Precise positioning method for hysteromyoma lesion in combination with image analysis and data fusion
By combining the method of fusion of image analysis and data, the problem that traditional image analysis methods are difficult to fully reveal the complexity of uterine fibroid lesions is solved, and a higher accuracy of lesion positioning and personalized treatment plans are achieved, improving the effectiveness and safety of ablation treatment.
Patent Information
- Application Number
- CN202510233500.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-28
- Publication Date
- 2025-06-20
AI Technical Summary
Traditional imaging analysis methods rely on single modal images, making it difficult to fully reveal the complexity of uterine fibroid lesions, affecting the accuracy and safety of treatment.
Combining the method of image analysis and data fusion, real-time navigation paths and ablation energy suggestions are generated by acquiring multimodal uterine image data, performing artifact suppression and spatial registration, performing three-dimensional reconstruction and lesion feature segmentation, performing hemodynamic analysis and biomechanical characteristic distribution map construction.
It significantly improves the accuracy of lesion localization, optimizes the treatment plan, improves the efficiency and success rate of ablation treatment, and reduces damage to surrounding healthy tissues.
Smart Images

Figure CN120182193A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of lesion localization, and particularly to a precise localization method for uterine fibroids lesions combining image analysis and data fusion. Background Art
[0002] Uterine fibroids are common benign tumors in women, and their occurrence and development have a significant impact on the health and quality of life of patients. Traditional treatment methods include drug treatment, surgical resection, etc., but these methods have certain invasiveness and recovery periods. In recent years, image-guided ablation therapy has become an emerging technology for non-invasive treatment of uterine fibroids; however, the accuracy of ablation therapy depends on the accurate localization of lesions and a comprehensive understanding of tissue characteristics. Traditional image analysis methods rely on single-modal images and may not fully reveal the complexity of lesions.
[0003] The precise localization method for uterine fibroids lesions combining image analysis and data fusion provides doctors with a more comprehensive lesion feature map by comprehensively utilizing different types of medical image data and biomechanical analysis. These technologies can not only accurately locate the position and size of uterine fibroids, but also further improve the accuracy of the treatment path by real-time analyzing the physical characteristics of the lesion area, such as hardness and blood flow characteristics. Through three-dimensional reconstruction and fusion analysis, a more accurate navigation path can be provided for ablation therapy, and ablation energy can be recommended according to the specific tissue characteristics of the lesion area, so as to optimize the treatment effect and reduce the impact on surrounding healthy tissues. Summary of the Invention
[0004] Based on the above purposes, the present invention provides a precise localization method for uterine fibroids lesions combining image analysis and data fusion.
[0005] The precise localization method for uterine fibroids lesions combining image analysis and data fusion includes the following steps:
[0006] S1. Obtain multi-modal uterine image data, suppress artifacts and perform spatial registration on the original images, and output a preprocessed multi-modal image data set;
[0007] S2. Perform three-dimensional reconstruction on the preprocessed multi-modal image data set to generate a three-dimensional uterine model including anatomical structure features;
[0008] S3. Implement lesion feature segmentation in the three-dimensional uterine model, extract the morphological parameters and texture features of the fibroid lesions, and output the lesion segmentation result;
[0009] S4. Perform hemodynamic analysis based on the lesion segmentation result to obtain the blood perfusion parameters and vascular distribution characteristics of the lesion area;
[0010] S5, fuse blood perfusion parameters with vascular distribution characteristics, and combine with a three-dimensional uterine model to construct a distribution map of the biomechanical characteristics of the lesion;
[0011] S6, generate a real-time navigation path based on the distribution map of biomechanical characteristics, and output a positioning report including the recommended ablation energy value.
[0012] Further, the S1 includes:
[0013] S11, obtain multi-modal uterine image data: obtain multi-modal data of the uterus through a variety of imaging devices, and the multi-modal data includes magnetic resonance imaging, ultrasound imaging, and computed tomography;
[0014] S12, artifact suppression of the original image data: perform wavelet transform artifact suppression processing on the obtained original image data to remove imaging interference;
[0015] S13, spatial registration: perform spatial registration on the multi-modal image data to align the spatial coordinate systems of each modal data, and use the mutual information maximization algorithm for registration;
[0016] S14, output the preprocessed multi-modal image data set: output the multi-modal image data set after artifact suppression and spatial registration processing.
[0017] Further, the S2 includes:
[0018] S21, input the preprocessed multi-modal image data set: receive the output data set from the multi-modal image data preprocessing step;
[0019] S22, use the voxel-based three-dimensional reconstruction method: use the voxel representation method to construct a preliminary three-dimensional uterine model;
[0020] S23, fuse multi-modal image information: perform weighted fusion on the multi-modal image data, and combine the data from different modalities in three-dimensional space;
[0021] S24, generate anatomical structure features: use the threshold segmentation method to extract the anatomical structure features of the uterus from the fused three-dimensional image data;
[0022] S25, three-dimensional model generation and visualization: based on the segmentation result and the fused image data, generate the final three-dimensional uterine model through three-dimensional reconstruction technology, and visualize the uterine model using the surface reconstruction method;
[0023] S26, output the three-dimensional uterine model: output the generated three-dimensional uterine model.
[0024] Further, the S3 includes:
[0025] S31. Perform lesion feature segmentation in the three-dimensional uterine model: Use the region growing algorithm to segment the lesions in the three-dimensional uterine model;
[0026] S32. Extract morphological parameters from the segmented lesion area to quantify the size characteristics of the fibroids;
[0027] S33. Extract texture features from the lesion area: Use the gray-level co-occurrence matrix texture analysis method to extract the texture features of the lesions;
[0028] S34. Output the lesion segmentation area, morphological parameters, and texture features.
[0029] Further, the S4 includes:
[0030] S41. Perform hemodynamic analysis based on the lesion segmentation result: Use the lesion segmentation result to calculate the blood perfusion parameters of the lesion area through a hemodynamic model;
[0031] S42. Extract the vascular distribution characteristics of the lesion area based on the hemodynamic analysis result of the lesion area;
[0032] S43. Output the blood perfusion parameters and vascular distribution characteristics.
[0033] Further, the S41 uses an image-based perfusion model to calculate the blood irrigation parameters.
[0034] Further, the S42 includes:
[0035] S421. Vascular distribution feature extraction: Use an edge detection algorithm to detect the boundaries of blood vessels in the image;
[0036] S422. Calculation of blood vessel diameter and branch characteristics: Calculate the blood vessel diameter according to the voxel density and geometric morphology of the blood vessel area.
[0037] Further, the S5 includes:
[0038] S51. Fuse the blood perfusion parameters and vascular distribution characteristics: Fuse the calculated blood perfusion parameters and vascular distribution characteristics to generate a comprehensive feature dataset;
[0039] S52. Combine with the three-dimensional uterine model to construct a distribution map of the biomechanical characteristics of the lesions.
[0040] Further, the S6 includes:
[0041] S61. Generate a navigation path according to the biomechanical characteristics distribution map: Generate an optimal ablation navigation path according to the biomechanical characteristics of the lesion area and the identification of the target area;
[0042] S62. Calculate the ablation energy and output a positioning report: Calculate the recommended value of the ablation energy for the lesion area according to the hardness and volume of the lesion area.
[0043] S63. Output a positioning report including the ablation energy and the navigation path.
[0044] Advantages of the present invention:
[0045] The present invention provides a precise positioning method for uterine fibroids by combining image analysis and data fusion, which can significantly improve the accuracy of lesion positioning. Traditional imaging techniques often rely on single-modal images and are difficult to comprehensively reflect the complex characteristics of the lesion area. However, through the fusion of multi-modal imaging data and the combination of biomechanical characteristics (such as elastic modulus, blood perfusion, etc.), a more detailed lesion feature map can be provided, thereby accurately identifying the size, location, and shape of uterine fibroids. This precise positioning ability not only improves the accuracy of diagnosis but also provides more reliable data support for subsequent treatment. Especially in image-guided ablation therapy, the lesion can be more accurately located, avoiding damage to surrounding healthy tissues.
[0046] The present invention further optimizes the treatment plan by combining the ablation path and energy recommendation generated by real-time biomechanical analysis. By formulating personalized ablation energy recommendations based on the biomechanical characteristics of the lesion area (such as tissue hardness, blood flow, etc.), the efficiency and success rate of ablation therapy can be effectively improved. Through this technology, doctors can monitor the treatment progress in real time and dynamically adjust the ablation energy to ensure the efficiency and safety of the treatment. Description of the Drawings
[0047] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only those of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.
[0048] Figure 1 It is a flowchart of the method according to an embodiment of the present invention;
[0049] Figure 2 It is a data processing flowchart according to an embodiment of the present invention. Detailed Embodiments
[0050] To make the objectives, technical solutions, and advantages of the present invention more clear and understandable, the following further details the present invention with specific embodiments.
[0051] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should have the ordinary meanings understood by those with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar terms used in the present invention do not denote any order, quantity or importance, but are only used to distinguish different components. The terms such as "including" or "comprising" mean that the elements or objects appearing before this term cover the elements or objects listed after this term and their equivalents, without excluding other elements or objects. The terms such as "connected" or "coupled" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The terms such as "upper", "lower", "left" and "right" are only used to represent relative positional relationships, and when the absolute position of the object being described changes, the relative positional relationship may also change accordingly.
[0052] As Figure 1 - Figure 2 shown, a precise positioning method for uterine fibroids lesions combining image analysis and data fusion includes the following steps:
[0053] S1. Obtain multi-modal uterine image data, perform artifact suppression and spatial registration on the original images, and output a preprocessed multi-modal image dataset;
[0054] S2. Perform three-dimensional reconstruction on the preprocessed multi-modal image dataset to generate a three-dimensional uterine model including anatomical structure features;
[0055] S3. Implement lesion feature segmentation in the three-dimensional uterine model, extract the morphological parameters and texture features of the fibroid lesions, and output the lesion segmentation result;
[0056] S4. Perform hemodynamic analysis based on the lesion segmentation result to obtain the blood perfusion parameters and vascular distribution features of the lesion area;
[0057] S5. Integrate the blood perfusion parameters and vascular distribution features, and combine with the three-dimensional uterine model to construct a distribution map of the biomechanical characteristics of the lesions;
[0058] S6. Generate a real-time navigation path according to the distribution map of the biomechanical characteristics, and output a positioning report including the recommended ablation energy value.
[0059] S1 includes:
[0060] S11. Obtain multi-modal uterine image data: Obtain multi-modal data of the uterus through multiple imaging devices, and the multi-modal data includes magnetic resonance imaging, ultrasound imaging, and computed tomography;
[0061] S12. Artifact suppression of the original image data: Perform wavelet transform artifact suppression processing on the obtained original image data to remove imaging interference. The specific steps are as follows:
[0062] (1) Perform a two-dimensional wavelet transform on the original image I m (x, y) to obtain multi-scale decomposition coefficients, expressed as: where k ∈ {1, 2,..., N}, W k (x, y) represents the decomposition coefficient at scale k;
[0063] (2) Apply threshold denoising to each scale coefficient to remove artifacts, expressed as:
[0064] W k ′(x, y) = threshold(W k (x, y), T k ), where T k is the threshold at scale k for denoising, threshold(·) is the threshold function, which is the hard threshold or soft threshold method;
[0065] (3) Synthesize the denoised coefficients back into the original image through inverse wavelet transform, expressed as:
[0066] where I m ′(x, y) is the denoised image;
[0067] S13, Spatial registration: Perform spatial registration on multi-modal image data to align the spatial coordinate systems of each modal data. Use the mutual information maximization algorithm for registration. The specific steps of the mutual information registration algorithm are as follows:
[0068] (1) Define mutual information as a similarity measure between two images, expressed as:
[0069]
[0070] where, The joint probability distribution of images I m and I n at pixel position (i, j), is the marginal probability distribution of pixel value i in image I m , is the marginal probability distribution of pixel value j in image I n , N m , N n are the total number of pixels in images I m and I n respectively;
[0071] (2) Maximize the mutual information value through an optimization algorithm to calculate the transformation parameter T mn , and achieve spatial registration from modality m to modality n, expressed as:
[0072] Among them, T mn is a transformation matrix used to align the image of modality m to the coordinate system of modality n;
[0073] (3) The calculated registration transformation T mn acts on all image data to complete the spatial alignment of each modality dataset, expressed as:
[0074] Among them, is the transformed image data;
[0075] S14, output the preprocessed multi-modal image dataset: Output the multi-modal image dataset after artifact suppression and spatial registration. The dataset D f has a unified spatial coordinate system and provides reliable image input for subsequent analysis, expressed as:
[0076] D f ={I1′, I2′,..., I M ′}, where D f is the preprocessed multi-modal image dataset, M is the number of modality types, and I m ′ is the m-th modality image dataset after artifact suppression and spatial registration, expressed as the processed image matrix.
[0077] S2 includes:
[0078] S21, input the preprocessed multi-modal image dataset: Receive the output dataset D f ={I1′, I2′,..., I M ′} from the multi-modal image data preprocessing step, where I m ′ represents the m-th modality image data with spatial coordinates (x, y, z), indicating the spatial coordinates of each pixel or voxel. After artifact suppression and spatial registration, it has a unified spatial coordinate system;
[0079] S22, use the voxel-based three-dimensional reconstruction method: Use the voxel representation method to map each image slice into the three-dimensional space and use the voxel value to represent the intensity or characteristics of the tissue structure to construct a preliminary three-dimensional uterine model. The specific steps are as follows:
[0080] (1) Convert the preprocessed multi-modal image data into a voxel grid: Each image slice (I m ′) represents two-dimensional data. By mapping each slice to different levels in the three-dimensional space, a three-dimensional voxel grid is formed. Each voxel represents a small area in the image and stores the intensity value (such as signal intensity) of this area. The value of each voxel is determined by the pixel intensity in the image, forming a three-dimensional voxel grid:
[0081] V = {V i,j,k | i = 1…N x , j = 1…N y , k = 1…N z};
[0082] Among them, V i,j,k is a voxel in three-dimensional space, representing the intensity value at the position (i, j, k);
[0083] (2) Three-dimensional surface reconstruction: Use the surface reconstruction algorithm (Marching Cubes) to generate a three-dimensional uterine model from the voxel grid: Based on the results of the voxel grid, use the three-dimensional surface reconstruction algorithm (Marching Cubes algorithm) to extract the outer surface of the uterus, convert the voxel grid into a visual triangular grid. The Marching Cubes algorithm scans the voxel grid and connects adjacent voxels that meet certain conditions to form triangles, thereby reconstructing the outer surface of the uterus, expressed as:
[0084] Vertices, Triangles = MarchingCubes(V), where the generated three-dimensional model consists of a vertex set Vertices and a triangular facet set Triangles, representing the outer surface of the uterus;
[0085] S23, fusing multi-modal image information: Perform weighted fusion on multi-modal image data, combine data from different modalities in three-dimensional space, and the weighted fusion algorithm is expressed as:
[0086] Among them, I fused (x, y, z) is the voxel value at the spatial position (x, y, z) of the fused three-dimensional image data, w m is the weight of the m-th modality, usually adjusted according to the quality and credibility of the modality, and M is the number of modalities;
[0087] S24, generating anatomical structure features: Use the threshold segmentation method to extract the anatomical structure features of the uterus from the fused three-dimensional image data. The simple segmentation based on the threshold is expressed as:
[0088]
[0089] Among them, S m (x, y, z) is the uterine region label at the spatial position (x, y, z), 1 represents the uterine region, 0 represents the non-uterine region, T is the threshold for segmentation, set according to the image intensity distribution, and the value range is 100 to 50;
[0090] S25, 3D Model Generation and Visualization: Based on the segmentation results and fused image data, the final 3D uterine model is generated through 3D reconstruction technology. The surface reconstruction (Marching Cubes) method is used to visualize the uterine model. The specific steps are as follows:
[0091] (1) Perform surface reconstruction on the segmentation results to generate a triangular mesh, expressed as:
[0092] Vertices, Triangles = MarchingCubes(I fused );
[0093] (2) Generate the final 3D uterine model based on the surface mesh data;
[0094] S26, Output 3D Uterine Model: Output the generated 3D uterine model.
[0095] S3 includes:
[0096] S31, Implement Lesion Feature Segmentation in the 3D Uterine Model: Use the region growing algorithm to segment the lesions in the 3D uterine model. Based on the known lesion seed points, expand the lesion area by gray value. The region growing process can be carried out recursively, expressed as:
[0097] R(x,y,z) = {(x′,y′,z′)||I fused (x′,y′,z′)-I seed |<T region}};
[0098] Among them, R(x,y,z) represents the expanded lesion area, I seed is the gray value of the seed point, T region is the specified region growing threshold, representing the allowable range of pixel intensity difference;
[0099] S32, Extract Morphological Parameters from the Segmented Lesion Area to Quantify the Size Characteristics of Fibroids. The morphological parameters include volume and surface area calculation, expressed as:
[0100] V lesion =∑ (x,y,z)∈S ΔV;
[0101] A lesion =∑ boundary ΔA;
[0102] Among them, V lesion is the volume of the lesion, S is the voxel set of the lesion area, S(x,y,z) = 1, ΔV is the volume of each voxel, A lesionis the surface area of the lesion, ΔA is the area of each triangular patch, and boundary represents the boundary surface of the lesion area;
[0103] S33. Extract texture features from the lesion area: Using the gray-level co-occurrence matrix texture analysis method, extract the texture features of the lesion, which are expressed as:
[0104] GLCM(x,y,z) = ∑ i,j p(i,j)·g(i,j);
[0105] Among them, p(i,j) is the joint probability distribution between gray values i and j, and g(i,j) is the calculation function of the gray-level co-occurrence matrix;
[0106] S34. Output the lesion segmentation area, morphological parameters, and texture features.
[0107] S4 includes:
[0108] S41. Perform hemodynamic analysis based on the lesion segmentation result: Using the lesion segmentation result, calculate the blood perfusion parameters of the lesion area through a hemodynamic model;
[0109] S42. Extract the vascular distribution features of the lesion area based on the hemodynamic analysis result of the lesion area;
[0110] S43. Output the blood perfusion parameters and vascular distribution features.
[0111] S41 uses an image-based perfusion model to calculate the blood irrigation parameters, which are expressed as:
[0112] Among them, P lesion (x,y,z) is the blood perfusion parameter at the lesion area S(x,y,z), Q blood (x,y,z) is the blood flow through the lesion area S(x,y,z) per unit time, and V lesion (x,y,z) is the volume of the lesion area, which can usually be obtained from the segmentation result.
[0113] S42 includes:
[0114] S421. Vascular distribution feature extraction: Use an edge detection algorithm to detect the boundaries of blood vessels in the image and segment the vascular area, which is expressed as:
[0115]
[0116] Among them, V vessel (x,y,z) is the binary result of the vascular distribution, 1 represents the vascular area, 0 represents the non-vascular area, and I fused(x, y, z) is the intensity value of each voxel in the fused image data, μ vessel is the average intensity value of the vascular region, T vessel is the vascular segmentation threshold to determine which pixels belong to the vascular region;
[0117] S422, Vessel diameter and branch feature calculation: Calculate the vessel diameter based on the voxel density and geometric morphology of the vascular region, expressed as:
[0118] where, D vessel is the diameter of the vessel, Area(V vessel ) is the area of the vascular region.
[0119] S5 includes:
[0120] S51, Fusing blood perfusion parameters and vascular distribution features: Fuse the calculated blood perfusion parameters and vascular distribution features to generate a comprehensive feature dataset, expressed as:
[0121] F lesion (x, y, z) = w1·P lesion (x, y, z) + w2·V vessel (x, y, z);
[0122] where, F lesion (x, y, z) is the fused lesion feature, P lesion (x, y, z) is the blood perfusion parameter of the lesion region, V vessel (x, y, z) is the vascular distribution feature of the lesion region, and w1 and w2 are the weighting coefficients of the blood perfusion parameter and the vascular distribution feature respectively;
[0123] S52, Combining the three-dimensional uterine model to construct a distribution map of the biomechanical properties of the lesion. The specific steps are as follows:
[0124] (1) Lesion region hardness calculation, expressed as:
[0125]
[0126] where, E lesion (x, y, z) is the elastic modulus of the lesion region, P lesion (x, y, z) is the blood perfusion parameter of the lesion region, V vessel (x, y, z) is the vascular distribution feature of the lesion region, and w1 and w2 are the weight coefficients,
[0127] It is a small constant used to avoid division-by-zero errors. In this model, there is an inverse relationship between the blood perfusion parameter and the elastic modulus (i.e., a high-perfusion region corresponds to low hardness), while the blood vessel distribution feature is directly proportional to the elastic modulus (i.e., a region rich in blood vessels usually has a higher hardness).
[0128] (2) Stress and strain calculation: The relationship between stress, elastic modulus, and strain is expressed as:
[0129] σ lesion (x,y,z) = E lesion (x,y,z)·ε lesion (x,y,z);
[0130] where σ lesion (x,y,z) is the stress in the lesion area, E lesion (x,y,z) is the elastic modulus in the lesion area, and ε lesion (x,y,z) is the strain in the lesion area.
[0131] S6 includes:
[0132] S61, generating a navigation path based on the biomechanical property distribution map: According to the biomechanical properties of the lesion area and the identification of the target area, generate the optimal ablation navigation path, expressed as:
[0133]
[0134] where S target (x,y,z) represents the lesion target area, 1 is the target area, and 0 is the non-target area;
[0135] S62, calculating the ablation energy and outputting a positioning report: According to the hardness (elastic modulus) and volume of the lesion area, calculate the recommended ablation energy value for the lesion area, expressed as:
[0136] E ablation = α·V lesion ·E lesion (x,y,z);
[0137] where α is the proportionality coefficient between the ablation energy and the hardness, V lesion is the lesion volume,
[0138] E ablation is the recommended ablation energy value for the lesion area, and E lesion (x,y,z) is the elastic modulus of the lesion;
[0139] S63, outputting a positioning report including the ablation energy and the navigation path.
[0140] Those of ordinary skill in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of the present invention is limited to these examples; under the concept of the present invention, the technical features in the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations in different aspects of the present invention as described above, which are not provided in detail for the sake of brevity.
Claims
1. A method for accurately locating uterine fibroid lesions by combining image analysis and data fusion, characterized in that: The following steps are involved: S1, acquiring multimodal uterine imaging data, performing artifact suppression and spatial registration on the original images, and outputting a preprocessed multimodal imaging data set; S2, performing three-dimensional reconstruction on the preprocessed multimodal image data set to generate a three-dimensional uterine model including anatomical structure features; S3, implement lesion feature segmentation in the three-dimensional uterine model, extract the morphological parameters and texture features of the myoma lesions, and output the lesion segmentation results; S4, perform hemodynamic analysis based on the lesion segmentation results to obtain the blood perfusion parameters and vascular distribution characteristics of the lesion area; S5, integrating blood perfusion parameters and vascular distribution characteristics, and combining with a three-dimensional uterine model to construct a distribution map of the biomechanical properties of the lesion; S6, generating a real-time navigation path according to the biomechanical property distribution map, and outputting a positioning report including ablation energy recommendation value.
2. The method for accurately locating uterine fibroid lesions by combining image analysis and data fusion according to claim 1, characterized in that: The S1 includes: S11, obtaining multimodal uterine imaging data: obtaining multimodal data of the uterus through a variety of imaging devices, wherein the multimodal data includes magnetic resonance imaging, ultrasound imaging, and computer tomography; S12, raw image data artifact suppression: performing wavelet transform artifact suppression processing on the acquired raw image data to remove imaging interference; S13, spatial registration: spatially registering the multimodal image data so that the spatial coordinate systems of each modality data are aligned, and registering using a mutual information maximization algorithm; S14, outputting the preprocessed multimodal image dataset: outputting the multimodal image dataset after artifact suppression and spatial registration processing.
3. The method for accurately locating uterine fibroid lesions by combining image analysis and data fusion according to claim 2, characterized in that: The S2 includes: S21, inputting a preprocessed multimodal image data set: receiving an output data set from a multimodal image data preprocessing step; S22, using voxel-based 3D reconstruction method: using voxel representation, construct a preliminary 3D uterus model; S23, fusing multimodal image information: performing weighted fusion on multimodal image data, combining data from different modalities in a three-dimensional space; S24, generating anatomical structure features: extracting anatomical structure features of the uterus from the fused three-dimensional image data using a threshold segmentation method; S25, 3D model generation and visualization: Based on the segmentation results and fused image data, the final 3D uterus model is generated by 3D reconstruction technology, and the uterus model is visualized by surface reconstruction method; S26, outputting the three-dimensional uterus model: outputting the generated three-dimensional uterus model.
4. The method for accurately locating uterine fibroid lesions by combining image analysis and data fusion according to claim 3, characterized in that: The S3 includes: S31, implementing lesion feature segmentation in the three-dimensional uterine model: segmenting the lesions in the three-dimensional uterine model using a region growing algorithm; S32, extracting morphological parameters from the segmented lesion area to quantify the size characteristics of the fibroid; S33, extracting texture features from the lesion area: using a gray level co-occurrence matrix texture analysis method to extract texture features of the lesion; S34, the output includes lesion segmentation area, morphological parameters and texture features.
5. The method for accurately locating uterine fibroid lesions by combining image analysis and data fusion according to claim 4, characterized in that: The S4 includes: S41, performing hemodynamic analysis based on the lesion segmentation result: using the lesion segmentation result, calculating the blood perfusion parameters of the lesion area through the hemodynamic model; S42, extracting the vascular distribution characteristics of the lesion area based on the hemodynamic analysis results of the lesion area; S43, output blood perfusion parameters and vascular distribution characteristics.
6. The method for accurately locating uterine fibroid lesions by combining image analysis and data fusion according to claim 5, characterized in that: The S41 calculates blood flow irrigation parameters using an image-based perfusion model.
7. The method for accurately locating uterine fibroid lesions by combining image analysis and data fusion according to claim 6, characterized in that: The S42 includes: S421, Blood vessel distribution feature extraction: Use edge detection algorithm to detect the boundaries of blood vessels in the image; S422, Calculation of blood vessel diameter and branch characteristics: Calculate the blood vessel diameter based on the voxel density and geometric morphology of the blood vessel area.
8. The method for accurately locating uterine fibroid lesions by combining image analysis and data fusion according to claim 7, characterized in that: The S5 includes: S51, fusion of blood perfusion parameters and vascular distribution characteristics: fusion of the calculated blood perfusion parameters and vascular distribution characteristics to generate a comprehensive feature data set; S52, combined with the three-dimensional uterine model, construct a distribution map of the biomechanical properties of the lesion.
9. The method for accurately locating uterine fibroid lesions by combining image analysis and data fusion according to claim 8, characterized in that: The S6 includes: S61, generating a navigation path according to the biomechanical property distribution map: generating an optimal ablation navigation path according to the biomechanical properties of the lesion area and the identification of the target area; S62, calculating ablation energy and outputting a positioning report: calculating a recommended ablation energy value of the lesion area according to the hardness and volume of the lesion area; S63, outputting a positioning report including ablation energy and navigation path.
Citation Information
Cited By
Focused ultrasound system with multiple treatment modes
CN121130332A
Breast duct lesion point identification method based on ultrasound contrast
CN121258923A
