Esophageal imaging analysis model and method based on mechanics and impedance characteristic reconstruction

Abstract

The application relates to an esophagus imaging analysis model and method based on mechanics and impedance characteristic reconstruction, and belongs to the field of biological tissue monitoring. In the method, a flexible balloon detection device composed of a PVDF piezoelectric sensor group, a detection electrode and a control circuit is used to collect esophagus resistance and mechanical signals; the analysis model uses the PVDF piezoelectric sensor group to obtain the stress distribution of the esophagus in real time, uses electrical data obtained by the detection electrode to perform esophagus electrical impedance imaging, and obtains a high-resolution three-dimensional esophagus image; a micro-variable piezoelectric matrix is obtained in the balloon, data processing is performed through balloon composite physical field calculation, key node mechanical reconstruction is obtained, and an esophagus mechanical distribution space-time diagram is obtained by using a visualization technology; potential data of an impedance collection array is obtained outside the balloon, a Tikhonov regularization algorithm of a trust region optimization is used for processing, and an electrical impedance image is obtained; and the node mechanical image reconstruction and the electrical impedance imaging result are combined.

Description

Technical Field

[0001] This invention belongs to the field of biological tissue monitoring and relates to an esophageal imaging analysis model and method based on mechanical and impedance feature reconstruction. Background Technology

[0002] Both gastroesophageal reflux disease (GERD) and achalasia are functional esophageal disorders. The former can cause symptoms such as acid reflux and heartburn, while the latter can cause digestive problems such as difficulty swallowing and food retention.

[0003] Currently, clinical diagnosis of these two diseases mainly relies on endoscopy or esophageal manometry. Additionally, barium meals can be used to detect esophageal diseases, but barium meals contain radioactivity, which poses certain risks to the patient's health.

[0004] Existing detection technologies mostly focus on single-dimensional information. Endoscopy only provides visual information about the esophagus, and esophageal manometry only provides mechanical information about the esophagus. Early achalasia may not show obvious morphological changes, making it difficult to distinguish it from a normal esophagus by its appearance, which has certain limitations. Functional luminal imaging probes are used to detect gastrointestinal dysfunction. They can provide information about the size, mucosal integrity, and stenosis of the gastrointestinal tract, but they cannot provide information about the dynamic stress and strain of the gastrointestinal tract. Summary of the Invention

[0005] In view of this, the purpose of the present invention is to provide an esophageal imaging analysis model and method based on mechanical and impedance feature reconstruction.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] An esophageal imaging analysis model and method based on mechanical and impedance characteristic reconstruction, the method including the following steps:

[0008] The micro-variable piezoelectric matrix was obtained from inside the balloon, and the data was processed by balloon composite physical field calculation to obtain the mechanical reconstruction of key nodes. The spatiotemporal map of esophageal mechanical distribution was generated using visualization technology.

[0009] Meanwhile, the potential data of the annular impedance acquisition array obtained from outside the balloon is processed using the trust region optimized Tikhonov regularization algorithm to obtain the impedance image;

[0010] By combining mechanical reconstruction images of key esophageal nodes with electrical impedance imaging results.

[0011] Furthermore, the balloon is a flexible closed balloon, with four PVDF piezoelectric sensors evenly distributed on the catheter inside the balloon, and four electrode arrays on the outer balloon, each array containing six detection electrodes; the PVDF piezoelectric sensor group reflects the pressure distribution on the balloon and obtains the stress distribution of the esophagus in real time.

[0012] Furthermore, the key node mechanical reconstruction includes: during esophageal peristalsis, controlling the operating frequency of the PVDF piezoelectric sensor within 1Hz to 20Hz, using four PVDF piezoelectric sensors to collect m frames of piezoelectric signals, integrating them into a 4×m time-series piezoelectric signal matrix, processing the data through a balloon composite physical field calculation model, obtaining the real-time pressure of the four key nodes of the esophagus, and obtaining the key node mechanical reconstruction image through visualization technology.

[0013] Furthermore, the trust region optimization Tikhonov regularization algorithm includes:

[0014] Step 1: Initialize system parameters; determine the measurement boundary voltage, set the prior conductivity, the initial radius of the trust region, and the stopping criterion threshold ε based on the pre-obtained esophageal impedance positive simulation data, set the initial number of iterations to 0, and preset the maximum number of iterations to a value determined based on experimental experience;

[0015] Step 2: Construct the Tikhonov regularization objective function J(x);

[0016] Step 3: Calculate the current solution x k The gradient and residual of the objective function at the given point;

[0017] Step 4: Determine the optimal search direction s k ;

[0018] Step 5: Dynamically adjust the size Δk of the trust region based on model predictions and actual performance;

[0019] Step 6: Update solution x k+1 ;

[0020] Step 7: Convergence check. If any convergence condition is met, stop iterating; otherwise, adjust the trust region size and return to step 3.

[0021] Step 8: Output the conductivity distribution obtained in the current iteration.

[0022] Furthermore, the search direction s k By minimizing the following model prediction error m k Determined by (s):

[0023]

[0024] Furthermore, try including the magnitude of the gradient ||g| in the convergence condition. k The three categories are: || less than the set threshold ε, the number of iterations reaches or exceeds the set maximum number of iterations, and the change of the objective function J(x) in continuous iterations is less than the set threshold ε.

[0025] The beneficial effects of this invention are as follows:

[0026] This invention utilizes a balloon-based detection structure to reconstruct both the mechanical and electrical impedance images of key esophageal nodes. Furthermore, it incorporates electrical impedance imaging technology into esophageal detection and analysis, obtaining the characteristic distribution of esophageal electrical impedance. A trust-region optimized Tikhonov regularization algorithm is employed to ensure high accuracy, stability, and rapid convergence of the imaging results. By combining the reconstructed mechanical and electrical impedance images of the esophagus, multi-dimensional data analysis is performed on two conditions: gastroesophageal reflux and achalasia.

[0027] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0028] To make the objectives, technical solutions, and advantages of the present invention clearer, the preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, wherein:

[0029] Figure 1 This is a schematic diagram of the flexible balloon catheter detection structure described in this invention;

[0030] Figure 2 This is a schematic diagram of the balloon detection structure placed in the normal esophagus in the embodiment;

[0031] Figure 3 This is a schematic diagram of adjacent excitation of the detection electrodes in the esophagus in the embodiment;

[0032] Figure 4 This is a flowchart of the esophageal disease analysis method based on mechanical and impedance characteristic reconstruction as described in this invention.

[0033] Figure 5 This is a flowchart of the Tikhonov regularization algorithm based on trust region optimization according to the present invention;

[0034] Figure 6 This is a flowchart illustrating the analysis and matching process for two diseases, gastroesophageal reflux and achalasia, as described in this invention. Detailed Implementation

[0035] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0036] The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual pictures. They should not be construed as limiting the invention. To better illustrate the embodiments of the invention, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0037] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," "front," and "rear" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present invention. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0038] Figure 1 and Figure 2 The balloon detection device of the analytical model and method described in this invention utilizes a balloon structure composed of a PVDF piezoelectric sensor array, detection electrodes, a control circuit, and a catheter to collect electrical impedance and mechanical signals of the esophagus. The balloon is a flexible, closed balloon. Four PVDF piezoelectric sensors are evenly distributed on the catheter inside the balloon, and four electrode arrays are located outside the balloon, each array containing six detection electrodes. The PVDF piezoelectric sensor array reflects the pressure distribution on the balloon, thereby acquiring the stress distribution of the esophagus in real time.

[0039] Preferably, the boundary voltage is measured using an adjacent driving mode. The four electrode arrays on the detection balloon are numbered into four groups: D1, D2, D3, and D4. Each electrode array is then numbered, with the first layer being D1. 1-1 …D 1-6 Second layer D 2-1 …D 2-6 The third layer D 3-1 …D 3-6Fourth layer D 4-1 …D 4-6 Then select label D 1-1 D 1-2 The electrode is used as the excitation electrode, and finally, the electrode labeled D is used as the excitation electrode. 1-3 —D 1-4 D 1-4 —D 1-5 D 1-5 —D 1-6 Boundary voltage values ​​are read from these adjacent pairs of electrodes; then the excitation electrodes are switched, such as D. 1-2 —D 1-3 D 1-6 —D 1-1 Similarly, measure the voltage of other adjacent electrode pairs. Figure 3 The diagram shown illustrates the adjacent excitation of detection electrodes within a normal esophagus.

[0040] Figure 4 This is a flowchart of the esophageal disease analysis method based on mechanical and impedance characteristic reconstruction of the present invention. A micro-variable piezoelectric matrix is ​​obtained from inside the balloon, and data processing is performed through balloon composite physical field calculations to obtain the mechanical reconstruction of key nodes. A spatiotemporal map of the esophageal mechanical distribution is then generated using visualization technology. Simultaneously, potential data from a ring-shaped impedance acquisition array is obtained from outside the balloon and processed using a trust-region optimized Tikhonov regularization algorithm to obtain an impedance image. By combining the mechanical reconstruction image of key esophageal nodes with the impedance imaging results, the analysis and matching of two esophageal diseases, gastroesophageal reflux and achalasia, are completed.

[0041] The timing piezoelectric signal matrix acquired by the detection system is defined as Q(4×m), whose four row vectors represent the timing signal outputs of the first to fourth piezoelectric sensors from left to right, respectively, and m is the frame rate of the signal within the motion cycle.

[0042]

[0043] In this embodiment, the calculation of the composite physical field of the detection balloon can be decomposed into three physical fields, including the balloon strain field, the catheter strain field, and the closed flow field inside the balloon. These physical fields have multi-level coupling.

[0044] Furthermore, the strain fields of the balloon and the catheter in this embodiment are analyzed. During the strain motion of the balloon under the pressure generated by the esophageal wall, the main strains of the balloon-catheter assembly are tensile strain, bending strain, and torsional strain.

[0045] Furthermore, the strain of the balloon and catheter in the embodiments is analyzed. First, under the pressure loading of the esophageal wall, stress and strain will be generated in the balloon cavity in both the axial and radial directions. and These are the radial stress matrix and axial stress matrix of the balloon, respectively. in This is the pressure loading matrix generated by the esophageal wall, where r1 and d1 are the inner diameter and thickness of the balloon when inflated, respectively. The equations for calculating the stress in the balloon cavity are as follows:

[0046]

[0047] The equations for calculating the conversion of stress to strain in a balloon are as follows:

[0048]

[0049] Where E1 and ν1 are Young's modulus and Poisson's ratio of the balloon material, respectively.

[0050] During the gas-liquid two-phase flow within the bladder, the catheter also experiences strain. The approximate formulas for calculating the axial and radial stresses and strains of the catheter are as follows:

[0051]

[0052] Among them, P t r2 and d2 represent the pressure on the outer surface of the catheter, the inner diameter of the catheter, and the thickness of the catheter, respectively. The elastic mechanical properties of the catheter material are the same as those of the balloon material. δ r2 δ θ2 E1, E2, and ν2 represent the axial stress, radial stress, axial strain, and radial strain of the catheter under pressure testing, respectively; E2 and ν2 represent the Young's modulus and Poisson's ratio of the balloon material, respectively.

[0053] Furthermore, the closed flow field inside the balloon in the embodiment was analyzed. The balloon catheter interacts with the fluid inside the balloon, and the total stress P exerted by the fluid on the balloon catheter system is... f The calculation formula is:

[0054]

[0055] Where ρ is the liquid density, g is the standard gravitational acceleration at the measurement site, L1 is the balloon length, and q liquid P represents the dynamic volume fraction of the fluid. f_d and τ f This refers to the dynamic pressure and shear stress exerted by the fluid on the wall of the balloon catheter system. The gradient of the flow velocity along the radial direction.

[0056] Preferably, the Tikhonov regularization algorithm with trust region optimization is as follows: Figure 5 As shown, it includes:

[0057] Step 1: Initialize system parameters. Determine the measurement boundary voltage b. Based on the pre-obtained esophageal impedance positive simulation data, set an appropriate prior conductivity x0, a reasonable initial radius of the trust region Δ0, and a stopping criterion threshold ε. Set the initial number of iterations k to 0, and preset the maximum number of iterations K to a value determined based on experimental experience.

[0058] Step 2: Construct the Tikhonov regularization objective function J(x);

[0059] Step 3: Calculate the current solution x k The gradient and residual of the objective function at the given point;

[0060] Step 4: Determine the optimal search direction s k ;

[0061] Step 5: Dynamically adjust the size Δk of the trust region based on model predictions and actual performance;

[0062] Step 6: Update solution x k+1 ;

[0063] Step 7: Convergence check. If any convergence condition is met, stop iterating; otherwise, adjust the trust region size and return to step 3.

[0064] Step 8: Output the conductivity distribution obtained in the current iteration. This distribution represents the electrical impedance characteristics of the esophagus and will be used for subsequent disease analysis.

[0065] The system model needs to convert the voltage data into a discrete system of linear equations:

[0066] Ax = b

[0067] Where A is the system matrix, x is the conductivity distribution to be determined, and b is the measured voltage data;

[0068] The expression for the Tikhonov regularization objective function J(x) is:

[0069]

[0070] Where L is the regularization matrix and is the regularization parameter. It is defined as a least squares problem with a Tikhonov regularization term, which controls the smoothness of the solution to prevent overfitting due to data noise;

[0071] Preferably, the objective function J(x) is calculated using the gradient descent method at the current solution x. k gradient g at point k And calculate the residual r k =A(x k-b. Gradient and residual calculations utilize numerical analysis methods to ensure high computational accuracy;

[0072] Preferably, the trust region method limits the length of the search steps by defining a local region within a predetermined range, ensuring that each iteration is mathematically and physically feasible. Search direction s k By minimizing the following model prediction error m k Determined by (s):

[0073]

[0074] Where A is the system matrix, b is the measured boundary voltage, and x k is the solution of the current iteration, s is the search direction, Δk is the current trust region radius, λ is the regularization parameter, L is the regularization matrix, and x0 is the prior conductivity distribution.

[0075] Optionally, the optimal search direction s is determined. k Next, the trust region size Δk is dynamically adjusted to ensure the effectiveness of the iterative steps and the stability of the algorithm. The adjustment process is as follows:

[0076] First, define the ratio ρ. k The ratio of the actual decrease to the predicted decrease is calculated as follows:

[0077]

[0078] Where J(x) is the objective function with a regularization term, m k (s) is the model prediction function within the trust region, s k It represents the search direction for the current iteration.

[0079] If ρ k A value close to 1 indicates that the predicted decrease matches the actual decrease, suggesting that the current confidence region radius is appropriate. In this case, increasing the confidence region size can accelerate convergence. Adjustment strategies could include:

[0080] Δ k+1 =γ inc ·Δ k

[0081] Where, γ inc It is a growth factor greater than 1.

[0082] such as ρ k If the predicted decrease is less than the preset threshold, it indicates that the predicted decrease is much greater than the actual decrease. In this case, the confidence region size needs to be reduced to avoid instability caused by an excessively large step size. Adjustment strategies could include:

[0083] Δ k+1 =γdec ·Δ k

[0084] Where, γ dec It is a growth factor greater than 1.

[0085] If ρ k If the value is greater than or equal to another preset threshold, but less than 1, it indicates that the trust region size is relatively appropriate and the current size can be kept unchanged.

[0086] Using a defined search direction s k The expression for updating the solution is:

[0087] x k+1 =x k +Δ k ·s k

[0088] Here, x k+1 This is the updated solution, x k This is the current solution, Δk is the trust region radius adjusted based on model predictions and actual performance, and s k It is the determined optimal search direction.

[0089] Convergence conditions include the magnitude of the gradient ||g k The convergence conditions fall into three categories: || less than a set threshold ε, || the number of iterations reaches or exceeds the set maximum number of iterations, || the change in the objective function J(x) during consecutive iterations is less than the set threshold ε; The last category's convergence condition expression is:

[0090] σ k =J(x) k )-J(x k +s k )

[0091] After obtaining two-dimensional electrical impedance section images of the esophagus through imaging algorithms, the process of obtaining high-resolution three-dimensional imaging includes the following steps:

[0092] Step 1: Perform grayscale transformation on the two-dimensional image data to adjust the brightness and contrast of the image in order to better display the esophageal structure.

[0093] Step 2: Edge detection and enhancement are performed. The Laplacian operator combined with Gaussian blur is used to enhance the image's edges, effectively highlighting edges while suppressing noise. A zero-crossing method is employed to precisely locate the edges detected by the Laplacian operator, ensuring boundary accuracy. Morphological operations are combined to optimize edge contours, and morphological gradients are used to further enhance the edges of the esophageal structure.

[0094] Step 3: Use cubic spline interpolation to interpolate the space between slices. This smoothly transitions and fills the space, ensuring image quality while providing continuous 3D structural data.

[0095] Figure 6 This invention presents an analytical matching flowchart for two diseases: gastroesophageal reflux and achalasia. It utilizes the results of esophageal electrical impedance image reconstruction and nodal mechanical reconstruction to achieve accurate diagnosis of esophageal conditions. Specifically, when the esophageal electrical impedance reconstruction image shows convexity or depression accompanied by decreased pressure or reverse peristalsis at the esophagogastric junction (EGJ), it corresponds to gastroesophageal reflux. When the electrical impedance reconstruction image shows thickening or increased diameter, accompanied by increased pressure, it corresponds to achalasia. Abnormally high pressure in the lower esophagus corresponds to type I, high pressure throughout the entire esophagus corresponds to type II, and premature onset of esophageal high pressure corresponds to type III. When both the morphology and pressure analysis of the esophageal reconstruction image are normal, it corresponds to a healthy esophagus without disease.

[0096] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. An esophageal imaging analysis method based on mechanical and impedance characteristic reconstruction, implemented using a balloon detection device, characterized in that, The balloon detection device includes a flexible closed balloon, a catheter, PVDF piezoelectric sensors, and detection electrodes. The four PVDF piezoelectric sensors are evenly distributed along the axial direction of the catheter on the catheter inside the flexible closed balloon. The outer layer of the flexible closed balloon is provided with four electrode arrays along the axial direction of the catheter. Each electrode array includes six detection electrodes. The four PVDF piezoelectric sensors and the four electrode arrays are respectively arranged in the axial direction of the catheter. The esophageal imaging analysis method includes the following steps: During the same detection period of esophageal peristalsis, the four PVDF piezoelectric sensors are controlled to collect mechanical signals at an operating frequency of 1Hz to 20Hz, while the boundary voltage is collected through the four electrode arrays. Data was collected using the four PVDF piezoelectric sensors. m The frame piezoelectric signal will... m The frame piezoelectric signal is integrated into 4× m The timing piezoelectric signal matrix Q, wherein the four row vectors of the timing piezoelectric signal matrix Q represent the timing signal outputs of the four PVDF piezoelectric sensors respectively; The time-series piezoelectric signal matrix Q is processed by a balloon composite physical field calculation model to obtain the real-time pressure of four key nodes in the esophagus. The spatial-temporal map of esophageal mechanical distribution is then formed through visualization processing. The balloon composite physical field includes the balloon strain field, the duct strain field, and the closed flow field inside the balloon. Adjacent driving mode is adopted for the four electrode arrays respectively. In each electrode array, a pair of adjacent detection electrodes is selected as excitation electrodes. Boundary voltage is read between the remaining pairs of adjacent detection electrodes, and the adjacent detection electrode pairs used as excitation electrodes are switched in turn to complete the boundary voltage acquisition. The collected boundary voltage is expressed as b The voltage data is converted into a discrete linear equation system. Ax = b The conductivity distribution of the esophagus is obtained using a trust region-optimized Tikhonov regularization algorithm, the objective function of which is: In the formula, J(x ) is the Tikhonov regularization objective function. A For the system matrix, b For the measured boundary voltage, x To determine the conductivity distribution, x 0 represents the prior conductivity distribution. λ For regularization parameters, L This is the regularization matrix; Calculate the gradient and residual of the objective function at the current iterative solution, determine the search direction within the trust region, dynamically adjust the radius of the trust region and update the iterative solution based on the actual decrease and the model-predicted decrease of the objective function, stop the iteration when the convergence condition is met, output the conductivity distribution obtained in the current iteration and form an impedance image; By combining the spatiotemporal map of esophageal biomechanical distribution with the impedance image, an esophageal imaging analysis result combining biomechanical and impedance characteristics is obtained.

2. The esophageal imaging analysis method based on mechanical and impedance characteristic reconstruction according to claim 1, characterized in that, The search direction within the trust region is determined by minimizing the model prediction error shown in the following equation: In the formula, For the first k The model prediction function at the next iteration s For the search direction, Δ k For the first k Trust region radius at the next iteration x k For the first k Conductivity distribution at the next iteration; The convergence condition includes any one of the following conditions: Gradient of the objective function in the current iteration g k The modulus is less than the stopping criterion threshold ε ,Right now: ; The current iteration count has reached the preset maximum iteration count; The change in the objective function between two consecutive iterations is less than the stopping criterion threshold ε, that is: .

3. The esophageal imaging analysis method based on mechanical and impedance characteristic reconstruction according to claim 1, characterized in that, After forming the impedance image, the method further includes constructing a three-dimensional impedance image based on multiple two-dimensional impedance slice images, specifically including: Perform grayscale conversion on the two-dimensional impedance slice image; The Laplacian operator is used to process the grayscale-converted two-dimensional electrical impedance slice image, and Gaussian blur is used to reduce image noise. Image edges are extracted by detecting zero cross points. Morphological processing was performed on the two-dimensional impedance slice image after the image edges were extracted, and morphological gradient enhancement was used to enhance the image edges. A cubic spline interpolation method was used to interpolate the image data between adjacent two-dimensional electrical impedance slices to obtain a three-dimensional electrical impedance image of the esophagus.

Images