Vascular hyperplasia quantitative grading method and device, electronic equipment and storage medium

Abstract

The invention provides a blood vessel hyperplasia quantitative grading method and device, electronic equipment and a storage medium, which are applied to the technical field of image processing, and provide double-view image data for subsequent steps by obtaining a stereoscopic image of a lesion area; the disparity map is utilized to construct the three-dimensional point cloud model, the first blood vessel region recognized in the two-dimensional image is positioned into the three-dimensional space to obtain the corresponding second blood vessel region, the geometrical characteristics of the blood vessel in the three-dimensional space are extracted by calculating the geometrical parameters of the blood vessel, and necessary parameters are provided for more accurate quantitative analysis of the blood vessel. By means of the three-dimensional density value obtained through calculation of the three-dimensional geometric parameters, objective and quantitative grading of the vascular hyperplasia degree is achieved, the defects that a traditional method is subjective and lacks physical distance guidance are overcome, and the accuracy and reliability of vascular hyperplasia evaluation are improved.

Description

Technical Field

[0001] This application relates to the field of image processing technologies, and particularly to a method, apparatus, electronic device, and storage medium for quantifying and grading vascular hyperplasia. Background Art

[0002] With the rapid and increasingly mature development of science and technology, minimally invasive surgery has become an important method in hospital surgeries. In the examination of gynecological patients, laparoscopic exploration is an important means. Doctors can conduct abdominal exploration to closely observe the internal details of the patients.

[0003] In the existing gynecological surgical explorations, people only focus on anatomical changes or the characteristics of masses, and do not pay special attention to blood vessels. There are neither relevant indicators nor corresponding standards. However, everyone also realizes that blood vessel changes have clinical significance.

[0004] The blood vessels seen under the laparoscope in gynecological surgeries are mainly on the surfaces of pelvic organs, such as the peritoneum, bladder, uterus, fallopian tubes, etc. However, the blood vessel distribution, density, morphology, color, filling, etc. are all ignored. The main reason is that there is no good objective evaluation method and means, let alone standards. And there is a clear relationship between vascular hyperplasia and inflammation. This leads to the inability to diagnose some non-specific inflammations. Even after surgery, they are ignored and missed due to the lack of objective indicators, thus affecting the understanding and treatment of diseases.

[0005] With the development of image recognition technology, by calculating and comparing various characteristics of the blood vessels under the mirror and combining clinical data, a diagnostic model is designed and deduced, making it possible to be used as a new diagnostic method and means. Especially important characteristics such as blood vessel density, as a tool for understanding diseases, can help clinicians diagnose more clearly and treat more pertinently.

[0006] However, the traditional method of using a 2D endoscope to take images and calculate the density of blood vessels at the pixel level lacks the guidance of real physical distance. It is easy to obtain incorrect judgments of blood vessel density due to the distance of the hand-held endoscope during shooting, thus having an adverse impact on disease judgment. Moreover, the traditional technology lacks an intraoperative blood vessel quantification and grading system based on three-dimensional spatial information, especially a solution that can integrate blood vessel physical dimensions, dynamic blood flow parameters, and clinical grading standards.

[0007] Therefore, there is a lack of a solution in the existing technology that can overcome the lack of guidance of real physical distance and, based on binocular ranging under a 3D laparoscope, more accurately obtain blood vessel density and other characteristics to judge the lesion results. Summary of the Invention

[0008] In view of the deficiencies of the above-mentioned prior art, a method, device, electronic device and storage medium for quantifying and grading vascular hyperplasia provided by the present application are applied to the field of image processing technology. This method can overcome the problem of lack of real physical distance guidance, introduce a binocular ranging scheme based on 3D laparoscopy, and has the beneficial effect of being able to more accurately obtain vascular density and other vascular geometric parameters to judge the lesion result.

[0009] In the first aspect, a method for quantifying and grading vascular hyperplasia, the method includes: S1: Obtain a stereoscopic image of the lesion area, where the stereoscopic image at least includes left-eye image information and right-eye image information; S2: Perform stereoscopic correction on the left-eye image information and the right-eye image information to obtain a disparity map between the two images, and construct a three-dimensional point cloud model according to the disparity map; S3: Segment the first vascular region according to the corrected left-eye image information, and convert the first vascular region into a corresponding second vascular region in the three-dimensional point cloud model; S4: Calculate the vascular geometric parameters of the second vascular region, where the vascular geometric parameters at least include: vascular centerline, vascular center point; S5: Calculate the vascular width according to the vascular center point, calculate the three-dimensional density value of the blood vessel according to the vascular centerline and the vascular width, and grade the blood vessel according to the three-dimensional density value.

[0010] The present application provides a method for quantifying and grading vascular hyperplasia. By obtaining the stereoscopic image of the lesion area, it provides dual-view image data for subsequent steps; by constructing a three-dimensional point cloud model through the disparity map, the conversion from a two-dimensional image to a three-dimensional space is realized, providing three-dimensional space information for subsequent vascular analysis. The technical contribution of this step lies in obtaining the three-dimensional structural information of the lesion area through stereoscopic vision technology, overcoming the limitation of the lack of depth information in traditional two-dimensional images. By positioning the first vascular region identified in the two-dimensional image into the three-dimensional space to obtain the corresponding second vascular region, the accurate conversion of the first vascular region from the two-dimensional image space to the three-dimensional point cloud space is realized, laying a foundation for subsequent calculation of vascular geometric parameters. By calculating the vascular geometric parameters, the geometric features of the blood vessel in the three-dimensional space are extracted, providing necessary parameters for more accurate vascular quantitative analysis. Using the three-dimensional density value calculated by the three-dimensional geometric parameters, an objective and quantitative grading of the degree of vascular hyperplasia is realized, overcoming the subjectivity of traditional methods and the defect of lack of physical distance guidance, and improving the accuracy and reliability of vascular hyperplasia assessment.

[0011] Further, step S2 includes: S21: Obtain calibration parameters according to the left-eye image information and the right-eye image information; S22: Perform epipolar correction on the left-eye image information and the right-eye image information according to the calibration parameters to eliminate lens distortion; S23: Calculate the disparity map between the left-eye image information and the right-eye image information after eliminating lens distortion; S24: Construct the three-dimensional point cloud model according to the disparity map.

[0012] This application provides a method for quantifying and grading vascular hyperplasia. Calibration parameters are obtained, which include camera internal parameters, external parameters, and distortion parameters. By using the obtained calibration parameters, epipolar correction is performed on the left-eye image information and the right-eye image information, thereby effectively eliminating or reducing lens distortion and improving image quality. By calculating the disparity map between the left-eye image information and the right-eye image information, since the input images have completed distortion correction, the obtained disparity map is more accurate; finally, a high-precision disparity map is used to construct a three-dimensional point cloud model, thus ensuring the accuracy and reliability of the three-dimensional point cloud model.

[0013] Further, step S24 includes: S241: Obtain the disparity value of each pixel point in the left-eye image information and the right-eye image information according to the disparity map; S242: Obtain the focal length, baseline distance, and noise compensation term of the camera; S243: Calculate the depth value of each pixel point according to the focal length, the baseline distance, the noise compensation term, and the disparity value; S244: Convert the depth value into a depth map and construct the three-dimensional point cloud model according to the depth map.

[0014] This application provides a method for quantifying and grading vascular hyperplasia. The disparity value of each pixel point is obtained through the disparity map; the disparity value reflects the position difference of the pixel point in the left and right-eye images, and thus can be used to calculate the depth value; by obtaining the focal length, baseline distance, and noise compensation term of the camera; the focal length and baseline distance are important parameters of the stereo vision system, which determine the proportional relationship of depth calculation; the noise compensation term takes into account the possible errors in practical applications and improves the accuracy of depth calculation. By calculating the depth value of each pixel point and converting the depth value into a depth map, a three-dimensional point cloud model can be constructed according to the depth map. This method provides a complete set of steps for constructing a three-dimensional point cloud model with strong operability, making it possible to construct an accurate three-dimensional point cloud model according to the disparity map and laying a foundation for the subsequent three-dimensional analysis of the vascular region.

[0015] Further, step S3 includes: S31: Perform CLAHE contrast enhancement on the corrected left-eye image information; S32: Perform semantic segmentation on the enhanced left-eye image information to extract the binary mask of blood vessels; S33: Perform morphological opening operation on the binary mask of blood vessels to obtain the pixel coordinate set of the first blood vessel region; S34: Map the pixel coordinate set of the first blood vessel region to the three-dimensional point cloud model to obtain the second blood vessel region.

[0016] The present application provides a method for quantitative grading of blood vessel hyperplasia. First, perform CLAHE contrast enhancement on the corrected left-eye image information, which can effectively improve the local contrast of the image, make blood vessels clearer in the image, and overcome the problem of insufficient contrast that may exist in the original image. Then, use semantic segmentation technology to process the enhanced left-eye image information, accurately identify the blood vessel region in the image, and generate the binary mask of blood vessels. Next, perform morphological opening operation on the binary mask of blood vessels to eliminate small noise points and burrs in the image, smooth the blood vessel contour, further optimize the blood vessel segmentation result, and obtain a more accurate pixel coordinate set of the first blood vessel region. Finally, map the pixel coordinate set of the first blood vessel region obtained through the above processing to the three-dimensional point cloud model to obtain the second blood vessel region corresponding to the first blood vessel region, achieving the purpose of accurately converting the two-dimensional image segmentation result to the three-dimensional space. Through the above series of steps, this technical solution can more accurately and reliably segment and extract the blood vessel region, lay a foundation for subsequent calculation of blood vessel geometric parameters and blood vessel grading, and effectively improve the accuracy and reliability of quantitative grading of blood vessel hyperplasia.

[0017] Further, step S34 includes: S341: Perform positioning calculation on the pixel coordinate set of the first blood vessel region based on the normalized cross-correlation algorithm to obtain the two-dimensional blood vessel pixel coordinates of the pixel coordinate set of the first blood vessel region; S342: Perform three-dimensional matrix transformation on the two-dimensional blood vessel pixel coordinates to obtain a three-dimensional blood vessel point cloud subset, and the position of the three-dimensional blood vessel point cloud subset in the three-dimensional point cloud model is the second blood vessel region.

[0018] Further, step S4 includes: S41: Perform DBSCAN clustering calculation on the binary mask to generate multiple independent blood vessel point cloud clusters, and the independent blood vessel point cloud clusters are the blood vessel centerlines of each blood vessel in the second blood vessel region; S42: Calculate the three-dimensional centroid of each independent blood vessel point cloud cluster to obtain the blood vessel center point of each blood vessel in the second blood vessel region.

[0019] Further, step S5 includes: S51: Arbitrarily select a cubic region in the three-dimensional point cloud model, where the cubic region contains one or more blood vessels; S52: Perform cylindrical fitting on the central points of the blood vessels to obtain the blood vessel width, and calculate the three-dimensional density value of the blood vessels in the cubic region according to the blood vessel center line and the blood vessel width; S53: Grade the blood vessels according to the three-dimensional density value.

[0020] In a second aspect, a device for quantifying and grading blood vessel hyperplasia is applied to the steps of any one of the above methods. The device includes: An image acquisition module: configured to acquire a stereoscopic image of a lesion area, where the stereoscopic image at least includes left-eye image information and right-eye image information; A point cloud construction module: configured to perform stereoscopic correction on the left-eye image information and the right-eye image information to obtain a disparity map between the two images, and construct a three-dimensional point cloud model according to the disparity map; A blood vessel segmentation module: configured to segment a first blood vessel area according to the corrected left-eye image information, and convert the first blood vessel area into a corresponding second blood vessel area in the three-dimensional point cloud model; A parameter calculation module: configured to calculate blood vessel geometric parameters of the second blood vessel area, where the blood vessel geometric parameters at least include: a blood vessel center line and a blood vessel central point; A blood vessel grading module: configured to calculate the blood vessel width according to the blood vessel central point, calculate the three-dimensional density value of the blood vessel according to the blood vessel center line and the blood vessel width, and grade the blood vessel according to the three-dimensional density value.

[0021] In a third aspect, the present application provides an electronic device, including a processor and a memory. The memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the steps in any one of the above methods are run.

[0022] In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps in any one of the above methods are run.

[0023] Beneficial effects: A method, device, electronic device, and storage medium for quantifying and grading angiogenesis proposed in this application provide dual-view image data for subsequent steps by acquiring a stereoscopic image of a lesion area; a three-dimensional point cloud model is constructed through a disparity map, realizing the conversion from a two-dimensional image to a three-dimensional space and providing three-dimensional space information for subsequent blood vessel analysis. The technical contribution of this step lies in obtaining the three-dimensional structural information of the lesion area through stereoscopic vision technology, overcoming the limitation of traditional two-dimensional images lacking depth information. By positioning the first blood vessel area identified in the two-dimensional image into the three-dimensional space to obtain the corresponding second blood vessel area, an accurate conversion of the first blood vessel area from the two-dimensional image space to the three-dimensional point cloud space is achieved, laying a foundation for subsequent calculation of blood vessel geometric parameters. By calculating blood vessel geometric parameters, the geometric features of blood vessels in the three-dimensional space are extracted, providing necessary parameters for more accurate blood vessel quantification analysis. Using the three-dimensional density value calculated from the three-dimensional geometric parameters, an objective and quantitative grading of the degree of angiogenesis is achieved, overcoming the subjectivity of traditional methods and the defect of lacking physical distance guidance, and improving the accuracy and reliability of angiogenesis assessment. Brief Description of the Drawings

[0024] Figure 1 It is a flowchart of a method for quantifying and grading angiogenesis proposed in this application.

[0025] Figure 2 It is a structural diagram of a device for quantifying and grading angiogenesis proposed in this application.

[0026] Figure 3 It is a structural diagram of an electronic device proposed in this application.

[0027] Label Description: 201, Image Acquisition Module; 202, Point Cloud Construction Module; 203, Blood Vessel Segmentation Module; 204, Parameter Calculation Module; 205, Blood Vessel Grading Module; 301, Processor; 302, Memory; 303, Communication Bus; 3, Electronic Device. Detailed Embodiments

[0028] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Usually, the components of the embodiments of the present application described and marked in the accompanying drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application to be protected, but only represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present application.

[0029] It should be noted that: Similar reference numerals and letters denote similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. At the same time, in the description of this application, terms such as "first, second" are only used for distinguishing descriptions and cannot be construed as indicating or implying relative importance.

[0030] The following disclosure provides many different embodiments or examples for achieving the purpose of the present invention, solving the problems in the traditional technology such as the lack of real physical distance guidance and the lack of an intraoperative blood vessel quantification and grading system based on three-dimensional space information, especially the solution that can integrate blood vessel physical dimensions, dynamic blood flow parameters and clinical grading criteria. Therefore, this application proposes a method, device, electronic device and storage medium for quantifying and grading blood vessel hyperplasia, which are specifically as follows: Please refer to Figure 1 , in the first aspect, a method for quantifying and grading blood vessel hyperplasia, the method includes: S1: Obtain a stereoscopic image of the lesion area, and the stereoscopic image at least includes left-eye image information and right-eye image information; S2: Perform stereoscopic correction on the left-eye image information and the right-eye image information to obtain a disparity map between the two images, and construct a three-dimensional point cloud model according to the disparity map; S3: Segment the first blood vessel area according to the corrected left-eye image information, and convert the blood vessel area to the corresponding second blood vessel area in the three-dimensional point cloud model; S4: Calculate the blood vessel geometric parameters of the second blood vessel area, and the blood vessel geometric parameters at least include: blood vessel centerline, blood vessel center point; S5: Calculate the blood vessel width according to the blood vessel center point, calculate the three-dimensional density value of the blood vessel according to the blood vessel centerline and the blood vessel width, and grade the blood vessel according to the three-dimensional density value.

[0031] Among them, in step S1, the acquisition of the stereoscopic image can be completed by a binocular vision system. This system is equipped with two cameras, which respectively collect the left-eye image information and the right-eye image information of the lesion area. The two cameras are set to simulate the human eye perspective and capture images of the same lesion area from different angles, thereby obtaining a stereoscopic image pair containing disparity information. This binocular vision system can be a 3D laparoscope with a resolution of 1920×1080@30fps.

[0032] In step S2, stereo rectification is an image preprocessing technique used to eliminate or reduce the geometric distortion of images caused by camera lens distortion and installation errors. Epipolar rectification is a possible stereo rectification method, through which the scan lines of the left and right eye images can be aligned, simplifying the subsequent disparity calculation. The generation of the disparity map can be achieved through stereo matching algorithms such as the block matching algorithm or the optical flow method. These algorithms aim to find the position deviation, i.e., the disparity value, between corresponding pixel points in the left and right eye images.

[0033] The construction of the three-dimensional point cloud model is based on the disparity map and camera calibration parameters. The three-dimensional coordinates of each pixel point can be calculated from the disparity value and calibration parameters. The set of three-dimensional coordinates of all pixel points constitutes the three-dimensional point cloud model, which is a discretized representation of the three-dimensional structure of the lesion area.

[0034] In step S3, various image segmentation algorithms can be used for the segmentation of the first vascular region, such as threshold segmentation, region growing, edge detection, or semantic segmentation. Semantic segmentation is a preferred image segmentation method that can achieve pixel-level vascular region recognition. The conversion of the vascular region from a two-dimensional image to a three-dimensional point cloud model can be realized through the coordinate mapping relationship, that is, according to the camera imaging model and calibration parameters, the two-dimensional image pixel coordinates are converted into three-dimensional point cloud coordinates, thereby obtaining the second vascular region in three-dimensional space.

[0035] In step S4, the calculation of vascular geometric parameters can be assisted by point cloud processing and analysis techniques. The extraction of the vascular centerline can be achieved through the skeleton extraction algorithm or topological analysis method. The vascular center point can be the average coordinate or centroid coordinate of the points on the vascular centerline.

[0036] In step S5, the vascular width can be obtained by fitting a cylinder with the vascular center point; the calculation of the three-dimensional vascular density value can be defined as the length or volume of the blood vessels per unit volume. The cube region can be a regular region set artificially in the three-dimensional point cloud model or an irregular region adaptively determined according to the anatomical structure or pathological characteristics of the lesion area. The formulation of the vascular grading standard can refer to the existing vascular hyperplasia grading standards or clinical guidelines, or a new grading standard can be determined through statistical analysis and experimental verification. The grading results can be used to evaluate the degree of vascular hyperplasia and the severity of the lesion.

[0037] Specifically, the working principle of this method for quantifying and grading angiogenesis is as follows: First, a stereo image of the lesion area is obtained through a binocular vision system, providing a data basis for subsequent 3D reconstruction and analysis. Then, the stereo image is corrected and the disparity is calculated to eliminate image distortion, obtaining a disparity map. Based on the disparity map, a 3D point cloud model of the lesion area is constructed to achieve the conversion from a 2D image to 3D space. Next, the blood vessel area is segmented in the corrected left-eye image, and the segmentation result is mapped onto the 3D point cloud model to obtain the blood vessel point cloud data in 3D space. On this basis, geometric parameters of the blood vessels are calculated, including the centerline, center point, and width, which are quantitative descriptions of the blood vessel morphology and size. Finally, the 3D density value of the blood vessels is calculated based on the geometric parameters of the blood vessels, and the degree of angiogenesis is graded according to the density value to achieve the quantitative evaluation of angiogenesis. Through the above steps, this method can accurately quantify the degree of angiogenesis in 3D space, overcome the limitation of the traditional 2D method lacking depth information, improve the objectivity and accuracy of angiogenesis evaluation, and provide a more reliable basis for clinical diagnosis and treatment.

[0038] In some specific embodiments, the stereo image of the lesion area is synchronously captured by the dual cameras of a 3D laparoscope. The optical axes of the two cameras are parallel and the baseline distance is 10 mm, and the image resolution is 1920x1080@30fps.

[0039] The Bouguet calibration method is used to calibrate the cameras to obtain the internal parameter matrix, distortion parameters, and external parameter matrix. The distortion parameters include radial distortion coefficients and tangential distortion, and the external parameter matrix includes a rotation matrix and a translation matrix. The epipolar correction algorithm is used to correct the left-eye and right-eye images, and the horizontal disparity range of the corrected images is between 0 and 200 pixels.

[0040] The Semi-Global Block Matching (SGBM) algorithm is used to calculate the disparity map. The disparity search range is 1 - 256 pixels, and the penalty parameters are P1 = 8 and P2 = 32.

[0041] Based on the disparity map and calibration parameters, the triangulation method is used to calculate the 3D coordinates of each pixel point to construct a 3D point cloud model of the lesion area, and the point cloud density is 1000 points per cubic millimeter.

[0042] In the step of cutting the first vascular region, first, CLAHE contrast enhancement is performed on the left-eye image, and then a pre-trained deep learning model (such as U-Net) is used for vascular semantic segmentation to obtain a vascular binary mask. Morphological opening operation with a 3x3 elliptical kernel is performed on the binary mask to remove small noise and artifacts, and a pixel coordinate set of the first vascular region is obtained. The normalized cross-correlation algorithm is used to search for the corresponding pixel coordinates of the first vascular region in the right-eye image to realize the conversion of the vascular region from a two-dimensional image to a three-dimensional point cloud model, and a second vascular region is obtained.

[0043] Then, DBSCAN clustering is performed on the point cloud data of the second vascular region, with a clustering radius epsilon = 0.5 mm and a minimum number of points minPts = 5, generating multiple independent vascular point cloud clusters, and each point cloud cluster represents the centerline of a blood vessel. The three-dimensional centroid of each point cloud cluster is calculated as the vascular center point. Cylindrical fitting is performed on each point cloud cluster to obtain the vascular radius, and the vascular width is twice the radius.

[0044] Finally, a 10x10x10 mm cube region can be manually selected in the three-dimensional point cloud model, and the total volume of all blood vessels in this region is calculated. The three-dimensional vascular density value is defined as the ratio of the total vascular volume to the volume of the cube region. According to the three-dimensional density value, the degree of vascular hyperplasia is divided into four grades: Grade I - density value ρ ∈ [0, 0.2), Grade II - density value ρ ∈ [0.2, 0.4), Grade III - density value ρ ∈ [0.4, 0.6), Grade IV - density value ρ ≥ 0.6. This grading standard can be adjusted and optimized according to clinical pathological results. Thus, a quantitative grading assessment of the degree of vascular hyperplasia in the pelvic lesion area of gynecological patients can be realized.

[0045] Furthermore, step S2 includes: S21: Obtain calibration parameters according to the left-eye image information and the right-eye image information; S22: Perform epipolar correction on the left-eye image information and the right-eye image information according to the calibration parameters to eliminate lens distortion; S23: Calculate the disparity map between the left-eye image information and the right-eye image information after eliminating lens distortion; S24: Construct a three-dimensional point cloud model according to the disparity map.

[0046] Among them, in step S21, the acquisition of calibration parameters can be completed through a standard camera calibration process. For example, using the checkerboard calibration method, multiple calibration board images at different angles are taken, and through corner detection and parameter optimization, the camera internal parameter matrix, distortion coefficients, and external parameter matrices of the left and right cameras are calculated. The distortion parameters include radial distortion coefficients and tangential distortion, and the external parameter matrix includes a rotation matrix and a translation matrix.

[0047] In step S22, through calibration parameters, the left-eye image information and the right-eye image information are projected onto the same plane in epipolar rectification, so that the pixel rows of the two images are aligned. Thus, subsequent disparity search is constrained in the horizontal direction, improving the search efficiency and accuracy, and eliminating the influence of lens distortion.

[0048] In step S23, the calculation of the disparity map can adopt the Semi-Global Block Matching (SGBM) algorithm. For the rectified left-eye image information and right-eye image information, matching pixel points are searched in the horizontal direction, and the disparity values of the pixel points are calculated. The disparity value is inversely proportional to the distance of the pixel point.

[0049] In step S24, the construction of the three-dimensional point cloud model is based on the disparity map and camera parameters. Using the principle of triangulation, the coordinates of each pixel point in the three-dimensional space are calculated, and the set of these three-dimensional coordinates constitutes the three-dimensional point cloud model.

[0050] Furthermore, step S24 includes: S241: Obtain the disparity value of each pixel point in the left-eye image information and the right-eye image information according to the disparity map; S242: Obtain the focal length, baseline distance, and noise compensation term of the camera; S243: Calculate the depth value of each pixel point according to the focal length, baseline distance, noise compensation term, and disparity value; S244: Convert the depth value into a depth map and construct a three-dimensional point cloud model according to the depth map.

[0051] Among them, in step S241, each pixel value in the disparity map represents the horizontal displacement amount between the corresponding pixel points of the left-eye image and the right-eye image, and this displacement amount is the disparity value. The way to obtain the disparity value can be to directly read the pixel value of the disparity map.

[0052] In step S242, the focal length and baseline distance are the inherent parameters of the stereo vision system and can be obtained in advance through camera calibration. The noise compensation term is introduced to improve the accuracy of depth calculation. It may be a fixed value obtained through experiments or empirical estimates and is used to correct system errors.

[0053] In step S243, the depth value is calculated based on the principle of triangulation. The specific formula is: Depth value = (focal length * baseline distance) / disparity value + noise compensation term.

[0054] In step S244, a depth map is an image in which the grayscale value of each pixel represents the depth value corresponding to that pixel. The depth map can be obtained by arranging the depth values of each pixel point calculated in step S243 in matrix form. A three-dimensional point cloud model is a set of points in space, and each point is represented by three-dimensional coordinates. The specific method for constructing the three-dimensional point cloud model can be that for each pixel in the depth map, based on the coordinates and depth value of the pixel, combined with the internal parameters of the camera, the coordinates of the pixel point in three-dimensional space are calculated, and the set of three-dimensional coordinates of all pixel points constitutes the three-dimensional point cloud model. Or it is also possible to construct a three-dimensional point cloud model based on the depth map and RGB color mapping through the Poisson reconstruction algorithm.

[0055] Further, step S3 includes: S31: Perform CLAHE contrast enhancement on the corrected left-eye image information; S32: Perform semantic segmentation on the enhanced left-eye image information to extract the binary mask of blood vessels; S33: Perform morphological opening operation on the binary mask of blood vessels to obtain the pixel coordinate set of the first blood vessel region; S34: Map the pixel coordinate set of the first blood vessel region to the three-dimensional point cloud model to obtain the second blood vessel region.

[0056] Among them, CLAHE contrast enhancement realizes the improvement of local contrast of the image by analyzing the local region histogram of the image and redistributing pixel intensities.

[0057] Semantic segmentation uses a deep learning model, such as U-Net, to perform pixel-level classification on the enhanced image to generate an accurate binary mask of blood vessels. Specifically, the depth of the U-Net encoder is 5.

[0058] Morphological opening operation can eliminate small noise points and burrs in the image, smooth the blood vessel contour, further optimize the blood vessel segmentation result, and more accurately obtain the pixel coordinate set of the first blood vessel region.

[0059] Morphological opening operation includes erosion and dilation operations. Through the operation of eroding first and then dilating, noise points are eliminated and the blood vessel edges are smoothed.

[0060] The mapping of the pixel coordinate set is achieved through camera calibration parameters and the disparity map. The pixel coordinates of the first blood vessel region are converted to three-dimensional point cloud coordinates, and thus the second blood vessel region is obtained.

[0061] In some specific embodiments, the CLAHE contrast enhancement algorithm is used to enhance the corrected left-eye image information. For example, the parameters of the CLAHE algorithm, such as the tile size and contrast limit, can be adjusted according to the actual image quality and vascular characteristics to achieve the optimal enhancement effect. As a preferred embodiment, the tile size is set to 8x8 pixels, and the contrast limit is set to 2. The semantic segmentation model adopts the U-Net network structure and is trained using a medical image dataset containing vascular images. The trained model can accurately identify and segment the vascular region. As a preferred embodiment, the encoder part of the U-Net network uses ResNet34 as the backbone network, the decoder part uses deconvolution and upsampling operations to gradually restore the image resolution, and the loss function uses the cross-entropy loss function. The morphological opening operation uses a 3x3 elliptical structural element, and the number of iterations is set to 1 time to eliminate the noise points in the vascular binary mask and smooth the vascular contour. As a preferred embodiment, both the erosion operation and the dilation operation use a 3x3 elliptical structural element. First, the erosion operation is performed to remove the noise points, and then the dilation operation is performed to restore the original size of the vascular region.

[0062] The coordinate mapping process is achieved through the following steps: First, obtain the camera calibration parameters, including the internal parameter matrix, external parameter matrix, and distortion parameters. Then, calculate the two-dimensional coordinates of each pixel point according to the pixel coordinates of the left-eye image and the disparity map. Finally, convert each pixel coordinate in the pixel coordinate set of the first vascular region into a three-dimensional space coordinate to obtain a three-dimensional point cloud subset of the second vascular region.

[0063] Specifically, the pixel coordinate set is: , where is the pixel coordinate set, is the two-dimensional coordinate of each pixel, is the vascular binary mask.

[0064] Furthermore, step S34 includes: S341: Perform positioning calculation on the pixel coordinate set of the first vascular region based on the normalized cross-correlation algorithm to obtain the two-dimensional vascular pixel coordinates of the pixel coordinate set of the first vascular region; S342: Perform three-dimensional matrix transformation on the two-dimensional vascular pixel coordinates to obtain a three-dimensional vascular point cloud subset, and the position of the three-dimensional vascular point cloud subset in the three-dimensional point cloud model is the second vascular region.

[0065] Among them, in step S341, the normalized cross - correlation algorithm is adopted to perform positioning calculation on the pixel coordinate set of the first blood vessel region. Specifically, the normalized cross - correlation algorithm can use the pixel coordinate set of the first blood vessel region as a template in the left - eye image information, search for the region in the right - eye image information that best matches this template, and determine the best - matching position by calculating the normalized cross - correlation coefficient between the template and the search region. This best - matching position is the two - dimensional blood vessel pixel coordinates of the pixel coordinate set of the first blood vessel region. Thus, the corresponding relationship of the first blood vessel region in the left - and right - eye images is accurately determined.

[0066] In step S342, the three - dimensional matrix transformation can convert the two - dimensional blood vessel pixel coordinates into three - dimensional space. Specifically, camera calibration parameters, such as the camera internal parameter matrix, external parameter matrix, and distortion parameters, are used to construct the three - dimensional matrix transformation relationship. Through this three - dimensional matrix transformation relationship, the two - dimensional blood vessel pixel coordinates can be projected into the three - dimensional point cloud model to obtain the corresponding three - dimensional coordinates. The set of these three - dimensional coordinates constitutes a three - dimensional blood vessel point cloud subset, and this three - dimensional blood vessel point cloud subset is the second blood vessel region. Through the accurate positioning of the normalized cross - correlation algorithm and the accurate conversion of the three - dimensional matrix transformation, the blood vessel region in the two - dimensional image can be accurately mapped into the three - dimensional point cloud model.

[0067] Specifically, aiming at the problem that the calculation of blood vessel density in the prior art lacks the guidance of real physical distance, the technical solution of this application, through step S34, uses the normalized cross - correlation algorithm to accurately locate the position of the two - dimensional blood vessel region in the three - dimensional point cloud model and performs three - dimensional matrix transformation, thereby obtaining an accurate three - dimensional blood vessel point cloud subset, that is, the second blood vessel region. In this way, the blood vessel region segmented from the two - dimensional image can be accurately converted into three - dimensional space, overcoming the problem of insufficient positioning accuracy that may be caused by direct simple correspondence. The second blood vessel region obtained in this way can more accurately reflect the true distribution of blood vessels in three - dimensional space, providing a more reliable data basis for the subsequent calculation of blood vessel geometric parameters and blood vessel grading. Compared with the method of directly simply corresponding two - dimensional pixel coordinates to the three - dimensional point cloud model, the technical solution of this application can improve the accuracy and reliability of the conversion from the blood vessel region in the two - dimensional image to the blood vessel region in the three - dimensional point cloud model, thus providing more accurate basic data for the subsequent quantification and grading of blood vessel hyperplasia.

[0068] Specifically, the calculation method of the three - dimensional blood vessel point cloud subset is as follows:

[0069] Among them, is the three - dimensional blood vessel point cloud subset, is the three - dimensional space coordinates of each pixel, is the two - dimensional coordinates of each pixel, is the coordinate transformation matrix.

[0070] Further, step S4 includes: S41: Perform DBSCAN clustering calculation on the binary mask to generate multiple independent vascular point cloud clusters, where the independent vascular point cloud clusters are the vascular centerlines of each blood vessel in the second vascular region; S42: Calculate the three-dimensional centroid of each independent vascular point cloud cluster to obtain the vascular center point of each blood vessel in the second vascular region.

[0071] Among them, in step S41, as a density-based clustering method, the DBSCAN clustering algorithm can effectively identify the vascular region in the binary mask. Specifically, the DBSCAN algorithm divides the pixel points in the binary mask into different clusters by setting two parameters: the neighborhood radius and the minimum number of neighborhood points. In the application scenario of vascular hyperplasia quantification and grading, each cluster in the binary mask can be considered as an independent blood vessel. Therefore, the multiple independent vascular point cloud clusters generated by DBSCAN clustering calculation can be defined as the vascular centerlines of each blood vessel in the second vascular region.

[0072] Specifically, when performing DBSCAN clustering on the binary mask, eps = 0.1mm, N independent vascular point cloud clusters are generated: , ... , where for each independent vascular point cloud cluster , k = 1, 2, 3... N.

[0073] In step S42, for each independent vascular point cloud cluster generated in step S41 , its three-dimensional centroid is obtained by calculating the average value of the three-dimensional coordinates of all pixel points in the cluster. This three-dimensional centroid can represent the central position of the corresponding vascular point cloud cluster in the three-dimensional space. Therefore, the three-dimensional centroid of each independent vascular point cloud cluster can be defined as the vascular center point of each blood vessel in the second vascular region.

[0074] The specific calculation formula for the three-dimensional centroid of each independent vascular point cloud cluster is: , where is each pixel point in the cluster, the three-dimensional coordinates of are ( , , is the vascular center point.

[0075] In some specific embodiments, in step S41, the specific parameters of the DBSCAN clustering algorithm can be set as follows: the neighborhood radius epsilon is 3 pixel units, and the minimum number of neighborhood points min_samples is 5.

[0076] Further, step S5 includes: S51: Arbitrarily select a cubic region in the three-dimensional point cloud model, where the cubic region contains one or more blood vessels; S52: Perform cylindrical fitting on the center points of the blood vessels to obtain the blood vessel width, and calculate the three-dimensional density value of the blood vessels within the cubic region based on the blood vessel centerline and the blood vessel width; S53: Classify the blood vessels according to the three-dimensional density value.

[0077] Among them, for the defined cubic region, as the spatial range for calculating the blood vessel density value, it can be selected manually or automatically. When selecting manually, the operator can delimit a cubic region in the three-dimensional point cloud model through the interactive interface, and this cubic region needs to be able to contain one or more blood vessels to be analyzed. When selecting automatically, the size of the cubic region can be preset, and then in the three-dimensional point cloud model, the cubic region can be slid with a certain step length, or a cubic region can be automatically generated with the blood vessel center point as the center.

[0078] Perform cylindrical fitting on each independent blood vessel point cloud cluster. Blood vessels usually present an approximately cylindrical structure in three-dimensional space. Cylindrical fitting is a method for estimating cylindrical parameters using point cloud data. Specifically, by performing cylindrical fitting on each independent blood vessel point cloud cluster, the cylindrical parameters that best fit the point cloud cluster can be obtained, such as the axis direction, radius, etc. Among them, the radius parameter obtained by cylindrical fitting can be used to characterize the width of the blood vessel (i.e., the diameter of the blood vessel). Thus, the cylindrical radius obtained by performing cylindrical fitting on each independent blood vessel point cloud cluster can be defined as the blood vessel width of each blood vessel in the second blood vessel region. In some specific embodiments, the cylindrical fitting algorithm can adopt the RANSAC algorithm, with the number of iterations set to 100 times and the inlier threshold set to 1 pixel unit. For example, for an independent blood vessel point cloud cluster, through DBSCAN clustering and cylindrical fitting calculations, the centerline of the blood vessel can be obtained as a series of three-dimensional point coordinates, the center point of the blood vessel as a three-dimensional coordinate value, and the blood vessel width as 0.5 mm. Through the setting of the above specific parameters and algorithm selection, accurate calculation of the blood vessel width can be achieved.

[0079] Specifically, in step S53, the three-dimensional density value can be calculated based on the Monte Carlo integration method: , where is the number of blood vessels within the cubic region, is the diameter of a single blood vessel, is the centerline length of a single blood vessel, is the volume of the cubic region, is the serial number of the blood vessel, k = 1, 2, 3... N.

[0080] For grading blood vessels according to three-dimensional density values, a mapping relationship between density values and blood vessel grades can be preset. For example, it can be set that the density value ρ ∈ [0, 0.2) is grade I, the density value ρ ∈ [0.2, 0.4) is grade II, the density value ρ ∈ [0.4, 0.6) is grade III, and the density value ρ ≥ 0.6 is grade IV. Then, according to the calculated three-dimensional density value, the corresponding blood vessel grade can be found.

[0081] In summary, through the 3D laparoscopic binocular stereoscopic vision system, the present application fundamentally overcomes the limitations of the traditional 2D solution: directly obtaining the true physical size of blood vessels (millimeter-level accuracy) based on physical ranging (binocular baseline distance + triangulation), replacing the estimation method relying on pixel distance under the traditional 2D laparoscope, and eliminating the magnification error caused by the change of the lens distance; and calculating the spatial blood vessel density through three-dimensional point cloud reconstruction and Monte Carlo integration method, solving the problem of density underestimation caused by blood vessel overlap and perspective tilt in the 2D projection method, and having the advantages of improving the accuracy and reliability of blood vessel hyperplasia evaluation.

[0082] Please refer to Figure 2 , a device for quantifying and grading blood vessel hyperplasia, which is applied to the steps of any of the above methods. The device includes: Image acquisition module 201: used to acquire a stereoscopic image of the lesion area, and the stereoscopic image at least includes left-eye image information and right-eye image information; Point cloud construction module 202: used to perform stereo correction on the left-eye image information and the right-eye image information to obtain a disparity map between the two images, and construct a three-dimensional point cloud model according to the disparity map; Blood vessel segmentation module 203: used to segment the first blood vessel area according to the corrected left-eye image information and convert the first blood vessel area into the corresponding second blood vessel area in the three-dimensional point cloud model; Parameter calculation module 204: used to calculate the blood vessel geometric parameters of the second blood vessel area, and the blood vessel geometric parameters at least include: blood vessel centerline, blood vessel center point; Blood vessel grading module 205: used to calculate the blood vessel width according to the blood vessel center point, calculate the three-dimensional density value of the blood vessel according to the blood vessel centerline and the blood vessel width, and grade the blood vessel according to the three-dimensional density value.

[0083] Among them, the image acquisition module 201 can specifically be a binocular camera system. The binocular camera system is placed in parallel to simulate human eye vision, and synchronously acquires the left-eye image information and the right-eye image information of the lesion area from different angles to provide data support for subsequent stereoscopic vision processing.

[0084] The point cloud construction module 202, where the stereo calibration specifically uses the Bouguet algorithm. The Bouguet algorithm corrects the distortion of the original images through calibration parameters and aligns the two images so that the pixel points of the two images are on the same row, simplifying the subsequent disparity calculation. The disparity map is obtained specifically by the semi-global matching algorithm to calculate the disparity values of the pixel points. The construction of the three-dimensional point cloud model is specifically based on the disparity values of each pixel point in the disparity map and the pre-calibrated camera parameters, and uses the principle of triangulation to calculate the three-dimensional spatial coordinates of the pixel points, thereby constructing the three-dimensional point cloud model of the lesion area.

[0085] The blood vessel segmentation module 203 performs CLAHE contrast enhancement on the corrected left-eye image information. CLAHE contrast enhancement can improve the local contrast of the image, making the blood vessel structure clearer. For semantic segmentation, the U-Net network is specifically used. The U-Net network is a deep learning network that can effectively segment the blood vessel area in the image to obtain the binary mask of the blood vessels. Morphological opening operation is used to remove the noise points and small connected areas in the binary mask of the blood vessels to obtain a more accurate pixel coordinate set of the first blood vessel area. The conversion of the blood vessel area to the three-dimensional point cloud model specifically uses the normalized cross-correlation algorithm to accurately locate the pixel coordinate set of the first blood vessel area to obtain the two-dimensional blood vessel pixel coordinates, and then through a three-dimensional matrix transformation, the two-dimensional blood vessel pixel coordinates are converted into the three-dimensional point cloud space to obtain the three-dimensional blood vessel point cloud subset, and the three-dimensional blood vessel point cloud subset is the second blood vessel area.

[0086] The parameter calculation module 204 performs DBSCAN clustering calculation on the three-dimensional blood vessel point cloud subset of the second blood vessel area, clustering the point clouds belonging to the same blood vessel into a cluster to form multiple independent blood vessel point cloud clusters, and each independent blood vessel point cloud cluster represents the blood vessel center line of a blood vessel. The blood vessel center point is obtained by calculating the three-dimensional centroid of each independent blood vessel point cloud cluster.

[0087] The blood vessel grading module 205. The selection of the cube region can be manually specified or automatically selected to include the region containing the blood vessels. The least squares method is used to perform cylindrical fitting on each independent blood vessel point cloud cluster to obtain the cylindrical parameters closest to the blood vessel shape, and the cylindrical diameter is the blood vessel width. It is obtained by calculating the ratio of the sum of the volumes of all blood vessels in the cube region to the volume of the cube region. The volume of the blood vessel is calculated as being proportional to the square of the blood vessel width and the length of the blood vessel center line. For blood vessel grading, the mapping relationship between the three-dimensional density value and the blood vessel hyperplasia grade is preset, and according to the calculated three-dimensional density value, the mapping relationship table is searched to obtain the blood vessel hyperplasia grade.

[0088] Specifically, when the vascular proliferation quantification and grading device proposed in this application is working, first, the image acquisition module 201 acquires a stereoscopic image of the lesion area. The stereoscopic image contains left-eye image information and right-eye image information, providing the original data for subsequent processing. Then, the point cloud construction module 202 receives the stereoscopic image, performs stereoscopic correction on the left-eye image information and the right-eye image information, eliminates image distortion and aligns the images. After that, it calculates the disparity map of the corrected image. The disparity map reflects the depth information of the image pixel points, and constructs a three-dimensional point cloud model of the lesion area based on the disparity map, realizing the conversion from a two-dimensional image to a three-dimensional space. Next, the blood vessel segmentation module 203 receives the corrected left-eye image information, segments the first blood vessel area in the left-eye image information. The first blood vessel area is the pixel set of the blood vessels in the two-dimensional image, and converts the first blood vessel area to the corresponding second blood vessel area in the three-dimensional point cloud model. The second blood vessel area is the point cloud set of the blood vessels in the three-dimensional space, realizing the three-dimensional positioning of the blood vessel area. The parameter calculation module 204 calculates the geometric parameters of the blood vessels in the second blood vessel area, obtaining key parameters of the blood vessels such as the blood vessel centerline and the blood vessel center point, providing a basis for subsequent density calculation and grading. Finally, the blood vessel grading module 205 calculates the blood vessel width based on the blood vessel center point, and then calculates the three-dimensional density value of the blood vessels according to the blood vessel centerline and the blood vessel width. The three-dimensional density value comprehensively reflects the quantity and thickness of the blood vessels, and grades the blood vessels according to the three-dimensional density value, realizing the quantitative evaluation of vascular proliferation.

[0089] Through the collaborative work of the above modules, the device proposed in this application can automatically, efficiently, and accurately achieve vascular proliferation quantification and grading. Compared with traditional methods, the device proposed in this application conducts blood vessel analysis based on a three-dimensional point cloud model, overcomes the defect of the lack of depth information in traditional two-dimensional image analysis methods, can more accurately obtain characteristic parameters such as blood vessel density, and thus more objectively and accurately evaluate the degree of vascular proliferation.

[0090] Please refer to Figure 3 , Figure 3A schematic structural diagram of an electronic device provided by an embodiment of the present application. The present application provides an electronic device 3, including: a processor 301 and a memory 302. The processor 301 and the memory 302 are interconnected and communicate with each other through a communication bus 303 and / or other forms of connection mechanisms (not marked). The memory 302 stores computer-readable instructions executable by the processor 301. When the electronic device runs, the processor 301 executes the computer-readable instructions to execute the method in any optional implementation manner of the above embodiment to implement the following functions: obtaining a stereoscopic image of a lesion area, where the stereoscopic image at least includes left-eye image information and right-eye image information; performing stereoscopic correction on the left-eye image information and the right-eye image information to obtain a disparity map between the two images, and constructing a three-dimensional point cloud model according to the disparity map; segmenting a first blood vessel area according to the corrected left-eye image information and converting the first blood vessel area into a corresponding second blood vessel area in the three-dimensional point cloud model; calculating blood vessel geometric parameters of the second blood vessel area, where the blood vessel geometric parameters at least include: a blood vessel center line and a blood vessel center point; calculating a blood vessel width according to the blood vessel center point, calculating a three-dimensional density value of the blood vessel according to the blood vessel center line and the blood vessel width, and grading the blood vessel according to the three-dimensional density value.

[0091] An embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it executes the method in any optional implementation manner of the above embodiment to implement the following functions: obtaining a stereoscopic image of a lesion area, where the stereoscopic image at least includes left-eye image information and right-eye image information; performing stereoscopic correction on the left-eye image information and the right-eye image information to obtain a disparity map between the two images, and constructing a three-dimensional point cloud model according to the disparity map; segmenting a first blood vessel area according to the corrected left-eye image information and converting the first blood vessel area into a corresponding second blood vessel area in the three-dimensional point cloud model; calculating blood vessel geometric parameters of the second blood vessel area, where the blood vessel geometric parameters at least include: a blood vessel center line and a blood vessel center point; calculating a blood vessel width according to the blood vessel center point, calculating a three-dimensional density value of the blood vessel according to the blood vessel center line and the blood vessel width, and grading the blood vessel according to the three-dimensional density value.

[0092] In the embodiments provided by the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For another example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some communication interfaces. The indirect coupling or communication connection of the devices or units can be in an electrical, mechanical or other form.

[0093] In addition, the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed over multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0094] Furthermore, in each embodiment of the present application, the various functional modules may be integrated together to form an independent part, or each module may exist alone, or two or more modules may be integrated to form an independent part.

[0095] In this text, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations.

[0096] The above description is only for the embodiments of the present application and is not intended to limit the protection scope of the present application. For those skilled in the art, the present application may have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for quantifying and grading angiogenesis, characterized in that, The method includes: S1: Obtain a stereoscopic image of the lesion area, where the stereoscopic image includes at least left-eye image information and right-eye image information; S2: Perform stereo calibration on the left-eye image information and the right-eye image information to obtain a disparity map between the two images, and construct a three-dimensional point cloud model based on the disparity map; S3: Segment a first blood vessel area according to the calibrated left-eye image information, and convert the first blood vessel area to a corresponding second blood vessel area in the three-dimensional point cloud model; S4: Calculate the blood vessel geometric parameters of the second blood vessel area, where the blood vessel geometric parameters include at least: the blood vessel centerline and the blood vessel center point; S5: Calculate the blood vessel width according to the blood vessel center point, calculate the three-dimensional density value of the blood vessel according to the blood vessel centerline and the blood vessel width, and classify the blood vessel according to the three-dimensional density value.

2. The method for quantifying and grading angiogenesis according to claim 1, wherein Step S2 includes: S21: Obtain calibration parameters according to the left-eye image information and the right-eye image information; S22: Perform epipolar calibration on the left-eye image information and the right-eye image information according to the calibration parameters to eliminate lens distortion; S23: Calculate the disparity map between the left-eye image information and the right-eye image information after eliminating lens distortion; S24: Construct the three-dimensional point cloud model according to the disparity map.

3. The method for quantifying and grading angiogenesis according to claim 2, wherein Step S24 includes: S241: Obtain the disparity value of each pixel point in the left-eye image information and the right-eye image information according to the disparity map; S242: Obtain the focal length, baseline distance, and noise compensation term of the camera; S243: Calculate the depth value of each pixel point according to the focal length, the baseline distance, the noise compensation term, and the disparity value; S244: Convert the depth value to a depth map, and construct the three-dimensional point cloud model according to the depth map.

4. A method for quantifying and grading angiogenesis according to claim 1, characterized in that Step S3 includes: S31: Perform CLAHE contrast enhancement on the calibrated left-eye image information; S32: Perform semantic segmentation on the enhanced left-eye image information to extract the binary mask of the blood vessel; S33: Perform morphological opening operation on the binary mask of the blood vessel to obtain the pixel coordinate set of the first blood vessel area; S34: Corresponding the pixel coordinate set of the first blood vessel area to the three-dimensional point cloud model to obtain the second blood vessel area.

5. A method for quantifying and grading angiogenesis according to claim 4, characterized in that Step S34 includes: S341: Perform positioning calculation on the pixel coordinate set of the first blood vessel area based on the normalized cross-correlation algorithm to obtain the two-dimensional blood vessel pixel coordinates of the pixel coordinate set of the first blood vessel area; S342: Perform three-dimensional matrix transformation on the two-dimensional blood vessel pixel coordinates to obtain a three-dimensional blood vessel point cloud subset, and the position of the three-dimensional blood vessel point cloud subset in the three-dimensional point cloud model is the second blood vessel area.

6. A method for quantifying and grading angiogenesis according to claim 5, characterized in that, Step S4 includes: S41: Perform DBSCAN clustering calculation on the binary mask to generate multiple independent blood vessel point cloud clusters, and the independent blood vessel point cloud clusters are the blood vessel centerlines of each blood vessel in the second blood vessel area; S42: Calculate the three-dimensional centroid of each independent blood vessel point cloud cluster to obtain the blood vessel center point of each blood vessel in the second blood vessel area.

7. A method for quantifying and grading angiogenesis according to claim 6, characterized in that, Step S5 includes: S51: Arbitrarily select a cubic region in the three-dimensional point cloud model, where the cubic region contains one or more blood vessels; S52: Perform cylindrical fitting on the center points of the blood vessels to obtain the blood vessel widths, and calculate the three-dimensional density values of the blood vessels in the cubic region according to the blood vessel centerlines and the blood vessel widths; S53: Classify the blood vessels according to the three-dimensional density values.

8. A device for quantifying and grading angiogenesis, characterized in that, Applied to the steps of the method according to any one of claims 1-7 above, the device includes: An image acquisition module: configured to acquire a stereoscopic image of a lesion area, where the stereoscopic image at least includes left-eye image information and right-eye image information; A point cloud construction module: configured to perform stereoscopic correction on the left-eye image information and the right-eye image information to obtain a disparity map between the two images, and construct a three-dimensional point cloud model according to the disparity map; A blood vessel segmentation module: configured to segment a first blood vessel area according to the corrected left-eye image information, and convert the first blood vessel area into a corresponding second blood vessel area in the three-dimensional point cloud model; A parameter calculation module: configured to calculate blood vessel geometric parameters of the second blood vessel area, where the blood vessel geometric parameters at least include: blood vessel centerlines, blood vessel center points; A blood vessel classification module: configured to calculate the blood vessel widths according to the blood vessel center points, calculate the three-dimensional density values of the blood vessels according to the blood vessel centerlines and the blood vessel widths, and classify the blood vessels according to the three-dimensional density values.

9. An electronic device, characterized in that, Including a processor and a memory, where the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the steps in the method according to any one of claims 1-7 are run.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the steps in the method according to any one of claims 1-7 are run.

Images