Artificial intelligence-based tricuspid valve annulus positioning and plane inclination angle automatic calculation method and system

Abstract

The invention provides a tricuspid valve annulus accurate positioning and plane inclination angle automatic calculation method based on artificial intelligence. The method comprises the steps that a three-dimensional medical image is acquired and preprocessed; key structures such as a right atrium, a tricuspid valve and a right ventricle are automatically segmented based on a deep learning model; extracting a valve leaflet junction point and a valve ring reference point, and constructing an initial valve ring curve; curve continuity and point distribution uniformity are improved through path optimization and resampling; using an iteration strategy to optimize valve ring point set space distribution so as to make the valve ring point set space distribution fit valve leaflet surface features; determining a valve ring plane by using a mathematical fitting method, calculating a normal vector, and further obtaining an included angle between the normal vector and the horizontal plane; according to the method, automatic identification and inclination angle quantification of the tricuspid valve annulus are realized, the accuracy and clinical applicability of morphological measurement are improved, and manual errors are reduced.

Description

Technical Field

[0001] The present invention relates to the technical fields of medical image processing and artificial intelligence (AI), and particularly relates to a method and system for automatically calculating the tricuspid annulus location and plane inclination angle based on an AI segmentation model. Technical Background

[0002] The tricuspid valve is located between the right atrium and the right ventricle, and its anatomical structure and functional status are crucial for the normal operation of the cardiovascular system. In clinical practice, tricuspid valve diseases (such as tricuspid regurgitation, tricuspid stenosis, and tricuspid valve prolapse) often require precise anatomical and functional evaluation by means of imaging techniques (such as echocardiography, CT, or MRI). However, traditional tricuspid valve image analysis methods rely on manual operations. Clinicians need to manually mark the tricuspid annulus position and estimate its plane inclination angle. However, this method is not only time-consuming and laborious but also easily affected by the operator's subjective experience, resulting in lower accuracy and consistency of the analysis results. In addition, due to the complex shape of the tricuspid valve and its large influence by the dynamic movement of the heart, the difficulty of manual analysis is further increased, thereby limiting the accuracy and repeatability of image analysis.

[0003] The plane inclination angle is a key indicator for describing the spatial location and geometric shape of the tricuspid annulus and has important application value in clinical practice. For example, in tricuspid annuloplasty, the plane inclination angle can help determine the optimal implantation position and angle of the annuloplasty ring to ensure good valve closure effect; during transcatheter tricuspid valve replacement (TTVR), the plane inclination angle helps select appropriate interventional devices and reduce the risk of offset and rotation during valve prosthesis implantation; in tricuspid valve repair, the plane inclination angle can be used as an important reference index for predicting the recovery of right heart function after surgery. However, most existing methods focus on the segmentation of valve anatomical structures and lack a systematic and automated solution for the spatial location of the annulus and the calculation of the plane inclination angle.

[0004] To address the above challenges, there is an urgent need to develop an artificial intelligence-based tricuspid valve image analysis technology to achieve automated, precise, and standardized measurement of valve morphology.

[0005] This paper proposes an innovative AI-based tricuspid valve image analysis method. By combining a deep learning segmentation model with a precise geometric measurement algorithm, it can automatically identify the tricuspid valve annulus and automatically derive the annular inclination angle through spatial geometric analysis. This method not only improves the efficiency and accuracy of image analysis but also provides clinicians with a more objective and accurate diagnostic basis, facilitating precise assessment and surgical planning of tricuspid valve disease. Summary of the Invention

[0006] The present invention combines artificial intelligence technology with image processing algorithms to provide an artificial intelligence-based method for accurate positioning of the tricuspid valve annulus and automatic calculation of plane inclination, including:

[0007] Step 1: Acquire 3D medical image data. Use imaging equipment such as CT, MRI, or echocardiography to collect patient cardiac images and perform preprocessing to improve data quality.

[0008] Step 2: Identify and segment the right atrium, tricuspid valve, and right ventricle. Based on a deep learning segmentation model, the cardiac anatomical structure is automatically extracted. Morphological operations and connected domain analysis are used to optimize the segmentation results and improve segmentation accuracy.

[0009] Step 3: Extract tricuspid valve anatomical key points. Specifically, based on the anatomical structure and spatial topological relationship of the tricuspid valve leaflets, extract the feature point set of the tricuspid valve leaflets, and automatically identify the leaflet intersection key points through distance measurement and anatomical position information;

[0010] Step 4: Construct the initial annulus. Based on the distribution of key points and the characteristics of the leaflet region, generate the annulus curve. Improve the continuity and uniformity of the annulus curve through path optimization and resampling.

[0011] Step 5: Iteratively optimize the annulus positioning. Using a positioning method based on the leaflet surface features, the spatial distribution of the annulus point set is adjusted. Smoothing and optimization strategies are then applied to ensure that the annulus morphology conforms to the leaflet anatomy.

[0012] Step 6: Determine the tricuspid valve annulus plane. Based on the optimized annulus point set, use a mathematical fitting method to determine the annulus plane and calculate its normal vector.

[0013] Step 7: Calculate the annular plane inclination.

[0014] Furthermore, in step 1, the preprocessing includes but is not limited to noise removal, normalization and registration of the image.

[0015] Furthermore, the key structure segmentation in step 2 includes, but is not limited to, the identification and segmentation of the right atrium, anterior leaflet of the tricuspid valve, posterior leaflet of the tricuspid valve, septal leaflet of the tricuspid valve, and right ventricle. The segmentation method can be manual annotation, traditional image segmentation algorithms, or an automatic segmentation model based on deep learning. The automatic segmentation model includes, but is not limited to, nnUNet and SwinUNETR, and morphological operations and connected component analysis are combined to optimize the analysis results to remove noise and artifacts and ensure the continuity and accuracy of the segmented regions.

[0016] Furthermore, in step 3, based on the anatomical structure characteristics, the tricuspid valve region is divided into the anterior leaflet, posterior leaflet, and septal leaflet, and the corresponding segmented result point sets are extracted respectively; then, the Euclidean distance between any two adjacent leaflet point sets is calculated, and when

[0017] d(p,q)<α

[0018] where α is a preset threshold, the point farthest from the right ventricle is selected from the candidate matching point sets to determine the junction key points, including the septal-anterior leaflet junction point P A 、the anterior-posterior leaflet junction point P B and the posterior-septal leaflet junction point P C .

[0019] Further, in step 4, an initial annulus is constructed. The specific method is as follows:

[0020] 4.1: Based on the segmentation results of the tricuspid valve, binary segmentation maps of the anterior leaflet, posterior leaflet, and septal leaflet are generated in sequence;

[0021] 4.2: According to the leaflet junction positions, key point pairs P A -P B 、P B -P C 、P C -P A are constructed in sequence as the initial reference points of the annulus;

[0022] 4.3: Inside the leaflet structure, for each pair of key points, a path P(t) (t ∈ [0,1]) is found. The skeletonization algorithm is combined with the dynamic programming algorithm, and a reverse weight based on distance transformation is introduced to obtain an approximate medial axis path, which is the central path connecting (P A -P B 、P B -P C 、P C -P A );

[0023] 4.4: Resample the central path, and for the anterior leaflet and posterior leaflet (P A -P B 、PB -P C ) Sample 10 points on the path for the septal leaflet (P C -P A ) Sample 20 points on the path;

[0024] 4.5: Connect all the resampled points in sequence to form a closed initial annulus curve, and obtain the initial annulus point set

[0025] Furthermore, in the said step 5, the specific method for optimizing the annulus positioning includes:

[0026] 5.1: Obtain the three-dimensional voxel data of the tricuspid valve segmentation area, and use the surface reconstruction algorithm to convert the leaflet area into a continuous triangular mesh or point cloud; calculate the unit normal vector of each vertex or face element on the leaflet surface, and ensure that the normal points to the atrial side to ensure the correctness of the subsequent stepping direction;

[0027] 5.2: For each point in the initial annulus point set R0 obtained in step 4 Use the nearest point projection or interpolation method to determine the nearest point of each point on the leaflet mesh surface Obtain The unit normal vector n at i as the direction of optimized stepping for this point;

[0028] 5.3: Adopt the iterative stepping update method. For each point The step size δ = 1mm, the maximum number of iterations k max For each point Perform the iterative update operation.

[0029] Furthermore, step 5.3 specifically includes the following steps:

[0030] 5.3.1: Calculate the nearest point of the current point on the leaflet surface to ensure that the path converges to the leaflet surface;

[0031] 5.3.2: Obtain the local normal

[0032] 5.3.3: Perform iterative update according to the formula ;

[0033] 5.3.4: Judge whether the new position is still within the leaflet structure. If it is within the leaflet, continue the iteration. If it exceeds the leaflet area or the mask value is lower than the set threshold, stop the iteration and record the current as the final positioning point;

[0034] 5.3.5 If the number of iterations reaches the maximum value k max , it is forced to stop to prevent abnormal situations from affecting the optimization process;

[0035] 5.4: After connecting all the final iteration points , an optimized tricuspid annulus curve is formed, and curve smoothing is performed on the annulus point set to eliminate the noise introduced by discrete iterations; at the same time, equidistant arc length resampling is performed on the smoothed curve to ensure uniform distribution of the annulus points.

[0036] Furthermore, in the step 6, the least squares method is used to fit the plane equation

[0037] Ax + By + Ca + D = 0

[0038] for the annulus point set obtained in step 5, and the normalized normal vector n = (A, B, C) is solved by using the singular value decomposition or covariance matrix method.

[0039] Furthermore, in the step 7, the angle between the fitted plane normal vector and the horizontal plane is calculated, and the calculation method is as follows:

[0040] θ = arccos(|C|)

[0041] The calculation result is converted to degree system to reflect the inclination of the annulus plane.

[0042] The present invention also provides an artificial intelligence-based tricuspid annulus positioning and inclination angle automatic calculation system, including:

[0043] A medical image data processing module, which is used to acquire and preprocess three-dimensional medical image data, perform noise removal, normalization and registration on CT, MRI or echocardiogram to improve data quality;

[0044] An anatomical structure segmentation module, which is used to automatically segment the right atrium, anterior leaflet of the tricuspid valve, posterior leaflet of the tricuspid valve, septal leaflet of the tricuspid valve and right ventricle based on a deep learning model, and optimize the segmentation result by combining morphological operations and connected component analysis;

[0045] A key point extraction module, which is used to extract leaflet feature points based on the leaflet anatomical structure, and identify the key points of leaflet junction through distance measurement and anatomical position information;

[0046] An annulus construction module, which is used to generate an annulus curve based on the key point distribution, and improve the continuity and uniformity of the annulus through path optimization and resampling processing;

[0047] An annulus optimization module, which is used to adjust the spatial distribution of the annulus point set based on the leaflet surface features, and adopt smoothing processing and optimization strategies to ensure that the annulus morphology conforms to the leaflet anatomical structure;

[0048] The annulus plane calculation module is used to fit the annulus plane based on the optimized annulus point set and calculate its normal vector;

[0049] The inclination angle calculation module is used to calculate the angle between the normal vector of the annulus plane and the horizontal plane to determine the spatial inclination of the annulus.

[0050] Technical effects

[0051] Through the combination of artificial intelligence and geometric modeling technology, the present invention realizes the automatic and precise analysis of the tricuspid annulus positioning and plane inclination angle, and its technical effects are as follows:

[0052] 1. High efficiency: The automated process shortens the time-consuming of traditional manual analysis to a few seconds, significantly improving the diagnostic efficiency.

[0053] 2. High precision: Based on the deep learning segmentation model and optimization algorithm, the spatial positioning accuracy at the sub-millimeter level is achieved, and the calculation error of the plane inclination angle is significantly lower than that of traditional methods.

[0054] 3. Consistency: Eliminating the influence of human factors ensures a high degree of consistency and repeatability of the analysis results, which is applicable to multi-center research and clinical applications.

[0055] 4. Adaptability to complex anatomy: Through iterative optimization and spatial geometric analysis, it adapts to the morphological differences of the tricuspid valves of different patients, generates personalized annulus models, and supports dynamic cardiac motion analysis.

[0056] 5. Clinical practicability: The output result (plane inclination angle) is directly applied to valvuloplasty, transcatheter replacement, and repair surgery planning, improving the surgical success rate and the ability to predict postoperative recovery.

[0057] In summary, the present invention significantly improves the diagnostic efficiency and treatment level of tricuspid valve-related diseases, provides reliable support for clinical decision-making, and has important application value. Description of the drawings

[0058] Figure 1 is a flowchart of an automatic calculation method for tricuspid annulus positioning and plane inclination angle based on artificial intelligence.

[0059] Figure 2 is a schematic diagram of the segmentation of the right atrium, tricuspid valve, and right ventricle described in step 2.

[0060] Figure 3 is a schematic diagram of extracting the anatomical key points of the tricuspid valve described in step 3.

[0061] Figure 4 is a schematic diagram of the initial annulus constructed in step 4.

[0062] Figure 5 is a schematic diagram of the final annulus after iterative optimization described in step 5.

[0063] Figure 6 It is a schematic diagram of the annulus iteration process described in step 5.

[0064] Figure 7 It is a schematic diagram of the determined tricuspid annulus plane described in step 6.

[0065] Figure 8 It is a schematic diagram of the finally obtained inclination angle of the tricuspid annulus plane described in step 7. Detailed implementation manners

[0066] The present invention will be further described below in conjunction with specific embodiments. It should be understood that the embodiments are only used to illustrate the present invention and not to limit the protection scope of the present invention. In addition, it should be understood that after reading the content disclosed in the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms also fall within the protection scope defined by the present invention.

[0067] The present invention combines artificial intelligence technology with image processing algorithms, and provides a method and system for accurately positioning the tricuspid annulus and automatically calculating the plane inclination angle based on artificial intelligence. Among them, the process of the method is as Figure 1 shown, and specifically includes the following steps:

[0068] Step 1: Use a high-precision medical imaging device (such as CT, MRI or echocardiogram, etc.) to scan the patient's heart to obtain its three-dimensional image data. In order to improve the accuracy of subsequent segmentation and calculation, preprocess the original data, including but not limited to:

[0069] Noise removal: Adopt a filtering algorithm (such as Gaussian filtering or median filtering) to reduce the random noise introduced in the imaging process, while retaining important anatomical boundary information.

[0070] Normalization: Map the image gray value to a unified range ([0,1]) to eliminate the differences brought by different devices or scanning parameters, and ensure the consistency of the model input.

[0071] Registration: For multi-modal image data (such as the combined analysis of CT and MRI), use rigid and non-rigid registration techniques to align images from different sources to make them spatially consistent.

[0072] Step 2: In response to the requirements of tricuspid image analysis, automatically identify and segment the tricuspid valve and its surrounding key anatomical structures that can form links, and focus on extracting the spatial position relationships of structures such as the right atrium, tricuspid valve, and right ventricle to provide a complete basis for annulus positioning. The segmentation result is as Figure 2As shown. In this step, a deep learning-based segmentation algorithm, such as nnUnet or SwinUNETR, can be adopted to automatically generate an accurate label map of each anatomical structure through the segmentation results of the deep learning model, and combined with morphological optimization and connected component analysis for certain smoothing processing to ensure the continuity and integrity of the boundaries of each structure.

[0073] Step 3: Extract the key anatomical points of the tricuspid valve. Specifically: Based on the segmentation results of Step 2, automatically identify the key anatomical points of the anterior, posterior, and septal leaflets of the tricuspid valve. The results are as Figure 3 shown. Specifically, it includes the following sub-steps:

[0074] Step 3.1: Leaflet division

[0075] According to the anatomical structure, the tricuspid valve area is subdivided into the anterior, posterior, and septal leaflets, and their spatial point sets: M1, M2, and M3 are extracted respectively.

[0076] Step 3.2: Neighborhood matching and threshold determination

[0077] Calculate the Euclidean distance between adjacent leaflet point sets

[0078] d(p,q) = ||p - q||

[0079] When a pair of points satisfies

[0080] d(p,q) < α

[0081] where α is a preset threshold, then it is considered that the pair of points p and q has an anatomical proximity relationship.

[0082] Step 3.3: Determine the junction key points

[0083] Among the candidate matching points that satisfy the proximity relationship, according to the principle of "far from the right ventricle", select the points with a farther distance from the right ventricle (obtained through segmentation) in each group. Respectively determine the tricuspid valve leaflet junction key points, including the septal-anterior leaflet junction point P A , the anterior-posterior leaflet junction point P B and the posterior-septal leaflet junction point P C .

[0084] Step 4: Construct the initial annulus. Specifically: Based on the key points extracted in Step 3, construct the initial shape of the tricuspid valve annulus. The results are as Figure 4 shown. Specifically, it includes the following sub-steps:

[0085] Step 4.1: Generate a binary segmentation map

[0086] Based on the segmentation results of the tricuspid valve, generate a binary image B(x, y, z), where:

[0087]

[0088] Furthermore, the anterior leaflet B ant (x, y, z), posterior leaflet B post (x, y, z) and septal leaflet B sep (x, y, z) binary segmentation maps can be obtained respectively.

[0089] Step 4.2: Construct key point pairs

[0090] To describe the anatomical structure of the tricuspid valve, the following three pairs of key points are defined:

[0091] Anterior leaflet to posterior leaflet path point pair: B ant-post =(P A -P B ), representing the starting and ending points of the anterior leaflet.

[0092] Posterior leaflet to septal leaflet path point pair: P post-sep =(P B -P C ), representing the starting and ending points of the posterior leaflet. <(

[0093] Septal leaflet to anterior leaflet path point pair: P sep-ant =(P C -P A ), representing the starting and ending points of the septal leaflet.

[0094] Step 4.3: Find the central line path

[0095] For each pair of key points, find an approximate central axis line path within the corresponding leaflet structure. The specific method is as follows:

[0096] Step 4.3.1: Skeletonize the binary segmentation map B(x, y, z) of each leaflet and extract its topological center line. The skeletonization algorithm retains the core path of the connected region through iterative erosion operations, generating a single-pixel-wide skeleton structure S(x, y, z).

[0097] Step 4.3.2: Calculate the distance transform field D(x, y, z) of the segmentation map, defined as follows:

[0098]

[0099] where (u, v, w) represents the coordinates of the background region. The distance transform field reflects the distance from each foreground pixel to the nearest background pixel.

[0100] Step 4.3.3: Combine the distance transform field D(x, y, z) and the skeletonization result S(x, y, z), and use the dynamic programming algorithm to search for the optimal path from the key point pair (P A -P B ). The objective function of the path is:

[0101]

[0102] where p i is the i-th point on the path, and n is the path length. By minimizing the objective function C(path), a smooth path close to the center of the valve leaflet can be obtained. Finally, a "central path" connecting the key point pairs is formed.

[0103] Step 4.4: Resample the path

[0104] To unify the representation forms of different valve leaflet paths and ensure the uniform distribution of path points, resample the obtained central path. Specifically, for the paths (P ant-post and P post-sep ) of the anterior leaflet and posterior leaflet, sample uniformly and select 10 points. For the path P sep-ant of the septal leaflet, sample uniformly and select 20 points. The sampling method is based on path arc length parameterization, and the arc length accumulation function is defined as:

[0105]

[0106] where q j is the j-th point on the path. According to the target number of points K, the total arc length L(m) is equally divided into K segments, and interpolation is performed within each segment to obtain new sampling points.

[0107] Step 4.5: Connect the three groups of resampled points in sequence to form a closed initial annulus curve, ensuring the uniformity of annulus point distribution and the continuity of the curve.

[0108] Step 5: Iteratively optimize the annulus positioning, specifically: based on the initial annulus obtained in Step 4, the final annulus is obtained through iterative optimization, and the result is as Figure 5 shown. Specifically, it includes the following sub-steps:

[0109] Step 5.1: Obtain the three-dimensional voxel data of the tricuspid valve segmentation area, and use a surface reconstruction algorithm (such as Marching Cubes or Poisson Surface Reconstruction) to convert the valve leaflet area into a continuous triangular mesh or point cloud. Calculate the unit normal vector of each vertex or face element on the valve leaflet surface, and the normal direction points to the atrial side to ensure the correctness of the subsequent stepping direction.

[0110] Step 5.2: For each point in the initial annulus obtained in Step 4 determine its nearest point on the valve leaflet mesh surface through the nearest point projection or interpolation method i and obtain the unit normal vector n i at this point, as the direction of the optimization step.

[0111] Step 5.3: Iterative step-by-step update, as Figure 6 shown, for each point set the step size δ = 1mm, the maximum number of iterations k max , for each point perform iterative optimization update, the specific steps are as follows:

[0112] Step 5.3.1: Calculate the nearest point of the current point on the leaflet surface to ensure that the path converges to the leaflet surface;

[0113] Step 5.3.2: Obtain the local normal to guide the optimization direction;

[0114] Step 5.3.3: Perform iterative update, by calculating the new point position;

[0115] Step 5.3.4: Judge whether the new position is still within the leaflet structure. If it is within the leaflet, continue the iteration; if it exceeds the leaflet area or the mask value is lower than the set threshold, force the iteration to stop and record the final positioning point:

[0116] Step 5.3.5: If the number of iterations reaches the maximum value k max , then force it to stop to prevent abnormal situations from affecting the optimization process.

[0117] Step 5.4: Connect all the final iterative points to form an optimized tricuspid annulus curve. Perform curve smoothing processing (such as cubic spline interpolation or moving average filtering) on the annulus point set to eliminate the noise introduced by discrete iteration. Perform equal arc length resampling on the smoothed curve to ensure uniform distribution of the annulus points.

[0118] Step 6: Determine the tricuspid annulus plane, specifically: Based on the optimized annulus point set obtained in Step 5, use the least squares method to fit the annulus plane to obtain the spatial orientation of the annulus, and the result is as Figure 7 shown.

[0119] The final annulus point set is:

[0120]

[0121] The plane equation fitted by the least squares method is:

[0122] Ax + By + Cz + D = 0

[0123] Constrained by the objective function:

[0124]

[0125] Using singular value decomposition (SVD), the eigenvector corresponding to the smallest eigenvalue is taken as the plane normal vector n = (A, B, C).

[0126] Step 7: Calculate the angle between the plane normal vector of the fitted plane in Step 6 and the horizontal plane to quantify the spatial inclination of the annulus plane, and the result is as Figure 8 shown. Specific calculation method: The horizontal plane is defined as z = 0, and its normal vector is n h = (0, 0, 1). The angle with the horizontal plane

[0127] θ = arccos(|n·n h |) = arccos(|C|)

[0128] Convert the calculation result to degree measure

[0129]

[0130] The obtained angle θ reflects the inclination degree of the annulus plane relative to the horizontal plane, providing key geometric parameters for preoperative evaluation and surgical planning.

[0131] The present invention also provides an artificial intelligence-based automatic calculation system for tricuspid annulus positioning and plane inclination angle, including:[[]]

[0132] A medical image data processing module, configured to acquire and preprocess three-dimensional medical image data, receive CT, MRI or echocardiogram (3DE) data, and perform noise removal, normalization and registration to improve data quality and image consistency;

[0133] An anatomical structure segmentation module, configured to automatically segment the right atrium, anterior leaflet of the tricuspid valve, posterior leaflet of the tricuspid valve, septal leaflet of the tricuspid valve and right ventricle based on a deep learning model, and optimize the segmentation result by combining morphological operations and connected component analysis to improve the segmentation accuracy and the continuity of the segmentation region;

[0134] A key point extraction module, configured to extract leaflet feature points based on the leaflet anatomical structure, and identify the key points of leaflet junction through distance measurement and anatomical position information to determine the starting point of annulus construction;

[0135] An annulus construction module, configured to generate an annulus curve based on the key point distribution, and improve the continuity and uniformity of the annulus curve through path optimization, skeletonization and resampling processing to ensure that the annulus conforms to anatomical characteristics;

[0136] An annulus optimization module, configured to adjust the spatial distribution of the annulus point set based on the leaflet surface features, and adopt iterative stepping, surface normal calculation and smoothing processing to ensure that the annulus shape conforms to the leaflet anatomical structure and improve the model stability;

[0137] An annulus plane calculation module, which is used to determine the annulus plane based on the optimized annulus point set by using the least squares fitting, singular value decomposition (SVD) or covariance matrix method, and calculate its normal vector to obtain the spatial orientation of the annulus;

[0138] An inclination angle calculation module, which is used to calculate the angle between the normal vector of the annulus plane and the horizontal plane, determine the spatial inclination of the annulus, and provide support for preoperative planning and surgical plan optimization.

[0139] This system can automatically execute the complete process from medical image input to annulus inclination angle calculation, reduce manual operations, improve the accuracy and stability of annulus shape analysis, and provide accurate imaging support for tricuspid valve plasty, valve repair, transcatheter valve replacement, etc.

Claims

1. An artificial intelligence-based method for precise positioning of the tricuspid annulus and automatic calculation of the plane inclination angle, characterized in that The following steps are involved: Step 1: Acquire 3D medical image data. Use CT, MRI, or echocardiography equipment to capture the patient's heart image and perform preprocessing to improve data quality. Step 2: Identify and segment the right atrium, tricuspid valve, and right ventricle. Based on a deep learning segmentation model, the cardiac anatomical structure is automatically extracted. Morphological operations and connected domain analysis are used to optimize the segmentation results and improve segmentation accuracy. Step 3: Extract tricuspid valve anatomical key points. Specifically, based on the anatomical structure and spatial topological relationship of the tricuspid valve leaflets, extract the feature point set of the tricuspid valve leaflets, and automatically identify the leaflet intersection key points through distance measurement and anatomical position information; Step 4: Construct the initial annulus. Based on the distribution of key points and the characteristics of the leaflet region, generate the annulus curve. Improve the continuity and uniformity of the annulus curve through path optimization and resampling. Step 5: Iteratively optimize the annulus positioning. Using a positioning method based on the leaflet surface features, the spatial distribution of the annulus point set is adjusted. Smoothing and optimization strategies are then applied to ensure that the annulus morphology conforms to the leaflet anatomy. Step 6: Determine the tricuspid valve annulus plane. Based on the optimized annulus point set, use a mathematical fitting method to determine the annulus plane and calculate its normal vector. Step 7: Calculate the annular plane inclination.

2. According to the method for precise positioning of the tricuspid valve annulus and automatic calculation of plane inclination angles according to claim 1, in step 1, the preprocessing includes but is not limited to noise removal, normalization and alignment of the image.

3. According to the method for precise positioning of the tricuspid valve annulus and automatic calculation of plane inclination angles according to claim 2, the key structure segmentation in step 2 includes but is not limited to the identification and segmentation of the right atrium, the anterior tricuspid valve leaflet, the posterior tricuspid valve leaflet, the tricuspid valve septum leaflet and the right ventricle. The segmentation method can adopt manual annotation, traditional image segmentation algorithm or automatic segmentation model based on deep learning. The automatic segmentation model includes but is not limited to nnUNet and SwinUNETR, and combines morphological operations and connected domain analysis to optimize the analysis results to remove noise and artifacts and ensure the continuity and accuracy of the segmented area.

4. The method for precise positioning of the tricuspid valve annulus and automatic calculation of plane inclination according to claim 3, wherein in step 3, the tricuspid valve region is divided into the anterior leaflet, the posterior leaflet, and the septal leaflet based on anatomical structural features, and corresponding segmentation result point sets are extracted respectively; then, the Euclidean distance between any two adjacent leaflet point sets is calculated, and if the following conditions are met: d(p,q)<α where α is a preset threshold, select the point farthest from the right ventricle from the candidate matching point set, and determine the handover key points respectively, including the septal valve - anterior valve handover point P A , the anterior valve - posterior valve handover point P B and the posterior valve - septal valve handover point P C .

5. The method for accurate positioning of the tricuspid valve annulus and automatic calculation of the plane inclination angle according to any one of claims 1 to 4, wherein the initial annulus is constructed in step 4 by: 4.1: Based on the segmentation results of the tricuspid valve, binary segmentation maps of the anterior leaflet, posterior leaflet, and septal leaflet are generated in sequence; 4.2: According to the positions of the leaflet commissures, key point pairs P A -P B 、P B -P C 、P C -P A are constructed in sequence as the initial reference points of the annulus; 4.3: Inside the leaflet structure, find a path P(t) (t ∈ [0, 1]) for each pair of key points. Using a skeletonization algorithm combined with a dynamic programming algorithm and introducing a reverse weight based on distance transformation, an approximate medial axis path is obtained, and this path is the central path connecting (P A -P B 、P B -P C 、P C -P A ); 4.4: Resample the central path, sample 10 points for the front and rear lobe (P A -P B 、P B -P C ) paths, and sample 20 points for the alternate lobe (P C -P A ) path; 4.5: Connect all resampled points in sequence to form a closed initial annulus curve, obtaining an initial annulus point set 6. The method for accurate positioning of the tricuspid valve annulus and automatic calculation of plane inclination according to any one of claims 1 to 5, wherein in step 5, the specific method for optimizing the annulus positioning comprises: 5.1: Obtain 3D voxel data of the tricuspid valve segmentation area and use a surface reconstruction algorithm to convert the leaflet area into a continuous triangular mesh or point cloud; Calculate the unit normal vector of each vertex or surface element on the leaflet surface, and ensure that the normal points to the atrial side to guarantee the correctness of the subsequent stepping direction; 5.2: For each point in the initial annulus point set R0 obtained in step 4 Using the nearest point projection or interpolation method, determine the nearest point of each point on the leaflet mesh surface Obtain The unit normal vector n at i , as the direction of the optimized step of this point; 5.3: Adopt the iterative step-by-step update method. For each point The step size δ = 1mm, and the maximum number of iterations k max , for each point Perform iterative update operations; 5.4: Connect all the final iteration points to form an optimized tricuspid annulus curve, and perform curve smoothing on the annulus point set to eliminate the noise introduced by discrete iterations; meanwhile, perform equi-arc length resampling on the smoothed curve to ensure uniform distribution of the annulus points.

7. The method for accurate positioning of the tricuspid annulus and automatic calculation of the plane inclination angle according to claim 6, step 5.3 specifically includes the following steps: 5.3.1: Calculate the current point The nearest point on the leaflet surface To ensure that the path converges to the leaflet surface; 5.3.2: Obtaining the local normal 5.3.3: Iteratively update according to the formula ; 5.3.4: Determine the new position Check whether it is still within the leaflet structure. If it is within the leaflet, continue the iteration. If it exceeds the leaflet area or the mask value is lower than the set threshold, stop the iteration and record the current as the final positioning point; 5.3.5 If the number of iterations reaches the maximum value k max , then force to stop to prevent abnormal situations from affecting the optimization process.

8. The method for accurate positioning of the tricuspid annulus and automatic calculation of the plane inclination angle according to any one of claims 1-7, in the step 6, use the least squares method to fit the plane equation of the annulus point set obtained in step 5 Ax + By + Cz + D = 0 and solve the normalized normal vector n = (A, B, C) by using the singular value decomposition or covariance matrix method.

9. The method for accurate positioning of the tricuspid annulus and automatic calculation of the plane inclination angle according to any one of claims 1-8, in the step 7, calculate the angle between the normal vector of the fitting plane and the horizontal plane, and its calculation method is as follows: θ = arccos(|C|) Convert the calculation result to degrees to reflect the inclination of the annulus plane.

10. A system for implementing the method for accurately positioning the tricuspid annulus and automatically calculating the plane inclination angle according to any one of claims 1-9, characterized in that, Including: A medical image data processing module for acquiring and preprocessing three-dimensional medical image data, performing noise removal, normalization, and registration on CT, MRI, or echocardiogram to improve data quality; An anatomical structure segmentation module for automatically segmenting the right atrium, anterior leaflet of the tricuspid valve, posterior leaflet of the tricuspid valve, septal leaflet of the tricuspid valve, and right ventricle based on a deep learning model, and optimizing the segmentation result by combining morphological operations and connected component analysis; A key point extraction module for extracting leaflet feature points based on the leaflet anatomical structure, and identifying the key points of leaflet junction through distance measurement and anatomical position information; An annulus construction module for generating an annulus curve based on the key point distribution, and improving the continuity and uniformity of the annulus through path optimization and resampling processing; An annulus optimization module for adjusting the spatial distribution of the annulus point set based on the leaflet surface features, and using smoothing processing and optimization strategies to ensure that the annulus morphology conforms to the leaflet anatomical structure; An annulus plane calculation module for fitting the annulus plane based on the optimized annulus point set and calculating its normal vector; An inclination calculation module for calculating the angle between the normal vector of the annulus plane and the horizontal plane to determine the spatial inclination of the annulus.

Images