Anesthesia depth monitoring method based on ultrasonic image recognition

Abstract

The application relates to the technical field of clinical anesthesia, and discloses an anesthesia depth monitoring method based on ultrasonic image recognition. The method comprises the following steps: collecting an ultrasonic image before anesthesia for image processing, using a contour tracking algorithm to obtain contour points of a target nerve anesthesia area and contour points of a shielding tissue area, calculating the areas of the two, defining initial conditions and boundary conditions, establishing a mathematical model of anesthesia diffusion by using Fick's second law, simulating a pre-anesthesia area position and area, generating a corresponding prediction image, processing the image, comparing a binary prediction image with a binary pre-anesthesia ultrasonic image, calculating an overlapping area of the target nerve anesthesia area and the pre-anesthesia area, obtaining a ratio of the overlapping area to the area of the target nerve anesthesia area, judging whether the pre-anesthesia area area contains the target nerve anesthesia area area, and determining the anesthesia depth, so that anesthesia depth monitoring is realized.

Description

Technical Field

[0001] This invention relates to the field of clinical anesthesia technology, specifically to a method for monitoring the depth of anesthesia based on ultrasound image recognition. Background Technology

[0002] Ultrasound imaging anesthesia, also known as ultrasound visualization anesthesia, is an anesthesia method that closely combines ultrasound imaging technology with anesthetic drug infusion technology. During anesthesia, doctors can observe the patient's internal anatomical structures, such as blood vessels, nerves, and muscles, in real time through the reflection and propagation characteristics of high-frequency ultrasound waves. This makes the injection of anesthetic drugs more precise and ensures that the drugs act directly on the target area.

[0003] The Chinese Invention Patent Publication No. CN116531089A proposes a data processing method for ultrasound guidance in nerve block anesthesia based on image enhancement. It uses an image enhancement algorithm to preprocess the acquired ultrasound images, making the images clearer and more recognizable. It then uses computer vision technology to extract and segment features from the preprocessed ultrasound images, identify and locate the relevant anatomical structures required for nerve block anesthesia, and finally uses an image registration algorithm to achieve accurate calculation of guidance and positioning.

[0004] The above invention, combined with existing methods, suggests that while image registration algorithms can effectively guide the puncture needle for safe operation, the concentration of anesthetic agent varies for each patient, and the location of tissue obstruction during injection differs. This instability in the diffusion of the anesthetic agent after injection can lead to incomplete anesthesia of the nerve segments requiring anesthesia, thus affecting the anesthetic effect. Therefore, an anesthesia depth monitoring method based on ultrasound image recognition is proposed to solve the above problems. Summary of the Invention

[0005] (a) Technical problems to be solved

[0006] To address the shortcomings of existing technologies, this invention provides a method for monitoring the depth of anesthesia based on ultrasound image recognition, which has advantages such as simulating the location of anesthesia diffusion and solves the problem of poor performance of existing ultrasound anesthesia depth monitoring.

[0007] (II) Technical Solution

[0008] To achieve the above objectives, the present invention provides the following technical solution: a method for monitoring anesthesia depth based on ultrasound image recognition, characterized by comprising the following steps:

[0009] S1. Determine the target area and acquire preanesthesia ultrasound images of that area. original Data, using image processing techniques to enhance image clarity;

[0010] S2. Based on the image processing data in S1, the contour tracking algorithm is used to extract the contour of the target nerve anesthesia region and the contour of the occluded tissue region, respectively, to obtain the contour point g of the target nerve anesthesia region and the contour point g' of the occluded tissue region. The area A1 of the target nerve anesthesia region is calculated using the contour point g of the target nerve anesthesia region.

[0011] S3. Initial conditions are defined based on the image processing data and physical properties of anesthetic drugs in S1. Boundary conditions are defined based on the image processing data, physical properties of anesthetic drugs, physiological structure of the occluded tissue, and contour points g' of the occluded tissue region in S1. A mathematical model of anesthetic diffusion is established based on Fick's second law. The initial and boundary conditions are substituted into the mathematical model to simulate the location and area of ​​the pre-anesthetic region, generating the corresponding predicted image I. pre .

[0012] S4, Based on predicted image I pre The data uses a contour tracking algorithm to extract the contour of the pre-anesthesia area and obtain the contour points k of the pre-anesthesia area;

[0013] S5, predict the image I pre Compared with pre-anesthesia ultrasound images I original By comparing the contour points k of the pre-anesthesia region and g of the target neuroanesthesia region, the overlapping area A is calculated. overlap Calculate the ratio of the pre-anesthetized area to the target nerve anesthesia area A1, determine whether the pre-anesthetized area includes the target nerve anesthesia area. If it does, continue the anesthesia injection while monitoring the anesthetic effect and patient response. If it does not, adjust the injection site of the anesthetic and proceed to step S3 to adjust the initial conditions.

[0014] Preferably, the image processing technology includes an adaptive histogram equalization algorithm, a bilateral filtering algorithm, and image binarization. The adaptive histogram equalization algorithm is used to enhance image contrast, the bilateral filtering algorithm is used to remove image noise and preserve edge details, and the image binarization converts the image into a binary image for segmentation and feature extraction.

[0015] Preferably, the method for calculating the area of ​​the target nerve anesthesia region and the area of ​​the obscured tissue region in S2 is as follows:

[0016] A convolutional neural network was used to segment the pre-anesthesia ultrasound image after image processing, resulting in binary images of the target nerve anesthesia region and the occluded tissue region. Each pixel represented a unit area. Contour tracking was used to extract the contours of the target nerve anesthesia region and the occluded tissue region, thereby obtaining the contour points g(x) of the target nerve anesthesia region. g y g ) and the contour point g'(x) of the occluded tissue region g’ y g’The area of ​​the target nerve anesthesia region is calculated using Green's formula:

[0017] ,

[0018] Where A1 is the area of ​​the target nerve anesthesia region, (x p+1 y p+1 )=(x1,y1), where p is the number of points on the contour of the target nerve anesthesia area.

[0019] ,

[0020] Where A2 is the area of ​​the obscured tissue region, (x g’ y g’ (x) is the g'th point on the outline of the occluded tissue region. q+1 y q+1 ) = (x1, y1), where q is the number of points on the outline of the occluded tissue region.

[0021] Preferably, in step S3, a mathematical model for the spread of anesthesia is established to generate a predicted image I. pre Way:

[0022] 1) Define initial conditions

[0023] Binarized preanesthesia ultrasound image I original Suppose that the anesthetic agent is injected into a specific location (x0, y0) at time t=0, with an initial concentration of C0. Then the initial conditions are expressed as:

[0024] ;

[0025] 2) Obtain the diffusion time propagation equation

[0026] According to Fick's second law, the diffusion equation for anesthetic drugs is:

[0027] ,

[0028] Where C(x, y, t) is the concentration at position (x, y) and time t, and D is the diffusion coefficient;

[0029] Considering the influence of the obscured tissue area, the diffusion coefficient D is modified to a position-dependent diffusion coefficient D0. Simultaneously, the space is discretized into grid points, and the time is discretized into a time step ∆t. Let the grid spacing, i.e., the spatial step, be ∆x and ∆y. The position of each grid point is represented as (i, j). Therefore, the position (x, y) is the same as the grid point position (i, j). Then, the position-dependent diffusion coefficient D0 is substituted into the anesthetic drug diffusion equation, and this equation is discretized to obtain the explicit finite difference formula:

[0030] ,

[0031] in, It is the concentration at position (i∆x, j∆y) at time t;

[0032] 3) Define boundary conditions

[0033] Free boundary condition: After the anesthetic is injected, the drug concentration gradient at the boundary is zero, i.e., the free boundary condition is: ;

[0034] Boundary condition for the anesthetic drug: During the diffusion process of the injected anesthetic drug, it cannot penetrate the boundary of the anesthetic drug area, i.e., C(x,y,t)=0.

[0035] When it encounters a certain degree of obstruction at the boundary of the occluded tissue area, it can be expressed as:

[0036] ,

[0037] Where R represents the penetration resistance, which is a constant, and C O It refers to the drug concentration within the obscured tissue area;

[0038] 4) Mathematical model simulation process

[0039] Substituting the initial and boundary conditions into the explicit finite difference formula, the concentration at position (i∆x, j∆y) at each time t is calculated, and the concentration threshold for each grid point is set to C. th The area where all grid points exceeding this threshold are summed is considered the effective pre-anesthesia area A. di Statistical analysis of all concentrations greater than C th The number of grid points N di Therefore, the area of ​​the effective pre-anesthesia region is:

[0040] A di =N di ×A p ,

[0041] Among them, A p This represents the area at each grid point;

[0042] Check the concentration at each grid point Is it greater than C? th ,if Then record the coordinates (i, j) of the grid point. Next, create a blank image of the same size as the pre-anesthesia ultrasound image, and define the effective pre-anesthesia area A. di Grid points in certain areas are marked in white, while grid points in other areas are marked in black, thus generating the predicted image I. pred .

[0043] Preferably, the method for correcting the diffusion coefficient D to a position-dependent diffusion coefficient D0 is as follows:

[0044] Because the outline of the occluded tissue region is composed of a series of points g'(x g’ y g’ Given f(x, y) to represent whether a point is located within the occluded tissue region, D eff It is the diffusion coefficient of the occluded tissue area;

[0045] When the obscured tissue area hinders the diffusion of anesthetic agents, the diffusion coefficient D of the obscured tissue area... eff =0, therefore, according to the binary mask, the unoccluded region f(x,y)=0, and the occluded region f(x,y)=1, thus we get:

[0046] ;

[0047] When the occluded tissue area slows down the diffusion of the anesthetic, the position-dependent diffusion coefficient D0 needs to vary according to position to ensure that D0 can be applied within the occluded tissue area. eff When applying D in the unobstructed area, it can be expressed by the formula:

[0048] .

[0049] Preferably, the method for obtaining the contour points of the pre-anesthesia region in the predicted image is as follows:

[0050] Using convolutional neural networks to generate predicted image I pred Segmentation is performed to obtain the predicted image I. pred The binary image is used to extract the contour of the pre-anesthesia region in the prediction image using a contour tracking method, thereby obtaining the contour point k(x). k y k ).

[0051] Preferably, the method for determining whether the area of ​​the pre-anesthesia region in S5 includes the area of ​​the target neuroanesthesia region is as follows:

[0052] Set the pre-anesthesia ultrasound image as I original Then, the binarized prediction image I pred Superimposed onto the binarized preanesthesia ultrasound image I original superior;

[0053] Generate a binary mask of the pre-anesthesia region, M pre (x) k y k ) = 1 indicates that it belongs to the point (x) k y k ) belongs to the pre-anesthesia area, otherwise M pre (x) k yk ) = 0,

[0054] Generate a binary mask of the target neuroanesthesia region, M targt (x) g y g ) = 1 indicates that it belongs to the point (x) g y g ) belongs to the target nerve anesthesia area, otherwise M targt (x) g y g ) = 0,

[0055] Substituting the contour points k of the pre-anesthesia region and g of the target neuroanesthesia region into a binary mask yields the overlapping area:

[0056] ,

[0057]

[0058] Where k∩g represents the intersection of the pre-anesthesia area and the target neuroanesthesia area, W and H are the width and height of the image, respectively, and ⋀ is the logical AND operator. For each pixel position (x, y), if M pre (x) k y k ) = 1 and M targt (x) g y g If M = 1, then M pre (x) k y k )⋀M targt (x) g y g ) = 1, otherwise, M pre (x) k y k )⋀M targt (x) g y g ) = 0;

[0059] Therefore, the ratio of the overlapping area to the area A1 of the target nerve anesthesia region is:

[0060] ,

[0061] When C ratio =1 indicates that the pre-anesthesia area and the target neuroanesthesia area completely overlap. When C... ratio If the value is less than 1, it means that the pre-anesthesia area did not cover the target neuroanesthesia area.

[0062] (III) Beneficial Effects

[0063] Compared with the prior art, the present invention provides a method for monitoring the depth of anesthesia based on ultrasound image recognition, which has the following beneficial effects:

[0064] This anesthesia depth monitoring method based on ultrasound image recognition acquires pre-anesthesia ultrasound images and processes them. A contour tracking algorithm is used to obtain the location and area data of the target nerve anesthesia area and the obscuring tissue area in the pre-anesthesia ultrasound image. A mathematical model of anesthesia diffusion is established using Fick's second law, generating corresponding predictive images, which are then processed. The binarized predictive images are compared with the binarized pre-anesthesia ultrasound images to determine the overlap between the pre-anesthesia area and the target nerve anesthesia area, thus achieving anesthesia depth monitoring. The method utilizes contour tracking algorithms, Fick's second law, and the anesthesia diffusion equation to rapidly process large amounts of image data and generate predictive results. This allows for precise control of the anesthetic drug diffusion process, real-time monitoring of the target nerve anesthesia status, improved anesthesia accuracy, and reduced risks of complications such as nerve damage, ensuring patient safety. Attached Figure Description

[0065] Figure 1 This is a schematic diagram of the process of the present invention;

[0066] Figure 2 This is a schematic diagram simulating the diffusion of the anesthetic agent of the present invention. Detailed Implementation

[0067] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. 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.

[0068] A method for monitoring anesthesia depth based on ultrasound image recognition includes the following steps:

[0069] S1. Determine the target area and acquire preanesthesia ultrasound images of that area. original Data, using image processing techniques to enhance image clarity;

[0070] S2. Based on the image processing data in S1, the contour tracking algorithm is used to extract the contour of the target nerve anesthesia region and the contour of the occluded tissue region, respectively, to obtain the contour point g of the target nerve anesthesia region and the contour point g' of the occluded tissue region. The area A1 of the target nerve anesthesia region is calculated using the contour point g of the target nerve anesthesia region.

[0071] S3. Initial conditions are defined based on the image processing data and physical properties of anesthetic drugs in S1. Boundary conditions are defined based on the image processing data, physical properties of anesthetic drugs, physiological structure of the occluded tissue, and contour points g' of the occluded tissue region in S1. A mathematical model of anesthetic diffusion is established based on Fick's second law. The initial and boundary conditions are substituted into the mathematical model to simulate the location and area of ​​the pre-anesthetic region, generating the corresponding predicted image I. pre .

[0072] S4, Based on predicted image I pre The data uses a contour tracking algorithm to extract the contour of the pre-anesthesia area and obtain the contour points k of the pre-anesthesia area;

[0073] S5, predict the image I pre Compared with pre-anesthesia ultrasound images I original By comparing the contour points k of the pre-anesthesia region and g of the target neuroanesthesia region, the overlapping area A is calculated. overlap Calculate the ratio of the pre-anesthetized area to the target nerve anesthesia area A1, determine whether the pre-anesthetized area includes the target nerve anesthesia area. If it does, continue the anesthesia injection while monitoring the anesthetic effect and patient response. If it does not, adjust the injection site of the anesthetic and proceed to step S3 to adjust the initial conditions.

[0074] In this embodiment, the image processing techniques include an adaptive histogram equalization algorithm, a bilateral filtering algorithm, and image binarization. The adaptive histogram equalization algorithm is used to enhance image contrast, the bilateral filtering algorithm is used to remove image noise and preserve edge details, and image binarization converts the image into a binary image for segmentation and feature extraction.

[0075] In this embodiment, the method for calculating the area of ​​the target nerve anesthesia region and the area of ​​the obscured tissue region in S2 is as follows:

[0076] A convolutional neural network was used to segment the pre-anesthesia ultrasound image after image processing, resulting in binary images of the target nerve anesthesia region and the occluded tissue region. Each pixel represented a unit area. Contour tracking was used to extract the contours of the target nerve anesthesia region and the occluded tissue region, thereby obtaining the contour points g(x) of the target nerve anesthesia region. g y g ) and the contour point g'(x) of the occluded tissue region g’ y g’ The area of ​​the target nerve anesthesia region is calculated using Green's formula:

[0077] ,

[0078] Where A1 is the area of ​​the target nerve anesthesia region, (x p+1 y p+1)=(x1,y1), where p is the number of points on the contour of the target nerve anesthesia area.

[0079] ,

[0080] Where A2 is the area of ​​the obscured tissue region, (x g’ y g’ (x) is the g'th point on the outline of the occluded tissue region. q+1 y q+1 ) = (x1, y1), where q is the number of points on the outline of the occluded tissue region.

[0081] In this embodiment, S3 establishes a mathematical model for anesthesia diffusion to generate a predicted image I. pre Way:

[0082] 1) Define initial conditions

[0083] Binarized preanesthesia ultrasound image I original Suppose that the anesthetic agent is injected into a specific location (x0, y0) at time t=0, with an initial concentration of C0. Then the initial conditions are expressed as:

[0084] ;

[0085] 2) Obtain the diffusion time propagation equation

[0086] According to Fick's second law, the diffusion equation for anesthetic drugs is:

[0087] ,

[0088] Where C(x, y, t) is the concentration at position (x, y) and time t, and D is the diffusion coefficient;

[0089] Considering the influence of the occluded tissue region, the diffusion coefficient D is corrected to a location-dependent diffusion coefficient D0, since the contour of the occluded tissue region consists of a series of points g'(x g’ y g’ Given f(x, y) to represent whether a point is located within the occluded tissue region, D eff It is the diffusion coefficient of the occluded tissue area;

[0090] When the obscured tissue area hinders the diffusion of anesthetic agents, the diffusion coefficient D of the obscured tissue area... eff =0, therefore, according to the binary mask, the unoccluded region f(x,y)=0, and the occluded region f(x,y)=1, thus we get:

[0091] ;

[0092] When the occluded tissue area slows down the diffusion of the anesthetic, the position-dependent diffusion coefficient D0 needs to vary according to position to ensure that D0 can be applied within the occluded tissue area. eff When applying D in the unobstructed area, it can be expressed by the formula:

[0093] ;

[0094] Simultaneously, the space is discretized into grid points, and the time is discretized into a time step ∆t. Let the grid spacing, i.e., the spatial step, be ∆x and ∆y. The position of each grid point is represented as (i, j). Therefore, the position (x, y) is the same as the grid point position (i, j). Then, the position-dependent diffusion coefficient D0 is substituted into the anesthetic drug diffusion equation, and this equation is discretized to obtain the explicit finite difference formula:

[0095] ,

[0096] in, It is the concentration at position (i∆x, j∆y) at time t;

[0097] 3) Define boundary conditions

[0098] Free boundary condition: After the anesthetic is injected, the drug concentration gradient at the boundary is zero, i.e., the free boundary condition is: ;

[0099] Boundary condition for the anesthetic drug: During the diffusion process of the injected anesthetic drug, it cannot penetrate the boundary of the anesthetic drug area, i.e., C(x,y,t)=0.

[0100] When it encounters a certain degree of obstruction at the boundary of the occluded tissue area, it can be expressed as:

[0101] ,

[0102] Where R represents the penetration resistance, which is a constant, and C O It refers to the drug concentration within the obscured tissue area;

[0103] 4) Mathematical model simulation process

[0104] Substituting the initial and boundary conditions into the explicit finite difference formula, the concentration at position (i∆x, j∆y) at each time t is calculated, and the concentration threshold for each grid point is set to C. th The area where all grid points exceeding this threshold are summed is considered the effective pre-anesthesia area A. di Statistical analysis of all concentrations greater than C th The number of grid points N di Therefore, the area of ​​the effective pre-anesthesia region is:

[0105] A di =N di ×A p ,

[0106] Among them, A p This represents the area at each grid point;

[0107] Check the concentration at each grid point Is it greater than C? th ,if Then record the coordinates (i, j) of the grid point. Next, create a blank image of the same size as the pre-anesthesia ultrasound image, and define the effective pre-anesthesia area A. di Grid points in certain areas are marked in white, while grid points in other areas are marked in black, thus generating the predicted image I. pred .

[0108] It should be noted that obtaining the diffusion time progression equation in step two above is to allow the anesthetic drug to diffuse over a certain period of time, so that the diffusion state can evolve over time. For image generation, time progression can be achieved through an iterative update mechanism, where each time step is adjusted based on the state of the previous moment. Defining boundary conditions in step three above helps to constrain the image generation process, making the output more stable and closer to real-world phenomena. Initialization in step one above refers to setting an initial state for the mathematical model, that is, defining the state of the anesthetic drug diffusion at the beginning. For image generation, this usually means providing an initial input image or random noise as a starting point.

[0109] When anesthetic drugs enter the obstructing tissue, the diffusion coefficient D needs to be corrected according to the location, taking into account the influence of the obstructing tissue area on the diffusion equation of anesthetic drugs, so as to reflect the different degrees of obstruction to drug diffusion by different tissues.

[0110] In this embodiment, the method for obtaining the contour points of the pre-anesthesia region in the predicted image is as follows:

[0111] Using convolutional neural networks to generate predicted image I pred Segmentation is performed to obtain the predicted image I. pred The binary image is used to extract the contour of the pre-anesthesia region in the prediction image using a contour tracking method, thereby obtaining the contour point k(x). k y k ).

[0112] In this embodiment, the method for determining whether the area of ​​the pre-anesthesia region in S5 includes the area of ​​the target nerve anesthesia region is as follows:

[0113] Set the pre-anesthesia ultrasound image as I original Then, the binarized prediction image I predSuperimposed onto the binarized preanesthesia ultrasound image I original superior;

[0114] Generate a binary mask of the pre-anesthesia region, M pre (x) k y k ) = 1 indicates that it belongs to the point (x) k y k ) belongs to the pre-anesthesia area, otherwise M pre (x) k y k ) = 0,

[0115] Generate a binary mask of the target neuroanesthesia region, M targt (x) g y g ) = 1 indicates that it belongs to the point (x) g y g ) belongs to the target nerve anesthesia area, otherwise M targt (x) g y g ) = 0,

[0116] Substituting the contour points k of the pre-anesthesia region and g of the target neuroanesthesia region into a binary mask yields the overlapping area:

[0117] ,

[0118] ,

[0119] Where k∩g represents the intersection of the pre-anesthesia area and the target neuroanesthesia area, W and H are the width and height of the image, respectively, and ⋀ is the logical AND operator. For each pixel position (x, y), if M pre (x) k y k ) = 1 and M targt (x) g y g If M = 1, then M pre (x) k y k )⋀M targt (x) g y g ) = 1, otherwise, M pre (x) k y k )⋀M targt (x) g y g ) = 0;

[0120] Therefore, the ratio of the overlapping area to the area A1 of the target nerve anesthesia region is:

[0121] ,

[0122] When C ratio =1 indicates that the pre-anesthesia area and the target neuroanesthesia area completely overlap. When C... ratio If the value is less than 1, it means that the pre-anesthesia area did not cover the target neuroanesthesia area.

[0123] It should be noted that C ratio =1 indicates that the pre-anesthesia area and the target nerve anesthesia area completely overlap, meaning that the anesthetic diffuses and covers the target nerve anesthesia area during injection. In actual injection, the nerve's anesthesia and relaxation can be visually observed on ultrasound images. Injection can be stopped when the nerve is fully relaxed and there is still anesthetic remaining to avoid overdose. When C... ratio If the value is less than 1, it means that the pre-anesthesia area failed to cover the target nerve anesthesia area. This indicates that the diffusion of the anesthetic drug dose could not cover the nerve that needs anesthesia during the simulated diffusion process. Since the anesthetic drug is quantitatively appropriate, the initial injection conditions need to be adjusted and the simulation repeated to ensure that the anesthetic drug injection diffusion completely covers the target nerve anesthesia and avoid the occurrence of incomplete anesthesia or no anesthesia.

[0124] 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 method for monitoring anesthesia depth based on ultrasound image recognition, characterized in that, Includes the following steps: S1. Determine the target area and acquire preanesthesia ultrasound images of that area. original Data, using image processing techniques to enhance image clarity; S2. Based on the image processing data in S1, the contour tracking algorithm is used to extract the contour of the target nerve anesthesia region and the contour of the occluded tissue region, respectively, to obtain the contour point g of the target nerve anesthesia region and the contour point g' of the occluded tissue region. The area A1 of the target nerve anesthesia region is calculated using the contour point g of the target nerve anesthesia region. S3. Initial conditions are defined based on the image processing data and physical properties of anesthetic drugs in S1. Boundary conditions are defined based on the image processing data, physical properties of anesthetic drugs, physiological structure of the occluded tissue, and contour points g' of the occluded tissue region in S1. A mathematical model of anesthetic diffusion is established based on Fick's second law. The initial and boundary conditions are substituted into the mathematical model to simulate the location and area of ​​the pre-anesthetic region, generating the corresponding predicted image I. pre ; S4, Based on predicted image I pre The data uses a contour tracking algorithm to extract the contour of the pre-anesthesia area and obtain the contour points k of the pre-anesthesia area; S5, predict the image I pre Compared with pre-anesthesia ultrasound images I original By comparing the contour points k of the pre-anesthesia region and g of the target neuroanesthesia region, the overlapping area A is calculated. overlap Calculate the ratio of the pre-anesthesia area to the area A1 of the target nerve anesthesia area, and determine whether the area of ​​the pre-anesthesia area includes the area of ​​the target nerve anesthesia area. The method for calculating the area of ​​the target nerve anesthesia region and the area of ​​the obscured tissue region in S2 includes the following steps: A convolutional neural network was used to segment the pre-anesthesia ultrasound image after image processing, resulting in binary images of the target nerve anesthesia region and the occluded tissue region. Each pixel represented a unit area. Contour tracking was used to extract the contours of the target nerve anesthesia region and the occluded tissue region, respectively, thereby obtaining the contour points g(x) of the target nerve anesthesia region. g y g ) and the contour point g'(x) of the occluded tissue region g’ y g The area of ​​the target nerve anesthesia region is calculated using Green's formula.

2. The method for monitoring anesthesia depth based on ultrasound image recognition according to claim 1, characterized in that, The image processing techniques include an adaptive histogram equalization algorithm, a bilateral filtering algorithm, and image binarization. The adaptive histogram equalization algorithm is used to enhance image contrast, the bilateral filtering algorithm is used to remove image noise and preserve edge details, and the image binarization converts the image into a binary image for segmentation and feature extraction.

3. The method for monitoring anesthesia depth based on ultrasound image recognition according to claim 1, characterized in that, In step S3, a mathematical model for the spread of anesthesia is established to generate a predicted image I. pre The method is as follows: 1) Define initial conditions Binarized preanesthesia ultrasound image I original Suppose that the anesthetic agent is injected into a specific location (x0, y0) at time t=0, with an initial concentration of C0. Then the initial conditions are expressed as: 2) Obtain the diffusion time propagation equation According to Fick's second law, the diffusion equation for anesthetic drugs is: Where C(x, y, t) is the concentration at position (x, y) and time t, and D is the diffusion coefficient; Considering the influence of the obscured tissue area, the diffusion coefficient D is modified to a position-dependent diffusion coefficient D0. Simultaneously, the space is discretized into grid points, and the time is discretized into a time step Δt. Let the grid spacing, i.e., the spatial step, be Δx and Δy. The position of each grid point is represented as (i, j). Therefore, the position (x, y) is the same as the grid point position (i, j). Then, the position-dependent diffusion coefficient D0 is substituted into the anesthetic drug diffusion equation, and this equation is discretized to obtain the explicit finite difference formula: in, It is the concentration at position (iΔx, jΔy) at time t; 3) Define boundary conditions Free boundary condition: After the anesthetic is injected, the drug concentration gradient at the boundary is zero, i.e., the free boundary condition is: Boundary condition for the anesthetic drug: During the diffusion process of the injected anesthetic drug, it cannot penetrate the boundary of the anesthetic drug area, i.e., C(x, y, t) = 0. When it encounters a certain degree of obstruction at the boundary of the occluded tissue area, it can be expressed as: Where R represents the penetration resistance, which is a constant, and C O ' represents the drug concentration within the obscured tissue area; 4) Mathematical model simulation process Substituting the initial and boundary conditions into the explicit finite difference formula, the concentration at position (iΔx, jΔy) at each time t is calculated, and the concentration threshold for each grid point is set to C. th The area where all grid points exceeding this threshold are summed is considered the effective pre-anesthesia area A. di Statistical analysis of all concentrations greater than C th The number of grid points N di Therefore, the area of ​​the effective pre-anesthesia region is: A di =N di ×A p , Among them, A p This represents the area at each grid point; Check the concentration at each grid point Is it greater than C? th ,if Record the coordinates (i, j) of the grid point. Then, create a blank image of the same size as the pre-anesthesia ultrasound image and mark the effective pre-anesthesia area A. di Grid points in certain areas are marked in white, while grid points in other areas are marked in black, thus generating the predicted image I. pred .

4. The method for monitoring anesthesia depth based on ultrasound image recognition according to claim 3, characterized in that, The location-dependent diffusion coefficient D0 is calculated as follows: Because the outline of the occluded tissue region is composed of a series of points g'(x g’ y g Given f(x, y) to represent whether a point is located within the occluded tissue region, D eff It is the diffusion coefficient of the occluded tissue area; When the obscured tissue area hinders the diffusion of anesthetic agents, the diffusion coefficient D of the obscured tissue area... eff =0, therefore, according to the binary mask, the unoccluded region f(x,y) = 0, and the occluded region f(x,y) = 1, thus we get: When the occluded tissue area slows down the diffusion of the anesthetic, the position-dependent diffusion coefficient D0 needs to vary according to position to ensure that D0 can be applied within the occluded tissue area. eff When applying D in the unobstructed area, it can be expressed by the formula: D0(x,y)=(1-f(x,y))·D+f(x,y)D eff 。 5. The method for monitoring anesthesia depth based on ultrasound image recognition according to claim 3, characterized in that, The method for obtaining the contour points of the pre-anesthesia region in the predicted image is as follows: Using convolutional neural networks to generate predicted image I pred Segmentation is performed to obtain the predicted image I. pred The binary image is used to extract the contour of the pre-anesthesia region in the prediction image using a contour tracking method, thereby obtaining the contour point k(x). k y k ).

6. The method for monitoring anesthesia depth based on ultrasound image recognition according to claim 4, characterized in that, The method for determining whether the area of ​​the pre-anesthesia region in S5 includes the area of ​​the target nerve anesthesia region is as follows: Set the pre-anesthesia ultrasound image as I original Then, the binarized prediction image I pred Superimposed onto the binarized preanesthesia ultrasound image I original superior; Generate a binary mask of the pre-anesthesia region, M pre (x k y k ) = 1 indicates that it belongs to the point (x) k y k ) belongs to the pre-anesthesia area, otherwise M pre (x k y k ) = 0, Generate a binary mask of the target neuroanesthesia region, M targt (x g y g ) = 1 indicates that it belongs to the point (x) g y g ) belongs to the target nerve anesthesia area, otherwise M targt (x g y g ) = 0, Substituting the contour points k of the pre-anesthesia region and g of the target neuroanesthesia region into the binary mask yields the overlapping area: A overlap =∑(k∩g), Where k∩g represents the intersection of the pre-anesthesia area and the target neuroanesthesia area, W and H are the width and height of the image, respectively, and ∧ is the logical AND operator. For each pixel position (x, y), if M pre (x k y k ) = 1 and M targt (x g y g If ) = 1, then M pre (x k y k )∧M targt (x g y g ) = 1, otherwise, M pre (x k y k )∧M targt (x g y g ) = 0; Therefore, the ratio of the overlapping area to the area of ​​the target nerve anesthesia region is: When C ratio =1 indicates that the pre-anesthesia area and the target nerve anesthesia area completely overlap. When C ratio If the value is less than 1, it means that the pre-anesthesia area did not cover the target neuroanesthesia area.

Images