] Secondly, a dynamic simulation of the lungs during the breathing process was performed. Using the solid mechanics module in COMSOL Multiphysics, the dynamic deformation process of the lungs from the end of inspiration to the end of expiration was simulated; the specific settings are as follows:

[0063] The load applies a dynamically distributed force that varies with time to the inner surfaces of the divided left and right lungs. This force is equivalent to the change in intrapulmonary pressure during respiration.

[0064] Boundary conditions: A three-dimensional fixed constraint is applied to the lung apex region to limit its displacement and simulate its connection with the top of the thorax. The strain is controlled below 5%. At the same time, a multi-axis dynamic force load is applied to the lung base region along the anatomical coordinate system to simulate the pushing effect of the diaphragm relaxation and rise on the lung base during exhalation. This setting aims to achieve a principal strain of 10%-15% in the lung base region, which is consistent with the physiological exhalation recoil characteristics.

[0065] Constitutive model: Lung tissue is assumed to be an isotropic linear elastic material, whose stress-strain relationship satisfies Hooke's law and is defined by Lamé constants λ and μ; by solving the linear elasticity equations... It can obtain displacement field and deformation data of the lungs and surrounding organs at a series of equally spaced time points during the respiratory cycle. These data accurately capture the main deformation process of the lungs from maximum volume to minimum volume.

[0066] Subsequently, disease model integration was performed. To enhance the diversity of the database, this embodiment also integrated several typical lung disease models.

[0067] Pneumonia Model: Inflammatory regions are delineated in the lung model, such as partial inflammation of the right upper lobe or complete inflammation of the right lung; these regions are assigned electrical parameters different from those of normal lung tissue, for example, conductivity is set to 0.704 S / m and dielectric constant is set to... In the mechanical simulation, a triaxial fixed constraint was applied to the inflamed area to limit its strain to less than 5% in order to simulate the characteristics of decreased lung tissue compliance caused by inflammation.

[0068] Pneumothorax model: An amorphous air cavity occupying 5%-30% of the lung volume is created in the apex region of the lung. This air cavity does not participate in deformation mechanically and is assigned electrical parameters close to those of air, specifically, conductivity of approximately 0.001 S / m and dielectric constant of approximately 1.

[0069] Tumor model: The pre-segmented geometric model of the tumor is imported into the normal lung model, and the electrical parameters of lung cancer tissue are assigned to it, for example, conductivity of 0.3597 S / m, approximately 1.6-3.3 times that of normal lung; dielectric constant of 1.11 × 10⁻⁶. 4 The strain is approximately 3-5 times that of a normal lung; in mechanical simulation, the tumor tissue is also subjected to fixed constraints to limit its strain to less than 5%.

[0070] Finally, a multiphysics coupling simulation of the respiratory process is performed. This step is one of the core aspects of this invention, coupling mechanical deformation with electrical properties. At the end of inspiration, the alveoli are fully expanded, with high air content and low conductivity; as exhalation occurs, lung volume decreases, air content decreases, and conductivity increases accordingly. This dynamic change is modeled using the following formula:

[0071]

[0072] in, yes Conductivity distribution at time t, It is the baseline conductivity distribution at the end of inhalation. Obtained through mechanical simulation lung volume at all times It is the maximum lung volume, also known as the end-inspiratory volume. It is the change in total conductivity throughout the entire respiratory cycle.

[0073] This coupled model transforms the geometric deformation data at each time point obtained from mechanical simulation into a dynamically changing, full-field three-dimensional conductivity distribution map.

[0074] Step S3: Method for establishing the dynamic virtual monitoring database for lung EIT

[0075] First, high-fidelity EIT sensor output voltage data generation is performed. Based on the established dynamic digital model, the EIT data acquisition process is simulated. Then, in the COMSOL AC / DC module, the current conservation equation, i.e., the Laplace equation, is solved. ,in It is the conductivity distribution that changes over time obtained in the previous step.

[0076] The simulation employs an adjacent excitation-measurement mode. For a single-layer 16-electrode EIT system, adjacent electrode pairs are used alternately as excitation electrodes to inject current, while voltage is measured on the remaining unexcited electrode pairs. A complete rotation measurement yields 16 × 13 / 2 = 104 independent voltage observations. Figure 10 As shown in the figure, this figure displays virtual measured voltage data of a single-layer 16-electrode EIT system generated at a certain moment.

[0077] For a dual-layer 32-electrode EIT system, the measurement modes are more complex, including independent measurements of the upper layer, independent measurements of the lower layer, and joint measurements across layers. For example, by designing a specific inter-layer excitation-measurement protocol, a single measurement can ultimately generate 328 sets of comprehensive observation datasets, such as... Figure 11 As shown in the figure, this diagram illustrates the virtual measured voltage data generated by the two-layer system.

[0078] Specifically, the comprehensive observation dataset includes 104 sets of data from independent measurements at the upper level, 104 sets of data from independent measurements at the lower level, and 120 sets of data from joint measurements across levels.

[0079] By repeating this process at each time point within the respiratory cycle, EIT voltage timing data that is completely synchronized with the dynamic changes in the lungs can be generated.

[0080] First, a multi-dimensional data augmentation strategy is implemented to improve the database's coverage and generalization ability. The following augmentation strategies are adopted:

[0081] Breathing pattern regulation: Different combinations of breathing parameters are set in the mechanical model, such as tidal volume of 300-800mL, respiratory rate of 12-30 breaths / min, and inspiratory-to-expiratory ratio of 1:1 to 1:3, to simulate various normal / abnormal breathing patterns such as calm breathing, deep breathing, and rapid breathing.

[0082] Pathological model injection: Combine the various pathological models described in step S2, third step, with different breathing patterns to generate a large amount of EIT data under pathological conditions.

[0083] Secondly, the virtual data was standardized and labeled. To make the virtual data closer to clinical reality and easier to use, the following processing was performed:

[0084] Noise injection: Based on the signal-to-noise ratio (SNR) characteristics of clinical EIT devices, which are typically 30-80dB, Gaussian white noise is added to the generated pure voltage data. At the same time, poor electrode contact can be simulated, introducing contact impedance fluctuation noise.

[0085] Spatiotemporal alignment and labeling: The virtual voltage measurement data at each time point is strictly timestamped with its corresponding simulation time, respiratory phase label, pathology type label, and the three-dimensional conductivity distribution map as the gold standard.

[0086] Data format: All data, including geometric, physical, and EIT measurement data, are organized and stored in HDF5 format files. This format supports a hierarchical structure for easy management and retrieval. A typical file structure is as follows:

[0087] / SubjectID

[0088] ├──Geometry (CT grid coordinates, electrode position matrix)

[0089] ├──Physics (Conductivity distribution tensor, dielectric constant mapping table)

[0090] ├──EIT_Data (Voltage measurement timing data, excitation mode parameters)

[0091] └──Metadata (Respiratory phase labels, pathological type markers, and mechanical parameter summaries)

[0092] Example 2

[0093] This embodiment specifically describes the application of a database in the training of an AI reconstruction algorithm. This part details how to use the database generated in Embodiment 1 to train a data-driven intelligent reconstruction algorithm.

[0094] The goal of this application is to build a deep neural network AI model that enables direct and rapid reconstruction of the internal three-dimensional conductivity variation distribution from EIT boundary voltage measurements.

[0095] First, a virtual database is constructed. On the digital twin platform, the method of Example 1 is used to simulate different breathing states such as calm breathing and deep breathing, and pathological states such as pneumonia of different locations and degrees in batches, generating a database containing tens of thousands of sets of virtual EIT data. Each set of data is a data pair of <boundary voltage vector, three-dimensional conductivity distribution map>.

[0096] Next, the AI ​​model is trained using a deep neural network architecture, such as U-Net, Transformer, or a custom encoder-decoder network. Boundary voltage vectors from the database are used as input to the network, and the corresponding 3D conductivity distribution map is used as the network output, i.e., the label or ground truth. During training, individualized anatomical parameters from the digital twin are incorporated into the network's loss function as physical constraints or prior information; specifically, the location can be the thoracic contour or organ location. This guides the network to learn a mapping relationship that not only fits the data but also conforms to the individual's anatomical patterns. Through backpropagation and optimization algorithms, specifically Adam, the difference between the network's predicted conductivity distribution and the ground truth distribution in the database is minimized until the model converges.

[0097] Then, virtual / real-time monitoring is performed. After training, the AI ​​model has the ability to reconstruct images from voltage. In virtual monitoring, newly generated voltage data from the digital twin can be input into the model to directly output the visualized distribution of the three-dimensional conductivity changes in the lungs during dynamic breathing.

[0098] In practical applications, EIT voltage data collected from real patients, along with a digital twin provided by the patient with parameters such as electrode configuration and tissue properties, are input into a trained AI model. The model can then output dynamic images of the patient's three-dimensional lung conductivity changes in real time, achieving efficient and intelligent bedside monitoring.

[0099] In summary, this invention constructs a highly realistic and dynamically evolving individualized digital phantom of the thoracic cavity, and generates a large-scale, high-quality virtual EIT database based on this. This not only provides an innovative technical path to solve the bottlenecks of traditional EIT technology, but also lays a solid foundation for developing the next generation of AI-based EIT systems that truly achieve three-dimensional, individualized, and real-time monitoring.

[0100] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A digital phantom and library construction method for four-dimensional EIT detection of the lungs at the whole organ system level in the thoracic cavity, characterized in that, Includes the following steps: Step S1: Construction of individualized three-dimensional geometric model: Based on the thoracic cavity medical imaging data obtained from the individual, multiple organs in the thoracic cavity are segmented, and three-dimensional geometric models of each organ are generated. The three-dimensional geometric models are then integrated into an individualized thoracic cavity three-dimensional geometric model containing multiple organs. An EIT electrode array model is then integrated at a preset position on the body surface of the thoracic cavity three-dimensional geometric model. Step S2, Multiphysics Dynamic Model Construction: Assign preset electrical parameters to each organ and tissue in the three-dimensional geometric model of the thoracic cavity; and use a solid mechanics model to simulate the dynamic deformation process of the lungs and related organs caused by respiratory motion during the respiratory cycle, and obtain the geometric morphology data of the organs that change over time. Step S3, Multiphysics Coupling and Dynamic Conductivity Mapping: Establish the coupling relationship between the dynamic deformation of the organ and the electrical parameters, and map the time-varying organ geometric morphology data obtained in Step S2 into a time-varying, spatially distributed three-dimensional conductivity distribution model. Step S4: Virtual EIT Sensing Data Generation: Based on the time-varying three-dimensional conductivity distribution model and the EIT electrode array model, the EIT measurement process is simulated by solving the electromagnetic field forward problem, and the boundary voltage measurement value generated on the EIT electrode array under the preset excitation mode is calculated, thereby generating a virtual EIT sensing database synchronized with the dynamic deformation process.

2. The digital phantom and library construction method for four-dimensional EIT detection of the lungs at the whole organ system level according to claim 1, characterized in that, In step S1, the thoracic medical imaging data is CT image data; the segmented multiple organs include at least the two lungs, heart, axial skeleton, main bronchus and esophagus.

3. The digital phantom and library construction method for four-dimensional EIT detection of the lungs at the whole organ system level according to claim 1, characterized in that, In step S1, when integrating the EIT electrode array model, the electrode-skin contact impedance characteristics are simulated, and the electrode array model is arranged circumferentially along the chest model at a height between the fourth and fifth ribs.

4. The digital phantom and library construction method for four-dimensional EIT detection of the lungs at the whole organ system level according to claim 2, characterized in that, In step S2, the solid mechanics model simulates the effect of respiratory pressure by applying a dynamically distributed force to the surface of the lungs, applying a fixed constraint to the apex region of the lungs, and applying a multiaxial dynamic force load to the base region of the lungs to simulate the deformation driven by diaphragmatic movement.

5. The digital phantom and library construction method for four-dimensional EIT detection of the lungs at the whole organ system level according to claim 4, characterized in that, The method further includes a state model integration step: introducing at least one lung state model into the three-dimensional geometric model of the lung, the lung state model including a healthy model and a pathological model, and assigning specific geometric constraints and electrical parameters to the corresponding regions of the state model.

6. The digital phantom and library construction method for four-dimensional EIT detection of the lungs at the whole organ system level according to claim 5, characterized in that: When the lung condition model is a pathological model, the pathological model is selected from at least one of a pneumonia model, a pneumothorax model, and a lung tumor model, wherein: When the pathological model is a pneumonia model, the defined pneumonia area is assigned a higher conductivity and dielectric constant than normal lung tissue, and a fixed constraint is applied to limit its strain. When the pathological model is a pneumothorax model, an invisible air cavity is set in the lung apex region, and electrical parameters close to those of air are assigned to the air cavity. When the pathological model is a lung tumor model, the segmented tumor geometric model is imported into the normal lung model, and electrical parameters higher than those of normal lung tissue are assigned to it, and fixed constraints are applied to limit its strain.

7. The digital phantom and library construction method for four-dimensional EIT detection of the lungs at the whole organ system level according to claim 3, characterized in that, In step S3, the electrical conductivity of the lung tissue changes with lung deformation according to the following relationship: in, for Time and Space Point conductivity at that point The conductivity at the end of the intake state. Maximum lung volume, for Lung volume at any given time This represents the change in electrical conductivity.

8. The digital phantom and library construction method for four-dimensional EIT detection of the lungs at the whole organ system level according to claim 1, characterized in that, In step S4, the EIT electrode array model includes a single-layer 16-electrode system or a dual-layer 32-electrode system; for a dual-layer 32-electrode system, the virtual EIT sensing data includes independent measurement data of the upper electrode, independent measurement data of the lower electrode, and cross-layer joint measurement data.

9. The digital phantom and library construction method for four-dimensional EIT detection of the lungs at the whole organ system level according to claim 1, characterized in that, The method also includes data augmentation and annotation steps: By adjusting the breathing parameters in the solid mechanics model, breathing patterns under different tidal volumes, respiratory rates, or inspiratory-expiratory ratios are simulated to expand the virtual EIT sensor database. Inject Gaussian white noise or contact impedance fluctuation noise into the generated virtual EIT sensing data; The generated boundary voltage measurements are time-stamped with the corresponding breathing phase and three-dimensional conductivity distribution to construct a triplet dataset containing breathing phase, conductivity distribution, and boundary voltage.

10. A method for training an artificial intelligence model using a database generated from a digital phantom and library construction method for four-dimensional EIT detection of the lungs at the whole organ system level according to any one of claims 1-9, characterized in that... include: The boundary voltage measurement value in the virtual EIT sensing database is used as input, and the corresponding three-dimensional conductivity distribution is used as output to train the artificial intelligence model, so as to establish a nonlinear mapping relationship from the EIT boundary voltage to the internal three-dimensional conductivity distribution. During the training process, the anatomical structural parameters in the individualized three-dimensional geometric model of the thoracic cavity are also incorporated as physical constraints into the artificial intelligence model.

Images