Tympanic membrane puncture positioning method and system based on clinical images
By acquiring layered and clinical images, extracting edge contour data, and constructing a triangular cone model, the problem of insufficient positioning accuracy for tympanic membrane puncture in existing technologies has been solved, achieving a more precise and stable puncture operation.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CAPITAL UNIVERSITY OF MEDICAL SCIENCES
- Filing Date
- 2026-01-14
- Publication Date
- 2026-06-19
Smart Images
Figure CN121606439B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the technical field of tympanic membrane puncture localization, and more particularly to a tympanic membrane puncture localization method and system based on clinical images. Background Technology
[0002] Tympanic membrane puncture localization is a core technology used in otolaryngology. Due to its intuitive and minimally invasive characteristics, the localization method based on clinical images has broad application prospects in the routine diagnosis and treatment of adults and patients with good cooperation. Its accurate image localization can provide a reliable basis for puncture operation and reduce medical risks.
[0003] Currently, existing localization methods based on clinical images mostly rely on doctors to make direct judgments through ordinary endoscopic images, or only perform edge extraction on single-dimensional images to determine the puncture area. Some methods use basic coordinate calculations to assist in localization, but lack a comprehensive analysis of the spatial structure of the tympanic membrane.
[0004] Therefore, these methods are easily affected by image clarity, doctors' subjective experience, etc., making it difficult to accurately capture the anatomical features and spatial relationship of the tympanic membrane, resulting in deviation of the puncture target point or inaccurate control of the puncture depth. Summary of the Invention
[0005] The purpose of this application is to provide a method and system for tympanic membrane puncture localization based on clinical images, so as to solve the problem of insufficient tympanic membrane puncture localization accuracy in the prior art.
[0006] To address the aforementioned technical problems, in a first aspect, this application provides a method for tympanic membrane puncture localization based on clinical images, comprising:
[0007] Acquire layered images of the target ear and clinical images of the tympanic membrane;
[0008] The edge contour data of the tympanic membrane is extracted from the layered image, and the light cone apex and umbilicus of the tympanic membrane are identified from the clinical image as the first reference point;
[0009] At least three non-collinear auxiliary positioning markers are set on the surface of the concha cavity as second reference points;
[0010] Based on the edge contour data and the first reference point, the puncture target point is determined, and the second reference point and the puncture target point are connected to construct a triangular pyramid model.
[0011] Based on the triangular pyramid model, a spatial geometric positioning method is used to calculate the needle insertion angle and puncture depth, so as to control the optical guidance device to indicate the puncture path and complete the positioning.
[0012] Optionally, the step of determining the puncture target point based on the edge contour data and the first reference point, connecting the second reference point and the puncture target point, and constructing a triangular pyramid model includes:
[0013] Based on the edge contour data, curve fitting is performed to obtain a closed contour curve, and the center projection point is determined within the area enclosed by the closed contour curve.
[0014] The center projection point and the first reference point are balanced to obtain the puncture target point, and an auxiliary line is constructed from the second reference point to the puncture target point.
[0015] Construct triangular facets between adjacent auxiliary lines, and combine all the triangular facets with the base formed by the auxiliary lines to form a triangular pyramid model.
[0016] Optionally, the step of performing curve fitting processing based on the tympanic membrane edge contour data to obtain a closed contour curve, and determining the center projection point within the area enclosed by the closed contour curve, includes:
[0017] Based on the edge contour data, the curvature distribution characteristics of the tympanic membrane edge contour are identified, and based on the curvature distribution characteristics, the contour stability area and the contour abnormal area within the tympanic membrane edge contour are identified.
[0018] Based on the stable contour region, a contour framework is constructed. Based on the contour framework, an elastic constraint algorithm is used to reshape the abnormal contour region to obtain the target segment.
[0019] The contour frame is combined with the target segment to obtain a closed contour curve. Based on the geometric features of the closed contour curve, the center projection point is calculated.
[0020] Optionally, the step of reshaping the contour anomaly region based on the contour framework using an elastic constraint algorithm to obtain the target segment includes:
[0021] Based on the geometric constraints of the contour frame, multiple deformation control points are set in the contour anomaly area, and an elastic constraint algorithm is used to establish an elastic connection relationship between adjacent deformation control points.
[0022] Based on the tension distribution corresponding to the elastic connection relationship, the abnormal contour area is subjected to deformation simulation processing to generate an initial segment;
[0023] The initial segment is aligned with the outline frame to obtain the intermediate segment;
[0024] The intermediate segment is optimized to obtain a target segment that matches the curvature characteristics of the contour stabilization region.
[0025] Optionally, the calculation of the needle insertion angle and puncture depth based on the triangular pyramid model and using a spatial geometric positioning method includes:
[0026] Extract tissue boundary data from the layered image;
[0027] Based on the tissue boundary data, a needle insertion depth threshold is determined in the vertical projection direction of the puncture target point;
[0028] Determine the normal vector of the puncture target point with respect to the tympanic membrane, and based on the normal vector of the puncture target point and the normal vector of the entrance plane of the external auditory canal, calculate the puncture needle insertion direction through the spatial constraints of the triangular pyramid model;
[0029] The average spatial distance from the second reference point to the puncture target point is calculated as the puncture depth, which is less than the needle insertion depth threshold.
[0030] Optionally, determining the sectional normal vector of the puncture target point with respect to the tympanic membrane, and calculating the puncture needle insertion direction based on the sectional normal vector and the normal vector of the entrance plane of the external auditory canal, using the spatial constraints of a triangular pyramid model, includes:
[0031] The tympanic membrane surface sampling points in the neighborhood of the puncture target point are subjected to surface fitting processing to obtain the tangent plane of the tympanic membrane, and the normal vector perpendicular to the tangent plane is taken as the first normal vector.
[0032] Plane fitting is performed based on the spatial distribution of the second reference point to obtain the entrance plane of the external auditory canal, and the normal vector perpendicular to the entrance plane is taken as the second normal vector.
[0033] The first normal vector is projected onto the entrance plane of the external auditory canal to obtain the first projection component, and the second normal vector is projected onto the tangent plane to obtain the second projection component;
[0034] The first projection component and the second projection component are vector-synthesized to generate the initial needle insertion direction;
[0035] Using a coordinate calibration method, the initial needle insertion direction is optimized through the spatial constraints of a triangular pyramid model to obtain the puncture needle insertion direction. The angle between the puncture needle insertion direction and the normal vector of the tangent plane satisfies a preset perpendicularity condition, and the spatial relationship with the second normal vector satisfies a preset coplanarity requirement.
[0036] Optionally, identifying the light cone apex of the tympanic membrane and the umbilicus from the clinical image as a first reference point includes:
[0037] Reflectance analysis is performed on the clinical image to determine the reflection region, and extreme points are located within the reflection region to obtain candidate vertex points;
[0038] The edge contour data is fitted to obtain the axis of symmetry. Based on the axis of symmetry, the depression features in the tympanic membrane surface region are searched to obtain candidate regions.
[0039] Curvature extremum analysis is performed within the candidate region to obtain candidate points for the umbilicus;
[0040] Based on preset spatial constraints, the vertex candidate points and the navel candidate points are verified to obtain the first reference point.
[0041] Secondly, this application provides a tympanic membrane puncture localization system based on clinical images, comprising:
[0042] The acquisition module is used to acquire layered images of the target ear and clinical images of the tympanic membrane;
[0043] The recognition module is used to extract edge contour data of the tympanic membrane from the layered image and to identify the light cone apex and umbilicus of the tympanic membrane from the clinical image as a first reference point.
[0044] The module is used to set at least three non-collinear auxiliary positioning markers on the surface of the concha cavity as second reference points;
[0045] The construction module is used to determine the puncture target point based on the edge contour data and the first reference point, connect the second reference point and the puncture target point, and construct a triangular pyramid model;
[0046] The calculation module is used to calculate the needle insertion angle and puncture depth based on the triangular pyramid model and spatial geometric positioning method, so as to control the optical guidance device to indicate the puncture path and complete the positioning.
[0047] Thirdly, this application provides an electronic device, comprising:
[0048] Memory, used to store computer programs;
[0049] A processor, configured to execute the computer program to implement the steps of the tympanic membrane puncture localization method based on clinical images as described in the first aspect above.
[0050] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps of the tympanic membrane puncture localization method based on clinical images as described in the first aspect above.
[0051] The tympanic membrane puncture localization method based on clinical images provided in this application first acquires layered images of the target ear and clinical images of the tympanic membrane. Then, it extracts the edge contour data of the tympanic membrane from the layered images and identifies the apex of the light cone and the umbilicus of the tympanic membrane from the clinical images as first reference points. Subsequently, at least three non-collinear auxiliary positioning markers are set on the surface of the concha cavity as second reference points. Then, based on the edge contour data and the first reference points, the puncture target point is determined, and the second reference points are connected to the puncture target point to construct a triangular pyramid model. Finally, based on the triangular pyramid model, a spatial geometric positioning method is used to calculate the needle insertion angle and puncture depth to control the optical guidance device to indicate the puncture path.
[0052] The technical solution of this application has the following beneficial effects:
[0053] This application provides comprehensive and fundamental data support for subsequent localization by acquiring layered images of the target ear and clinical images of the tympanic membrane. Next, it extracts the edge contour data of the tympanic membrane from the layered images and identifies a first reference point from the clinical images, accurately capturing the morphological features and key anatomical locations of the tympanic membrane and providing a reliable reference for determining the puncture target point. Then, a second reference point is set on the surface of the concha cavity, and surface spatial anchor points are established to fill the gap in image-based localization lacking physical spatial reference. The puncture target point is then determined based on the edge contour data and the first reference point, and a triangular pyramid model is constructed to clarify the spatial location of the puncture target. Finally, puncture parameters are calculated based on the triangular pyramid model, and the optical guidance device is controlled, making the puncture path easier to control and improving the practicality of localization.
[0054] Furthermore, this application performs curve fitting processing based on edge contour data to obtain a closed contour curve, thereby determining the center projection point within the enclosed area. The center projection point is then balanced with the first reference point to obtain the puncture target point. Subsequently, auxiliary lines are constructed from the second reference point to the puncture target point. Triangular facets are constructed between adjacent auxiliary lines. Finally, all triangular facets and the base formed by the auxiliary lines are combined to form a triangular pyramid model.
[0055] This application optimizes the tympanic membrane contour morphology through curve fitting and performs positional balancing processing in conjunction with the first reference point, which enables more accurate determination of the puncture target point. Furthermore, a triangular pyramid model is constructed through the orderly combination of triangular facets and auxiliary connecting lines, enabling the model to stably map the spatial relationship between the puncture target point and the body surface reference point, providing a reliable three-dimensional data foundation for subsequent accurate calculation of puncture parameters. Attached Figure Description
[0056] To more clearly illustrate the technical solutions of the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0057] Figure 1 A schematic flowchart of a tympanic membrane puncture localization method based on clinical images provided in an embodiment of this application;
[0058] Figure 2 A schematic diagram illustrating a specific implementation of a tympanic membrane puncture localization method based on clinical images, provided in this application embodiment;
[0059] Figure 3 A schematic diagram of a tympanic membrane puncture localization system based on clinical images provided in this application embodiment;
[0060] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation
[0061] Existing methods for tympanic membrane puncture localization based on clinical images mostly rely on doctors to directly observe images to determine the puncture location, or simply process images in a single dimension to determine the target area. However, this approach is not only easily affected by the doctor's personal experience and image clarity, but also makes it difficult to fully grasp the spatial structure and key anatomical features of the tympanic membrane, leading to deviation of the puncture target point and inaccurate control of the needle insertion angle or depth, which increases the operational risks and the difficulty of diagnosis and treatment.
[0062] To address this issue, this application provides a tympanic membrane puncture localization method based on clinical images. This method first acquires layered images of the target ear and clinical images of the tympanic membrane, extracting the edge shape and key anatomical locations of the tympanic membrane as internal references. Then, multiple markers are set on the surface of the concha as external references. A three-dimensional model is constructed using these internal and external references, allowing for precise calculation of the appropriate needle insertion angle and depth. Finally, an optical guidance device indicates the puncture path. This approach not only compensates for the lack of spatial reference in a single image but also reduces bias caused by subjective judgment, making puncture localization more accurate and the operation more evidence-based, effectively solving the problem of insufficient localization accuracy in existing methods.
[0063] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0064] The core of this application is to provide a method for tympanic membrane puncture localization based on clinical images, and a flowchart of one specific implementation is shown below. Figure 1 As shown, the method includes:
[0065] S101. Acquire layered images of the target ear and clinical images of the tympanic membrane.
[0066] In the above scheme, layered images refer to multiple consecutive images obtained by scanning the target ear layer by layer from the entrance of the external auditory canal to the surface of the tympanic membrane, containing detailed data of ear tissue at different depths; clinical images of the tympanic membrane refer to intuitive images that clearly present the surface morphology of the tympanic membrane and are taken by medical imaging equipment, used to observe the anatomical features of the tympanic membrane.
[0067] In this application example, firstly, a medical imaging device with depth-of-field adjustment function is used to start the layered scanning mode. The layered scanning mode scans the external auditory canal region of the target ear layer by layer from the outside to the inside, and a frame is captured at each preset distance. After continuously acquiring multiple frames of images at different depths, a layered image is formed. Then, the focal length of the device is adjusted so that the imaging focus is accurately aligned with the surface of the tympanic membrane, and an image that clearly presents the anatomical features of the tympanic membrane is captured to obtain a clinical image of the tympanic membrane. Then, the synchronous acquisition function of the device completes the acquisition of two types of images at the same time, providing basic image data for subsequent extraction of the edge contour of the tympanic membrane and identification of key anatomical locations.
[0068] In this embodiment of the invention, by simultaneously acquiring the depth layering information of the target ear and the intuitive morphological information of the tympanic membrane, comprehensive and high-quality basic image data can be provided, which provides a data foundation for subsequent extraction of the tympanic membrane edge contour and identification of key anatomical locations, thereby ensuring the orderly progress of the positioning process.
[0069] S102. Extract the edge contour data of the tympanic membrane from the layered image, and identify the light cone apex and umbilicus of the tympanic membrane from the clinical image as the first reference point.
[0070] In one specific implementation, step S102 includes:
[0071] Step 1021: Perform reflectance analysis on the clinical image to determine the reflectance region, locate the extreme point within the reflectance region, and obtain candidate vertex points.
[0072] Step 1022: Fit the edge contour data to obtain the axis of symmetry. Based on the axis of symmetry, search for depression features in the tympanic membrane surface region to obtain candidate regions. The tympanic membrane surface region is obtained based on clinical image recognition of the tympanic membrane.
[0073] Step 1023: Perform curvature extremum analysis within the candidate region to obtain candidate points for the umbilicus.
[0074] Step 1024: Based on preset spatial constraints, verify the vertex candidate points and the navel candidate points to obtain the first reference point.
[0075] In the above scheme, edge contour data refers to the line data describing the boundary of the tympanic membrane extracted from the layered image, used to clarify the spatial morphology of the tympanic membrane; the first reference point refers to the apex of the tympanic light cone and the umbilicus identified from the clinical image, used to provide internal anatomical reference for localization; the vertex candidate point refers to the point corresponding to the extreme value of light reflection intensity in the reflective area of the clinical image, used to screen the apex of the light cone; the axis of symmetry refers to the straight line through the center of the tympanic membrane obtained by fitting the edge contour data, used to locate the center of symmetry of the tympanic membrane; the candidate region refers to the region on the surface of the tympanic membrane that may contain the umbilicus, defined based on the axis of symmetry, used to narrow the search range for the umbilicus; the umbilicus candidate point refers to the point corresponding to the extreme value of curvature in the candidate region, used to screen the umbilicus of the tympanic membrane; the spatial constraint conditions of the tympanic membrane anatomical structure refer to the rules such as the relative position and distance between the apex of the light cone and the umbilicus set based on the anatomical features of the tympanic membrane, used to verify the accuracy of the candidate points.
[0076] In this application example, firstly, through step 1021, the clinical image is processed using a light reflection intensity analysis algorithm to determine the reflection area where the light reflection intensity is higher than the set standard, and the point with the highest light reflection intensity in the reflection area is found using extreme point localization technology to obtain vertex candidate points.
[0077] Next, in step 1022, the edge contour data of the tympanic membrane is extracted from the layered image using an edge extraction algorithm, and the edge contour data is fitted using a straight line fitting algorithm to obtain the axis of symmetry. Based on the axis of symmetry, a specific range is defined on the surface of the tympanic membrane, and a candidate region is obtained by searching within this range using a concave feature detection technique.
[0078] Then, in step 1023, the curvature value of each point in the candidate region is calculated using a curvature analysis algorithm to filter out the point with the largest curvature and obtain the candidate points for the navel.
[0079] Finally, in step 1024, based on the preset spatial constraints, the relative positions and distances of the vertex candidate points and the navel candidate points are verified to eliminate candidate points that do not meet the constraints and retain the points that meet the requirements as the first reference points.
[0080] In practical applications, in the ENT diagnosis and treatment scenarios of medical institutions, medical staff operate medical imaging equipment to acquire images of the patient's target ear. First, the depth D of the external auditory canal of the target ear is set to 12mm, the scanning step size d is 0.6mm, and the number of scanning frames is calculated according to the scanning frame number formula N=D / d, where D represents the depth of the external auditory canal, d represents the scanning step size, and N represents the number of scanning frames. The calculated N is 20 frames. Then, the medical imaging equipment completes the layered image acquisition according to the acquisition amount of 20 frames, and then switches to high-definition shooting mode to automatically focus on the surface of the tympanic membrane and capture 5 images. The 20 frames of data of this layered image will be used for subsequent extraction of the edge contour of the tympanic membrane. The medical imaging equipment has a built-in sharpness scoring mechanism. Specifically, each frame of image corresponds to a score value, and the sharpness threshold T is set to 85 points. Then, the image with a score value higher than T is selected as the clinical image of the tympanic membrane. This clinical image will be used for subsequent identification of the first reference point.
[0081] Next, medical staff processed the clinical and layered images of the patient's target ear. Specifically, they first used a light reflection intensity analysis algorithm to calculate the light reflection intensity I of each pixel in the clinical image and set an intensity threshold. =180, and filter out those with I greater than The region is designated as the reflection region, and the intensity extremes within this region are calculated using extreme point localization technology. =235, and the corresponding extreme points are used as vertex candidates; at the same time, edge contour data are extracted from the layered image, and based on the coordinates of feature points on the contour ( The linear fitting algorithm is used according to the formula for fitting the axis of symmetry. Fit the axis of symmetry, where, and The x and y coordinates of the contour feature points are represented by k, which represents the slope of the axis of symmetry (e.g., k = 0.12), and b represents the intercept of the axis of symmetry (e.g., b = 5.3). The axis of symmetry is then obtained. .
[0082] Subsequently, candidate regions were defined by extending 3 pixels to each side of this axis, and the curvature calculation formula was used. Calculate the curvature value of each point within the candidate region, where K represents the curvature value of the point, y' represents the first derivative of the curve at that point, and y'' represents the second derivative of the curve at that point. Select the point with the largest K as the umbilicus candidate point. Then, under the preset spatial constraint condition that the straight-line distance L between the vertex candidate point and the umbilicus candidate point is between 8 and 12 pixels, proceed according to the formula... Calculate the distance between the two, where, and Represents the x and y coordinates of the candidate vertex. and The x and y coordinates of the candidate points of the umbilicus are represented. The calculated value is L = 10.5 pixels. Since 10.5 is within the constraint of 8 to 12 pixels, these two points are determined as the first reference points. These first reference points will be used in subsequent steps to determine the puncture target points in combination with the edge contour data.
[0083] In this embodiment of the invention, by accurately extracting the contour data of the tympanic membrane edge and identifying a reliable first reference point, an accurate internal anatomical reference is provided for subsequently determining the puncture target point, thereby ensuring the accuracy of positioning.
[0084] S103. Set at least three non-collinear auxiliary positioning markers on the surface of the concha cavity as second reference points.
[0085] In the above scheme, the auxiliary positioning marker point refers to the identifiable mark set on the surface of the concha cavity to provide external spatial reference; the second reference point refers to at least three non-collinear auxiliary positioning marker points to construct external anchor points for spatial positioning.
[0086] In this application example, firstly, the surface of the concha cavity is cleaned, and a flat area that is not easily obscured is selected. Then, at least three marker points are set using medical identifiable marking material. Spatial position detection technology is used to determine whether each marker point is collinear. After confirming that each marker point is not collinear, these marker points are used as second reference points to provide external spatial reference for the subsequent connection with the puncture target point to construct a three-dimensional model.
[0087] In practical applications, in the ENT (Ear, Nose, and Throat) department of medical institutions, after cleaning the surface of the patient's concha, medical staff select three suitable locations and set auxiliary positioning markers, labeled P1, P2, and P3. The spatial coordinates of each point, P1(2, 3, 1), P2(5, 4, 2), and P3(3, 6, 4), are obtained using coordinate measuring tools. The result of checking if the determinant is equal to zero determines whether the three points are collinear. Here, x1 is the x-coordinate of P1, y1 is the y-coordinate of P1, z1 is the z-coordinate of P1, x2 is the x-coordinate of P2, y2 is the y-coordinate of P2, z2 is the z-coordinate of P2, x3 is the x-coordinate of P3, y3 is the y-coordinate of P3, and z3 is the z-coordinate of P3. Substituting the coordinate values, we first obtain the elements of the first row of the determinant: 5-2=3, 4-3=1, 2. -1=1, the elements of the second row are 3-2=1, 6-3=3, 4-1=3, and the determinant is calculated to be 3×(3×3-1×3)-1×(1×3-1×1)+1×(1×3-3×1)=3×6-1×2+1×0=16. The result of the determinant is not zero, which means that the three points are not collinear. Finally, these three marked points are determined as the second reference points. The second reference points will be used to connect with the puncture target point in the subsequent steps to construct the triangular pyramid model.
[0088] In this embodiment of the invention, establishing stable external spatial anchor points can fill the gap of lacking entity references in the simple image localization process, providing a reliable external reference for the subsequent construction of a three-dimensional model, thereby ensuring the stability of spatial localization.
[0089] S104. Based on the edge contour data and the first reference point, determine the puncture target point, connect the second reference point and the puncture target point, and construct a triangular pyramid model.
[0090] In one specific implementation, step S104 includes:
[0091] Step 1041: Perform curve fitting processing based on the edge contour data to obtain a closed contour curve, and determine the center projection point within the area enclosed by the closed contour curve.
[0092] Step 1041 may specifically include the following steps:
[0093] Step a1: Identify the curvature distribution characteristics of the tympanic membrane edge contour based on the edge contour data, and identify the contour stability area and contour abnormal area within the tympanic membrane edge contour based on the curvature distribution characteristics.
[0094] Step a2: Based on the stable contour region, construct a contour framework. Based on the contour framework, use an elastic constraint algorithm to reshape the abnormal contour region to obtain the target segment.
[0095] In an exemplary embodiment, step a2 may specifically include the following steps: based on the geometric constraints of the contour frame, multiple deformation control points are set in the contour anomaly area, and an elastic constraint algorithm is used to establish an elastic connection relationship between adjacent deformation control points; based on the tension distribution corresponding to the elastic connection relationship, deformation simulation processing is performed on the contour anomaly area to generate an initial segment; the initial segment is aligned with the contour frame boundary to obtain an intermediate segment; the intermediate segment is optimized to obtain a target segment that matches the curvature characteristics of the contour stable area.
[0096] Among them, deformation control points refer to key points set in the contour anomaly area to guide the reshaping direction of the anomaly segment; elastic connection relationship refers to the simulated elastic tension relationship established between adjacent deformation control points to control the deformation amplitude of the contour anomaly area; initial segment refers to the preliminary contour segment obtained after performing deformation simulation on the contour anomaly area based on the tension distribution of elastic connection relationship; intermediate segment refers to the transitional contour segment obtained after aligning the initial segment with the contour frame boundary; target segment refers to the final contour segment that matches the curvature characteristics of the contour stable area after curvature consistency optimization of the intermediate segment.
[0097] Furthermore, tension distribution refers to the magnitude and spatial distribution of tension generated at each connection point after establishing a simulated elastic connection between adjacent deformation control points in the contour anomaly area based on the elastic constraint algorithm. It reflects the deformation constraint intensity at different locations in the contour anomaly area. The tension distribution can be obtained by: first, setting multiple deformation control points in the contour anomaly area, establishing elastic connections between adjacent control points through the elastic constraint algorithm, and then, combining the geometric constraints of the contour frame, calculating the tension value of each connection point based on the difference between the actual distance and the standard distance between adjacent control points, the preset elastic coefficient, and other parameters, and finally integrating the tension values of all connection points to form the overall tension distribution.
[0098] Step a3: Combine the contour frame with the target segment to obtain a closed contour curve, and calculate the center projection point based on the geometric features of the closed contour curve.
[0099] Step 1042: Perform position balancing processing on the central projection point and the first reference point to obtain the puncture target point, and construct an auxiliary line from the second reference point to the puncture target point.
[0100] Step 1043: Construct triangular facets between adjacent auxiliary lines, and combine all the triangular facets with the base formed by the auxiliary lines to form a triangular pyramid model.
[0101] In the above scheme, the contour closure curve refers to the complete closed line formed by combining the stable segment of the tympanic membrane edge with the reshaped abnormal segment, which is used to accurately describe the overall shape of the tympanic membrane; the center projection point refers to the center point of the region calculated based on the geometric features of the contour closure curve, which is used to provide the core reference of the tympanic membrane region; the contour stable area refers to the segment in the tympanic membrane edge contour with a gentle curvature change and regular shape, which is used to construct the basic contour framework; the contour abnormal area refers to the segment in the tympanic membrane edge contour with abrupt curvature change and irregular shape, which needs to be reshaped and optimized; the contour framework refers to the basic structure constructed based on the contour stable area to define the overall shape of the tympanic membrane contour, which is used to ensure the accuracy of contour reshaping.
[0102] Position balancing refers to the process of comprehensively considering the spatial relationship between the central projection point and the first reference point to adjust and obtain the optimal position, which is used to determine the precise puncture target; auxiliary connecting lines refer to the straight line segments pointing from the second reference point to the puncture target point, which are used to construct the skeleton structure of the three-dimensional model; triangular facets refer to the triangular planes formed by two adjacent auxiliary connecting lines and the puncture target point, which are used to build the side of the triangular pyramid model; the triangular pyramid model refers to the three-dimensional geometric structure formed by the combination of all triangular facets and the base formed by the second reference point, which is used to map the spatial relationship between the puncture target point and the body surface reference point.
[0103] In this application example, the following three steps are performed through step 1041: First, based on the extracted tympanic membrane edge contour data, curvature analysis technology is used to identify its curvature distribution characteristics to distinguish between the contour stable area and the contour abnormal area; Second, a contour framework is built based on the contour stable area, multiple deformation control points are set in the contour abnormal area, and an elastic constraint algorithm is used to establish an elastic connection relationship between adjacent deformation control points. Based on the tension distribution corresponding to the connection relationship, the deformation of the contour abnormal area is simulated to obtain an initial segment. Then, the initial segment and the contour framework are aligned at the boundary to obtain an intermediate segment. The curvature consistency of the intermediate segment is then optimized to obtain a target segment that matches the curvature characteristics of the contour stable area, so as to complete the reshaping of the contour abnormal area; Third, the contour framework and the target segment are seamlessly combined to form a contour closed curve, and the center point of the region is calculated based on the geometric characteristics of the contour closed curve to obtain the center projection point.
[0104] In practical applications, within the ENT (Ear, Nose, and Throat) department of medical institutions, medical staff utilize the tympanic membrane edge contour data obtained from S102 and the first reference point, the light cone vertex (15, 20, 8), and the umbilicus (18, 22, 8). They then identify the stable and abnormal contour regions, and within the abnormal contour region, set three deformation control points: Q1 (12, 18, 8), Q2 (14, 16, 8), and Q3 (16, 17, 8). The formula used in the elastic constraint algorithm is then applied. Calculate the elastic tension between adjacent control points, where F is the elastic tension between adjacent deformation control points, k is the elastic coefficient with a value of 0.8, and ΔL is the difference between the actual distance and the standard distance between adjacent control points. First, calculate the actual distance between Q1 and Q2. Since the standard distance is set to 2, then ΔL = 2.83 - 2 = 0.83, F1 = 0.8 × 0.83 ≈ 0.66. The actual distance between Q2 and Q3 is then calculated as follows: Since the standard distance is set to 2, ΔL = 2.24 - 2 = 0.24 and F2 = 0.8 × 0.24 ≈ 0.19. Then, based on the tension distribution, the deformation simulation of the contour anomaly area is performed to obtain the initial segment. The initial segment is then aligned with the boundary of the contour frame to obtain the intermediate segment. Finally, the curvature consistency optimization of the intermediate segment is performed to obtain the target segment.
[0105] Next, the combined contour frame and the target segment form a closed contour curve, and the curve is then aligned according to the center projection point formula. Calculate the central projection point, where, As the central projection point, The x-axis coordinates of the feature points on the closed contour curve. Let be the y-axis coordinate of the feature point on the closed contour curve. Let be the z-axis coordinate of a feature point on the closed contour curve, and n be the number of feature points. For example, a value of 50 can be used. Then, the coordinates of 50 feature points are summed and divided by 50 to obtain the result. x=17, y=21, z=8; then, the position balance formula is applied. Position balancing is performed, where... Let α be the puncture target point, and α be the balancing weight, which is set to 0.6 here. First, calculate... and The midpoint coordinates are (16.5, 21, 8), and substituting them into the formula yields... Given x = 0.6 × 17 + 0.4 × 16.5 = 16.8, y = 0.6 × 21 + 0.4 × 21 = 21, and z = 0.6 × 8 + 0.4 × 8 = 8, the final second reference point is obtained. (2, 3, 1) (5, 4, 2) (3, 6, 4) are three auxiliary lines pointing to P. Three triangular facets are then constructed between these auxiliary lines. These three triangular facets, along with the base formed by P1, P2, and P3, form a triangular pyramid model. This is the puncture target point. The triangular pyramid model will be used in subsequent steps to calculate the needle insertion angle and puncture depth.
[0106] In this embodiment of the invention, a three-dimensional model that can accurately map the spatial relationship between the reference point on the body surface and the puncture target point is constructed by determining the puncture target point, providing a stable and reliable spatial data foundation for subsequent accurate calculation of puncture parameters.
[0107] S105. Based on the triangular pyramid model, the spatial geometric positioning method is used to calculate the needle insertion angle and puncture depth, so as to control the optical guidance device to indicate the puncture path and complete the positioning.
[0108] For example, in one specific implementation, step S105 includes:
[0109] Step 1051: Extract tissue boundary data from the layered image;
[0110] Step 1052: Based on the tissue boundary data, determine the needle insertion depth threshold in the vertical projection direction of the puncture target point;
[0111] Step 1053: Determine the tangential normal vector of the puncture target point with respect to the tympanic membrane, and based on the tangential normal vector and the normal vector of the entrance plane of the external auditory canal, calculate the puncture needle insertion direction through the spatial constraints of the triangular pyramid model.
[0112] Step 1053 may specifically include the following steps: performing surface fitting processing on the tympanic membrane surface sampling points in the neighborhood of the puncture target point to obtain the tangent plane of the tympanic membrane, and taking the normal vector perpendicular to the tangent plane as the first normal vector; performing plane fitting processing based on the spatial distribution of the second reference point to obtain the entrance plane of the external auditory canal, and taking the normal vector perpendicular to the entrance plane as the second normal vector; projecting the first normal vector onto the entrance plane of the external auditory canal to obtain the first projection component, and projecting the second normal vector onto the tangent plane to obtain the second projection component; performing vector synthesis on the first projection component and the second projection component to generate the initial needle insertion direction; using a coordinate calibration method, through the spatial constraints of the triangular pyramid model, optimizing the initial needle insertion direction to obtain the puncture needle insertion direction, wherein the angle between the puncture needle insertion direction and the tangent plane normal vector satisfies the preset perpendicularity condition, and the spatial relationship with the second normal vector satisfies the preset coplanarity requirement.
[0113] The puncture target point neighborhood refers to the spatial range defined centered on the puncture target point, based on the actual size of the tympanic membrane and the image resolution. This range must cover the effective area of the tympanic membrane surrounding the puncture target point to obtain sufficient curved surface sampling points to accurately fit the tympanic membrane tangential plane. For example, the determination method of the puncture target point neighborhood is as follows: first, a reasonable spatial radius is preset based on the clinically common tympanic membrane size and image pixel accuracy; then, with the three-dimensional coordinates of the puncture target point as the center, effective points located on the curved surface of the tympanic membrane are screened within this preset radius, and abnormal points that are detached from the tympanic membrane surface are removed, finally forming a puncture target point neighborhood containing multiple effective sampling points.
[0114] Step 1054: Calculate the average spatial distance from the second reference point to the puncture target point as the puncture depth, where the puncture depth is less than the needle insertion depth threshold.
[0115] In the above scheme, tissue boundary data refers to the boundary line data of deep tissue below the tympanic membrane extracted from layered images, which is used to determine the safe depth limit of puncture; the scission normal vector refers to the vector perpendicular to the scission plane of the tympanic membrane where the puncture target point is located, which is used to indicate the vertical direction of the tympanic membrane surface; the first normal vector refers to the normal vector perpendicular to the scission plane of the tympanic membrane, which is used to provide a basic reference for the puncture direction; the entrance plane of the external auditory canal refers to the plane that matches the contour of the entrance of the external auditory canal, which is obtained by fitting the spatial distribution based on the second reference point, and is used to define the starting plane of puncture;
[0116] The second normal vector is the normal vector perpendicular to the plane of the external auditory canal entrance, used to optimize the puncture direction in conjunction with the vertical direction of the tympanic membrane; the first projection component is the projection vector of the first normal vector onto the plane of the ear canal entrance, used to adapt to the spatial direction of the ear canal entrance; the second projection component is the projection vector of the second normal vector onto the tympanic membrane tangential plane, used to adapt to the spatial direction of the tympanic membrane surface; the initial needle insertion direction is the preliminary puncture direction obtained by synthesizing the vectors of the first and second projection components, used to provide the basic prototype of the puncture direction;
[0117] The puncture needle insertion direction refers to the final puncture direction that meets the preset perpendicularity and coplanarity requirements after coordinate calibration of the initial needle insertion direction, used to accurately guide the puncture path; the perpendicularity condition refers to the preset standard that the angle between the puncture needle insertion direction and the normal vector of the tympanic membrane tangential plane must meet, used to ensure that the puncture is perpendicular to the tympanic membrane surface; the coplanarity requirement refers to the preset spatial position standard that the puncture needle insertion direction and the second normal vector must meet, used to ensure that the puncture direction adapts to the shape of the ear canal entrance; the puncture depth refers to the average spatial distance from the second reference point to the puncture target point, used to determine the actual depth of the puncture operation; the needle insertion depth threshold refers to the maximum safe puncture depth that can be achieved based on tissue boundary data, used to avoid puncture damage to deep tissues.
[0118] In this application example, firstly, in step 1051, an edge extraction algorithm is used to extract the boundary line data of the deep tissue below the tympanic membrane from the layered image to obtain the tissue boundary data.
[0119] Next, in step 1052, based on the tissue boundary data, the maximum safe depth that can be reached by puncture is calculated in the vertical projection direction of the puncture target point to determine the needle insertion depth threshold.
[0120] Then, in step 1053, the sampling points of the tympanic membrane surface in the neighborhood of the puncture target point are processed by a surface fitting algorithm to obtain the tympanic membrane tangent plane. The vector perpendicular to this plane is taken as the first normal vector. At the same time, based on the spatial distribution of the second reference point, a plane fitting algorithm is used to obtain the entrance plane of the external auditory canal. The vector perpendicular to this plane is taken as the second normal vector. The first normal vector is projected onto the ear canal entrance plane to obtain the first projection component, and the second normal vector is projected onto the tympanic membrane tangent plane to obtain the second projection component. Then, the two projection components are vector synthesized to generate the initial needle insertion direction. The initial needle insertion direction is then optimized by a coordinate calibration algorithm to make the optimized direction meet the preset perpendicularity and coplanarity requirements, thereby obtaining the puncture needle insertion direction.
[0121] Finally, in step 1054, the spatial distance from each second reference point to the puncture target point is calculated using the spatial distance calculation formula, and the average of all distances is taken as the puncture depth. After verifying that the puncture depth is less than the needle insertion depth threshold, the final puncture depth is determined. Then, the puncture needle insertion direction and puncture depth parameters are transmitted to the optical guidance device, and the puncture path is indicated by the device's visualization guidance function to complete the positioning.
[0122] In practical applications, in the ENT diagnosis and treatment scenarios of medical institutions, firstly, tissue boundary data is extracted from the layered images, and the coordinates of the deep tissue boundary (16.8, 21, 1) in the vertical projection direction of the puncture target point are determined. Then, the needle insertion depth threshold formula is applied. Calculate the needle insertion depth threshold, where T is the needle insertion depth threshold. The z-coordinate of the puncture target point. Using the z-axis coordinates of the deep tissue boundary, the needle insertion depth threshold T = |8-11| = 3 was obtained through numerical calculation. Subsequently, five surface sampling points were selected in the neighborhood of the puncture target point, and the equation of the tympanic membrane tangential plane was obtained using a surface fitting algorithm. Therefore, the first normal vector =(0.1, 0.2, 1), the plane equation of the external auditory canal inlet is obtained based on the second reference point using a plane fitting algorithm. Therefore, the second normal vector =(0.2, 0.3, 0.5), then according to the projection formula Calculate the first projection component respectively Second projection component ,in, Let v be the projection vector and v be the normal vector. Let be the unit normal vector of the projection plane, then we can calculate... =(0.08, 0.17, 0.03), =(0.15, 0.25, 0), then according to the vector composition formula Generate the initial needle insertion direction, where, This is the initial needle insertion direction. The composite weights for the first projection component, The composite weights of the second projection components are all set to 0.5, and the calculation yields... =(0.115, 0.21, 0.015), and the puncture needle insertion direction is obtained after coordinate calibration and optimization. =(0.12, 0.21, 0.02).
[0123] Next, according to the spatial distance formula Calculate separately , , arrive The distance, where d is the spatial distance from a single second reference point to the puncture target point. The x-axis coordinates of the second reference point. The y-axis coordinate of the second reference point. The z-axis coordinate of the second reference point. The x-coordinate of the puncture target point. The y-coordinate of the puncture target point. Here is the z-axis coordinate of the puncture target point, for example. , , and according to the average distance formula Calculate the puncture depth, where, For the puncture depth, the calculation is as follows: If the value is 2.36 and it is verified that it is less than the needle insertion depth threshold of 3, the puncture needle direction and puncture depth are transmitted to the optical guidance device. The device uses a light beam to indicate the puncture path and completes the positioning.
[0124] In this embodiment of the invention, by accurately calculating puncture parameters that are safe and compatible with the shape of the ear canal and tympanic membrane, the optical guidance device intuitively presents the puncture path, which not only ensures the safety of the puncture operation, but also improves the accuracy and operability of the positioning.
[0125] The following is a complete embodiment for steps S101-S105, such as Figure 2As shown, in a routine ENT clinic setting, medical staff perform tympanic membrane puncture and localization for patients. First, they clean and pre-treat the target ear using a medical imaging device. Then, they activate the device's layered scanning mode, setting the external auditory canal scanning depth D to 10 mm and the scanning step size d to 0.5 mm. The number of scanning frames is calculated using the formula N = D / d, where N is the total number of layered images, D is the external auditory canal scanning depth, and d is the scanning step size. The calculated N is 20 frames. The device then scans and acquires 20 layered images from the external auditory canal entrance towards the tympanic membrane. After that, the device switches to high-definition shooting mode, automatically focuses on the tympanic membrane surface, and captures 3 images. The device's built-in clarity filtering mechanism selects the best frame as the clinical image of the tympanic membrane.
[0126] Next, the acquired layered images underwent slight noise reduction preprocessing. Then, an edge extraction algorithm was used to extract the tympanic membrane edge features frame by frame and integrate them to obtain complete tympanic membrane edge contour data. Simultaneously, light reflection intensity analysis was performed on the clinical images of the tympanic membrane to identify the apex of the light cone with the strongest light reflection and the umbilicus in the center of the tympanic membrane. The data was then analyzed according to the formula... Calculate the center coordinates of the first reference point, where, The center coordinates of the first reference point, and Find the x-coordinate and y-coordinate of the light cone vertex; for example, the coordinates of the light cone vertex are (15, 20). and Let x and y be the x and y coordinates of the navel, respectively. For example, the coordinates of the navel vertex are (18, 22). The center coordinates of the first reference point are then calculated. The value is (16.5, 21).
[0127] Then, a flat area of the patient's concha, not easily obscured by hands, and close to the entrance of the external auditory canal was selected. Three medical, non-toxic, and identifiable markers were affixed as auxiliary positioning points P1(2,3,1), P2(5,4,2), and P3(3,6,4). The three-dimensional spatial coordinates of each point were then obtained using a coordinate measuring tool. The result of whether the determinant value is equal to zero determines whether the three points are collinear, where (x1, y1, z1) are the spatial coordinates of P1, (x2, y2, z2) are the spatial coordinates of P2, and (x3, y3, z3) are the spatial coordinates of P3. If the calculated determinant result is not zero, it is confirmed that the three points are not collinear, and this is used as the second reference point.
[0128] Subsequently, based on the obtained tympanic membrane edge contour data, stable segments with regular shapes and slightly irregular abnormal segments in the contour were first identified. After simple correction of the abnormal segments, curve fitting was performed to obtain a smooth contour closure curve, according to the formula... Calculate the center projection point of the closed profile curve, where, The center projection point of the closed contour curve. , , Let be the coordinates of the feature points on the closed contour curve, and n be the number of feature points. For example, n=50, and the sum of the x-coordinates of the 50 feature points is 850, the sum of the y-coordinates is 1050, and the sum of the z-coordinates is 400. Substituting the values, we get... The value is (17, 21, 8), and the value is... Add z coordinate 8, and get (16.5, 21, 8), then follow the formula The puncture target point is obtained by performing position balancing processing, where, Let α be the puncture target point, and α be the balancing weight, which is set to 0.6 here. The calculation yields... The coordinates are (16.8, 21, 8); subsequently, the three second reference points obtained are connected to... In three-dimensional space, a triangular pyramid model is constructed that can map the positional relationship between the body surface and the puncture target.
[0129] Finally, based on this triangular pyramid model, the tympanic membrane region where the puncture target point is located is first sampled and fitted to obtain the tympanic membrane tangent plane and determine its normal vector. Then, combined with the normal vector of the external auditory canal entrance plane fitted by the second reference point, the needle insertion angle that adapts to the ear canal shape and tympanic membrane angle is calculated through vector projection and synthesis, and then the formula is applied. Calculate the puncture depth, where, d1, d2, and d3 represent the puncture depth, and d1, d2, and d3 represent the distances from each of the second reference points to the puncture depth. The spatial distance was determined, and the boundary data of the deep tissue below the tympanic membrane was extracted from the layered images. The needle insertion depth threshold T to avoid tissue damage was then determined and finally validated. Once the needle insertion angle and puncture depth parameters are less than T and meet clinical safety standards, they are transmitted to the optical guidance device. The device projects a puncture path guide onto the patient's ear surface using a precise beam of light. Medical staff can then determine the operation direction based on the beam indication and complete the entire tympanic membrane puncture positioning operation.
[0130] Figure 3 This is a schematic diagram of a specific embodiment of a tympanic membrane puncture localization system based on clinical images provided in this application. (Refer to...) Figure 3 The system may include:
[0131] The acquisition module 31 is used to acquire layered images of the target ear and clinical images of the tympanic membrane.
[0132] The recognition module 32 is used to extract the edge contour data of the tympanic membrane from the layered image and to identify the light cone apex and umbilicus of the tympanic membrane from the clinical image as the first reference point.
[0133] Setting module 33 is used to set at least three non-collinear auxiliary positioning markers on the surface of the concha cavity as second reference points.
[0134] The construction module 34 is used to determine the puncture target point based on the edge contour data and the first reference point, connect the second reference point and the puncture target point, and construct a triangular pyramid model.
[0135] The calculation module 35 is used to calculate the needle insertion angle and puncture depth based on the triangular pyramid model and the spatial geometric positioning method, so as to control the optical guidance device to indicate the puncture path and complete the positioning.
[0136] The tympanic membrane puncture positioning system based on clinical images in this application is used to implement the aforementioned tympanic membrane puncture positioning method based on clinical images. Therefore, the specific implementation of the tympanic membrane puncture positioning system based on clinical images can be found in the embodiment section of the tympanic membrane puncture positioning method based on clinical images above. The specific implementation can be referred to the description of the corresponding embodiments, and will not be repeated here.
[0137] like Figure 4 As shown, this application also provides an electronic device, including: a memory 41 for storing a computer program; and a processor 42 for executing the computer program to implement the steps of any of the above-described clinical image-based tympanic membrane puncture localization methods.
[0138] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of any of the above-described clinical image-based tympanic membrane puncture localization methods.
[0139] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, random access memory, portable hard drives, magnetic disks, or optical disks.
[0140] Embodiments of the present invention also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the embodiments of the tympanic membrane puncture localization method based on clinical images.
[0141] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.
[0142] The foregoing has provided a detailed description of a tympanic membrane puncture localization method and system based on clinical images provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.
Claims
1. A method for tympanic membrane puncture localization based on clinical images, characterized in that, include: Acquire layered images of the target ear and clinical images of the tympanic membrane; The edge contour data of the tympanic membrane is extracted from the layered image, and the light cone apex and umbilicus of the tympanic membrane are identified from the clinical image as the first reference point; At least three non-collinear auxiliary positioning markers are set on the surface of the concha cavity as second reference points; Based on the edge contour data and the first reference point, the puncture target point is determined, and the second reference point and the puncture target point are connected to construct a triangular pyramid model. Based on the triangular pyramid model, a spatial geometric positioning method is used to calculate the needle insertion angle and puncture depth, so as to control the optical guidance device to indicate the puncture path and complete the positioning. The step of determining the puncture target point based on the edge contour data and the first reference point, connecting the second reference point and the puncture target point, and constructing a triangular pyramid model includes: Based on the edge contour data, curve fitting is performed to obtain a closed contour curve, and the center projection point is determined within the area enclosed by the closed contour curve. The center projection point and the first reference point are balanced to obtain the puncture target point, and an auxiliary line is constructed from the second reference point to the puncture target point. Construct triangular facets between adjacent auxiliary lines, and combine all the triangular facets with the base formed by the auxiliary lines to form a triangular pyramid model.
2. The method according to claim 1, characterized in that, The step of performing curve fitting based on the edge contour data to obtain a closed contour curve, and determining the center projection point within the area enclosed by the closed contour curve, includes: Based on the edge contour data, the curvature distribution characteristics of the tympanic membrane edge contour are identified, and based on the curvature distribution characteristics, the contour stability area and the contour abnormal area within the tympanic membrane edge contour are identified. Based on the stable contour region, a contour framework is constructed. Based on the contour framework, an elastic constraint algorithm is used to reshape the abnormal contour region to obtain the target segment. The contour frame is combined with the target segment to obtain a closed contour curve. Based on the geometric features of the closed contour curve, the center projection point is calculated.
3. The method according to claim 2, characterized in that, Based on the contour framework, an elastic constraint algorithm is used to reshape the contour anomaly region to obtain the target segment, including: Based on the geometric constraints of the contour frame, multiple deformation control points are set in the contour anomaly area, and an elastic constraint algorithm is used to establish an elastic connection relationship between adjacent deformation control points. Based on the tension distribution corresponding to the elastic connection relationship, the abnormal contour area is subjected to deformation simulation processing to generate an initial segment; The initial segment is aligned with the outline frame to obtain the intermediate segment; The intermediate segment is optimized to obtain a target segment that matches the curvature characteristics of the contour stabilization region.
4. The method according to claim 1, characterized in that, The method for calculating the needle insertion angle and puncture depth based on the triangular pyramid model and using spatial geometric positioning includes: Extract tissue boundary data from the layered image; Based on the tissue boundary data, a needle insertion depth threshold is determined in the vertical projection direction of the puncture target point; Determine the normal vector of the puncture target point with respect to the tympanic membrane, and based on the normal vector of the puncture target point and the normal vector of the entrance plane of the external auditory canal, calculate the puncture needle insertion direction through the spatial constraints of the triangular pyramid model; The average spatial distance from the second reference point to the puncture target point is calculated as the puncture depth, which is less than the needle insertion depth threshold.
5. The method according to claim 4, characterized in that, The process of determining the slicing normal vector of the puncture target point with respect to the tympanic membrane, and calculating the puncture needle insertion direction based on the slicing normal vector and the normal vector of the entrance plane of the external auditory canal, using the spatial constraints of a triangular pyramid model, includes: The tympanic membrane surface sampling points in the neighborhood of the puncture target point are subjected to surface fitting processing to obtain the tangent plane of the tympanic membrane, and the normal vector perpendicular to the tangent plane is taken as the first normal vector. Plane fitting is performed based on the spatial distribution of the second reference point to obtain the entrance plane of the external auditory canal, and the normal vector perpendicular to the entrance plane is taken as the second normal vector. The first normal vector is projected onto the entrance plane of the external auditory canal to obtain the first projection component, and the second normal vector is projected onto the tangent plane to obtain the second projection component; The first projection component and the second projection component are vector-synthesized to generate the initial needle insertion direction; Using a coordinate calibration method, the initial needle insertion direction is optimized through the spatial constraints of a triangular pyramid model to obtain the puncture needle insertion direction. The angle between the puncture needle insertion direction and the normal vector of the tangent plane satisfies a preset perpendicularity condition, and the spatial relationship with the second normal vector satisfies a preset coplanarity requirement.
6. The method according to claim 1, characterized in that, The step of identifying the light cone apex of the tympanic membrane and the umbilicus from the clinical image as a first reference point includes: Reflectance analysis is performed on the clinical image to determine the reflection region, and extreme points are located within the reflection region to obtain candidate vertex points; The edge contour data is fitted to obtain the axis of symmetry. Based on the axis of symmetry, the depression features in the tympanic membrane surface region are searched to obtain candidate regions. Curvature extremum analysis is performed within the candidate region to obtain candidate points for the umbilicus; Based on preset spatial constraints, the vertex candidate points and the navel candidate points are verified to obtain the first reference point.
7. A tympanic membrane puncture localization system based on clinical images, characterized in that, To implement the tympanic membrane puncture localization method based on clinical images as described in claim 1, the method includes: The acquisition module is used to acquire layered images of the target ear and clinical images of the tympanic membrane; The recognition module is used to extract edge contour data of the tympanic membrane from the layered image and to identify the light cone apex and umbilicus of the tympanic membrane from the clinical image as a first reference point. The module is used to set at least three non-collinear auxiliary positioning markers on the surface of the concha cavity as second reference points; The construction module is used to determine the puncture target point based on the edge contour data and the first reference point, connect the second reference point and the puncture target point, and construct a triangular pyramid model; The calculation module is used to calculate the needle insertion angle and puncture depth based on the triangular pyramid model and spatial geometric positioning method, so as to control the optical guidance device to indicate the puncture path and complete the positioning.
8. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the steps of the tympanic membrane puncture localization method based on clinical images as described in any one of claims 1 to 6.
9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, enables the implementation of the tympanic membrane puncture localization method based on clinical images as described in any one of claims 1 to 6.