Transverse process positioning method, apparatus, computer equipment and storage medium

Abstract

This application relates to the field of medical imaging technology, and more particularly to a method, apparatus, computer device, and storage medium for locating transverse processes. The method includes: scanning the center of the transverse process of a target object from multiple directions to obtain first coordinates for each transverse process center in each direction; performing principal component analysis and cluster analysis based on each of the first coordinates to obtain a cluster center for each vertebra; determining the target center closest to the transverse process from the cluster centers based on second coordinates of each transverse process in the ultrasound image of the target object and the coordinates of the cluster centers; both the first and second coordinates are three-dimensional coordinates; and identifying the vertebra corresponding to the target center as the vertebra where the transverse process is located. This method enables precise location of the vertebra where the transverse process is located.

Description

Technical Field

[0001] This application relates to the field of medical imaging technology, and in particular to a transverse process localization method, apparatus, computer device, and storage medium. Background Technology

[0002] With the development of medical technology, image registration technology has emerged. Image registration technology is mainly used to merge images from different imaging modalities, thereby providing richer information to help doctors make diagnoses. For example, when merging images from different imaging modalities, image registration is required based on the common features in the images from different imaging modalities.

[0003] When performing image registration for vertebral regions, the location information of the transverse processes in ultrasound images and medical images can be used for image registration. However, since ultrasound probes cannot simultaneously detect the transverse processes of all vertebrae, it is impossible to determine the vertebra where the transverse process detected by the ultrasound probe is located. Summary of the Invention

[0004] Therefore, it is necessary to provide a transverse protrusion positioning method, device, computer equipment, and storage medium that can accurately locate the aforementioned technical problems.

[0005] In a first aspect, this application provides a method for locating transverse processes, the method comprising:

[0006] The transverse process center of the target object is scanned from multiple directions to obtain the first coordinate of each transverse process center in each direction;

[0007] Principal component analysis and cluster analysis were performed based on each of the first coordinates to obtain the cluster center of each vertebra;

[0008] Based on the second coordinates of each transverse process in the ultrasound image of the target object and the coordinates of the cluster center, the target center closest to the transverse process is determined from the cluster center; both the first coordinates and the second coordinates are three-dimensional coordinates;

[0009] The vertebra corresponding to the target center is identified as the vertebra where the transverse process is located.

[0010] In one embodiment, scanning the transverse process centers of the target object from multiple directions to obtain the first coordinates of each transverse process center in each of the directions includes:

[0011] Determine the center of the transverse process of the target object;

[0012] The transverse process center is scanned from multiple directions using an ultrasonic probe to obtain the two-dimensional coordinates of each transverse process center in each direction.

[0013] The two-dimensional coordinates of each transverse process center in each direction are converted into the first coordinates in electromagnetic space.

[0014] In one embodiment, determining the center of the transverse process of each vertebra of the target object includes:

[0015] The first transverse process image of multiple transverse processes is segmented from the ultrasound image of the vertebral region of the target object;

[0016] The first transverse process image is binarized to obtain a binarized image, and the edge contours of each transverse process in the binarized image are identified.

[0017] The centroid of the edge contour is determined as the center of the transverse process of the target vertebra, or a circumscribed shape containing the edge contour is determined, and the center of the circumscribed shape is determined as the center of the transverse process of the target vertebra.

[0018] In one embodiment, the step of performing principal component analysis and cluster analysis based on each of the first coordinates to obtain the cluster center of each vertebra includes:

[0019] Based on each of the first coordinates, a data matrix is ​​constructed, and the data matrix is ​​centered to obtain a centered matrix;

[0020] Calculate the covariance matrix of the centered matrix, and calculate the eigenvalues ​​of the covariance matrix and the eigenvectors corresponding to the eigenvalues;

[0021] Using the eigenvector corresponding to the largest eigenvalue as the projection axis, each first coordinate is projected onto the projection axis to obtain the third coordinate after projection of each first coordinate.

[0022] Cluster analysis was performed based on the aforementioned third coordinates to obtain the cluster center for each vertebra.

[0023] In one embodiment, the method further includes:

[0024] The cluster centers are sorted according to their coordinates on the principal axis of the projection.

[0025] Based on the sorting results of each cluster center, the vertebrae corresponding to each cluster center are determined.

[0026] In one embodiment, determining the target center closest to the transverse process from the cluster centers based on the second coordinates of each transverse process in the ultrasound image of the target object and the coordinates of the cluster centers includes:

[0027] Second transverse process images of multiple transverse processes are segmented from ultrasound images of the vertebral region of the target object;

[0028] Determine the two-dimensional coordinates of each transverse process in the second transverse process image, and convert the two-dimensional coordinates into second coordinates in electromagnetic space;

[0029] The second coordinate is centered to obtain the centered second coordinate;

[0030] The centered second coordinate is projected onto the projection principal axis to obtain the fourth coordinate of each transverse protrusion after projection.

[0031] Based on each of the fourth coordinates, the target center that is closest to each of the transverse processes is determined from the cluster centers.

[0032] In one embodiment, the method further includes:

[0033] The third transverse process image is segmented from the medical image of the vertebral region of the target object, and the vertebrae in which each transverse process is located in the third transverse process image are determined.

[0034] The transverse processes in the same vertebra are identified as a pair of transverse processes, based on the second and third transverse process images.

[0035] Based on the three-dimensional coordinate information of each transverse process in the transverse process alignment, the ultrasound image and the medical image are registered to obtain a registration matrix;

[0036] Based on the registration matrix, a cross-sectional image of the ultrasound image in the three-dimensional reconstruction model corresponding to the medical image is determined, so as to perform puncture guidance based on the cross-sectional image and the ultrasound image.

[0037] Secondly, this application provides a transverse process positioning device, the device comprising:

[0038] The scanning module is used to scan the transverse process center of the target object from multiple directions to obtain the first coordinates of each transverse process center in each of the directions;

[0039] The analysis module is used to perform principal component analysis and cluster analysis based on each of the first coordinates to obtain the cluster center of each vertebra.

[0040] The center determination module is used to determine the target center closest to the transverse process from the cluster centers based on the second coordinates of each transverse process in the ultrasound image of the target object and the coordinates of the cluster centers; both the first coordinates and the second coordinates are three-dimensional coordinates;

[0041] The positioning module is used to identify the vertebra corresponding to the target center as the vertebra where the transverse process is located.

[0042] Thirdly, this application provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described method.

[0043] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0044] The aforementioned transverse process localization method, device, computer equipment, and storage medium scan the transverse process center of the target object from multiple directions to obtain the first coordinates of each transverse process center in each direction. Principal component analysis and cluster analysis are then performed based on these first coordinates to obtain the cluster center of each vertebra. This avoids errors caused by a single viewpoint, allowing for more accurate determination of the cluster center of each vertebra. By using the second coordinates of each transverse process in the ultrasound image of the target object and the coordinates of the cluster centers, the target center closest to the transverse process is determined from the cluster centers. Since both the first and second coordinates are three-dimensional coordinates, the vertebra corresponding to the target center is identified as the vertebra containing the transverse process. This reduces data dimensionality while retaining the main features of the data, thereby improving the localization accuracy of the transverse process and enabling precise localization of the vertebra containing the transverse process. Attached Figure Description

[0045] Figure 1 This is a diagram illustrating the application environment of the transverse process localization method in one embodiment.

[0046] Figure 2 This is a flowchart illustrating a transverse process localization method in one embodiment;

[0047] Figure 3 This is a flowchart illustrating the lumbar spine arrangement in one embodiment;

[0048] Figure 4 This is a structural block diagram of the transverse protrusion positioning device in one embodiment;

[0049] Figure 5 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0051] The transverse process positioning method provided in this application embodiment can be applied to, for example... Figure 1The application environment shown is illustrated. Console 102 and terminal 104 are connected. Console 102 can be a device or a system for locating transverse processes. Console 102 scans the center of the transverse process of the target object from multiple directions, obtaining the first coordinates of each transverse process center in each direction. Console 102 performs principal component analysis and cluster analysis based on each first coordinate to obtain the cluster center of each vertebra. Console 102 determines the target center closest to the transverse process from the cluster centers based on the second coordinates of each transverse process in the ultrasound image of the target object and the coordinates of the cluster centers. Both the first and second coordinates are three-dimensional coordinates. Console 102 identifies the vertebra corresponding to the target center as the vertebra where the transverse process is located.

[0052] In one embodiment, such as Figure 2 As shown, a transverse process localization method is provided, which can be applied to... Figure 1 Taking the console in the browser as an example, the following steps are included:

[0053] S202, scan the transverse process center of the target object from multiple directions to obtain the first coordinate of each transverse process center in each direction.

[0054] The transverse process center refers to the center point of each transverse process in the vertebral region of the target object. The transverse process center can be obtained by identifying images of the vertebral region of the target object. The vertebral region can be the lumbar vertebrae or the thoracic vertebrae. If the vertebral region is the lumbar vertebrae, the transverse process center is the center point of each transverse process in the lumbar region; if the vertebral region is the thoracic vertebrae, the transverse process center is the center point of each rib in the thoracic region.

[0055] Scanning the transverse process center can be achieved using an ultrasound probe. As the ultrasound probe moves, the coordinates of the same transverse process center will deviate on different scanning planes. Therefore, scanning the transverse process center of the target object from multiple directions can reduce errors caused by scanning.

[0056] The first coordinate is a three-dimensional coordinate in electromagnetic space. Each transverse process center has a first coordinate in each direction; these first coordinates may or may not be the same in each direction. After scanning the transverse process center with an ultrasound probe, the two-dimensional coordinates of the transverse process center are first obtained, and then the two-dimensional coordinates are converted into three-dimensional first coordinates using a transformation matrix. The transformation matrix is ​​the transformation matrix from the ultrasound plane to electromagnetic coordinates.

[0057] Optionally, the console controls the ultrasound probe to scan the transverse process center of the target object from multiple directions to obtain the first coordinates of each transverse process center in each direction.

[0058] S204. Principal component analysis and cluster analysis are performed based on each first coordinate to obtain the cluster center of each vertebra.

[0059] Principal component analysis (PCA) is a statistical process that transforms a set of potentially correlated variables into a set of linearly independent variables through orthogonal transformations. These new variables are called principal components. Principal components are linear combinations of the original variables and are ordered by variance; that is, the first principal component has the largest variance, the second has the second largest, and so on. The main purpose of PCA is dimensionality reduction, i.e., reducing the dimensionality of the dataset while preserving as much variation information as possible from the original data. This helps simplify data analysis, improve computational efficiency, and also remove noise and redundant information.

[0060] Cluster analysis is a technique used to divide a dataset into multiple groups or clusters, such that objects within the same cluster are more similar to objects in other clusters to some extent. Clustering algorithms include, but are not limited to, K-means clustering and hierarchical clustering.

[0061] A vertebra is part of the spine of a target object. Vertebrae can be lumbar or thoracic. Lumbar vertebrae include the vertebral body, vertebral arch, and extending from the vertebral arch, such as transverse processes, spinous processes, superior articular processes, and inferior articular processes. Thoracic vertebrae include the vertebral body, pedicles, laminae, vertebral foramina, spinous processes, articular processes, and transverse processes. The cluster center of each vertebra can be determined based on the coordinate sorting results of each cluster center; that is, the vertebra corresponding to each cluster center is determined based on the coordinate sorting results of each cluster center.

[0062] Optionally, the console performs principal component analysis based on each first coordinate, and then performs cluster analysis based on the results of the principal component analysis to obtain the cluster center of each vertebra.

[0063] S206, Based on the second coordinates of each transverse process in the ultrasound image of the target object and the coordinates of the cluster center, determine the target center that is closest to the transverse process from the cluster center; both the first and second coordinates are three-dimensional coordinates.

[0064] The second coordinate is a three-dimensional coordinate in electromagnetic space. After the ultrasound probe scans the center of the transverse process, the two-dimensional coordinates of the transverse process can be obtained first, and then the two-dimensional coordinates are converted into the three-dimensional form of the second coordinate through a transformation matrix.

[0065] Each transverse process corresponds to a target center. For example, in an ultrasound image, there are transverse processes 1, 2, and 3. After calculation, it is found that cluster center 1 is closest to transverse process 1, cluster center 2 is closest to transverse process 2, and cluster center 3 is closest to transverse process 3. Therefore, cluster center 1 is the target center of transverse process 1, cluster center 2 is the target center of transverse process 2, and cluster center 3 is the target center of transverse process 3.

[0066] Optionally, the console calculates the distance between each transverse process and each cluster center based on the second coordinates of each transverse process and the coordinates of the cluster centers in the ultrasound image of the target object, and determines the cluster center closest to the transverse process as the target center corresponding to the transverse process.

[0067] S208, the vertebra corresponding to the target center is determined to be the vertebra where the transverse process is located.

[0068] The human body has five lumbar vertebrae, namely the first, second, third, fourth, and fifth lumbar vertebrae, which are connected sequentially as follows: Figure 3 As shown, L1 is the first lumbar vertebra, L2 is the second lumbar vertebra, L3 is the third lumbar vertebra, L4 is the fourth lumbar vertebra, and L5 is the fifth lumbar vertebra. The human body has a total of 12 thoracic vertebrae, designated as T1 (first thoracic), T2 (second thoracic), T3 (third thoracic), T4 (fourth thoracic), T5 (fifth thoracic), T6 (sixth thoracic), T7 (seventh thoracic), T8 (eighth thoracic), T9 (ninth thoracic), T10 (tenth thoracic), T11 (eleventh thoracic), and T12 (twelfth thoracic). The vertebra containing the transverse process refers to its position relative to the target vertebra, such as the lumbar or thoracic vertebra.

[0069] The vertebra corresponding to the target center is the same as the vertebra corresponding to the cluster center. For example, continuing the example above, the vertebra corresponding to cluster center 1 is the vertebra containing transverse process 1, the vertebra corresponding to cluster center 2 is the vertebra containing transverse process 2, and the vertebra corresponding to cluster center 3 is the vertebra containing transverse process 3.

[0070] Optionally, the console can determine the vertebrae corresponding to each target center as the vertebrae containing the transverse process.

[0071] In the aforementioned transverse process localization method, the transverse process centers of the target object are scanned from multiple directions to obtain the first coordinates of each transverse process center in each direction. Principal component analysis and cluster analysis are then performed based on these first coordinates to obtain the cluster centers of each vertebra. This avoids errors caused by a single viewpoint, allowing for more accurate determination of the cluster centers of each vertebra. By using the second coordinates of each transverse process in the ultrasound image of the target object and the coordinates of the cluster centers, the target center closest to the transverse process is determined from the cluster centers. Since both the first and second coordinates are three-dimensional coordinates, the vertebra corresponding to the target center is identified as the vertebra containing the transverse process. This reduces data dimensionality while preserving the main features of the data, thereby improving the localization accuracy of the transverse processes and accurately locating the vertebrae containing each transverse process.

[0072] In one embodiment, scanning the transverse process centers of the target object from multiple directions to obtain the first coordinates of each transverse process center in each direction includes:

[0073] Determine the center of the transverse process of the target object.

[0074] The center of each transverse process is scanned from multiple directions using an ultrasound probe to obtain the two-dimensional coordinates of each transverse process center in each direction.

[0075] The two-dimensional coordinates of each transverse process center in each direction are converted into the first coordinates in electromagnetic space.

[0076] The transverse process center is the center point of the transverse process. The transverse process center can be obtained from a target image including each vertebra of the target object, or it can be obtained by the operator through manual identification of the target image, and the result of manual identification is sent to the console.

[0077] The transformation of the two-dimensional coordinates of the transverse process center into the first coordinates in electromagnetic space can be achieved using a transformation matrix. Specifically, the two-dimensional coordinates of the transverse process center are: P ( x , y ), T w The transformation matrix from the ultrasonic plane to the electromagnetic space is obtained through... The two-dimensional coordinates of the transverse process center can be converted into a three-dimensional first coordinate. x w , y w , z w ).

[0078] Optionally, the console determines the transverse process center of the target object and scans each transverse process center of the target object from multiple directions using an ultrasound probe to obtain the two-dimensional coordinates of each transverse process center in each direction. The console obtains the transformation matrix from the ultrasound plane to electromagnetic space and uses the transformation matrix to convert the two-dimensional coordinates of each transverse process center in each direction into the first coordinates in electromagnetic space.

[0079] In this embodiment, by scanning the center of each transverse process from multiple directions and converting the two-dimensional coordinates into the first coordinates in electromagnetic space, the center of the transverse process can be accurately located in three-dimensional space.

[0080] In one embodiment, determining the center of the transverse process of each vertebra of the target object includes:

[0081] The first transverse process image of multiple transverse processes is segmented from the ultrasound image of the vertebral region of the target object;

[0082] The first transverse process image is binarized to obtain a binarized image, and the edge contours of each transverse process in the binarized image are identified.

[0083] The centroid of the edge contour is determined as the center of the transverse process of the target vertebra, or the circumscribed graph containing the edge contour is determined, and the center of the circumscribed graph is determined as the center of the transverse process of the target vertebra.

[0084] The ultrasound image is obtained by scanning the vertebral region of the target object with an ultrasound probe. Methods for segmenting the first transverse process image from the ultrasound image include, but are not limited to, thresholding, edge detection, region generation, and model-based segmentation. A combination of these methods can also be used. Thresholding segmentation divides pixels in the image into several parts by setting one or more thresholds, thus separating the first transverse process image from other tissues in the ultrasound image based on pixel values. Edge detection uses edge detection algorithms to identify the edge contours of the transverse process in the image, thereby locating and segmenting the first transverse process image. Model-based methods use a trained segmentation model to segment the first transverse process image from the ultrasound image.

[0085] Binarization is an image processing technique used to convert grayscale or color images into images with only black and white values, meaning each pixel's value becomes either 0 or 255. Binarization is typically based on a threshold. Specifically, if a pixel's grayscale value is greater than the threshold, the pixel is set to 255; if the pixel's grayscale value is less than or equal to the threshold, it is set to 0. The edge contours of each protrusion in the binarized image can be identified using image processing techniques.

[0086] The centroid of the edge contour can be determined based on the average coordinates of all points within the contour. Circumscribed figures include, but are not limited to, circumscribed circles and circumscribed rectangles.

[0087] In some embodiments, the centroid of the edge contour can be determined as the first center of the target vertebra, a circumscribed graph containing the edge contour can be determined, the center of the circumscribed graph can be determined as the second center of the target vertebra, and the transverse process center of the vertebra can be determined based on the first and second centers. Specifically, the transverse process center of the vertebra is determined based on the average coordinates of the first and second centers.

[0088] Optionally, the console segments multiple transverse process images from the ultrasound image of the vertebrae of the target object using one or more segmentation methods. The console binarizes the segmented first transverse process images to obtain binarized images. The console identifies the edge contours of each transverse process in the binarized image and determines the centroid of the edge contour using the average coordinates of all points within the edge contour. The console determines the centroid of the edge contour as the center of the transverse process of the target object vertebra, or determines the circumscribed shape containing the edge contour and determines the center of the circumscribed shape as the center of the transverse process of the target object vertebra.

[0089] In this embodiment, by binarizing the first transverse protrusion image and identifying the edge contours of each transverse protrusion in the binarized image, the boundary information of the transverse protrusion can be effectively highlighted, the influence of background noise can be reduced, and the range of the transverse protrusion can be defined more clearly and accurately. By determining the centroid of the edge contour or the center of the circumscribed figure as the transverse protrusion center, the transverse protrusion center can be determined for transverse protrusions of different shapes, improving the consistency and accuracy of positioning.

[0090] In one embodiment, principal component analysis and cluster analysis are performed based on each first coordinate to obtain the cluster center of each vertebra, including:

[0091] Based on each first coordinate, a data matrix is ​​constructed, and the data matrix is ​​centered to obtain a centered matrix.

[0092] Calculate the covariance matrix of the centered matrix, and then calculate the eigenvalues ​​and corresponding eigenvectors of the covariance matrix.

[0093] Using the eigenvector corresponding to the largest eigenvalue as the projection principal axis, each first coordinate is projected onto the projection principal axis to obtain the third coordinate after projection of each first coordinate.

[0094] Cluster analysis was performed based on each third coordinate to obtain the cluster center for each vertebra.

[0095] In this data matrix, each row is defined by a first coordinate. Specifically, the data matrix... middle( x n , y n , z n ) indicates the first n The first coordinate.

[0096] The centralization process includes: calculating the average of each column in the data matrix to obtain a mean vector; and calculating the difference between the data matrix and the mean vector to obtain the centralized matrix. Specifically, the centralized matrix... X is a data matrix. It is the mean vector.

[0097] The formula for calculating the covariance matrix is: , C Let covariance matrix be the variance matrix. n The number of first coordinates, X c For a centered matrix, It is the transpose of the centered matrix.

[0098] Eigenvalues ​​and their corresponding eigenvectors are obtained through singular value decomposition of the covariance matrix. Specifically, the singular value decomposition of the covariance matrix... , V It is the eigenvector matrix. It is the transpose of the eigenvector matrix. It is a diagonal matrix containing eigenvalues.

[0099] The formula for calculating the third coordinate is: , V 1 is the projection axis. X c For a centered matrix, X p This is the third coordinate.

[0100] Optionally, the console constructs a data matrix based on multiple first coordinates. X Calculate the average value of each column in the data matrix to obtain the mean vector. Calculate the data matrix X Subtract the mean vector The difference yields the centered matrix. X c The console calculates the covariance matrix of the centered matrix and performs singular value decomposition on the covariance matrix. C The covariance matrix is ​​obtained. C The eigenvalues ​​and their corresponding eigenvectors are given. The console projects the eigenvector corresponding to the largest eigenvalue as the principal axis. V 1. Project each first coordinate onto the projection principal axis. V 1. Obtain the third coordinate after projecting each first coordinate. The console is based on each third coordinate. X p Cluster analysis was performed to obtain the cluster centers for each vertebra.

[0101] In this embodiment, by using the eigenvector corresponding to the largest eigenvalue as the projection principal axis, the main direction of variation in the data can be captured. This helps to filter out noise, emphasize the main pattern of the data, thereby improving the accuracy of the third coordinate estimation and thus improving the accuracy of the cluster center.

[0102] In one embodiment, the transverse process localization method further includes:

[0103] The cluster centers are sorted according to their coordinates on the principal axes of projection.

[0104] Based on the ranking results of each cluster center, the corresponding vertebrae for each cluster center are determined.

[0105] In some embodiments, the direction of the projection principal axis is from the feet to the head of the target object. Since the vertebrae are arranged sequentially from the head to the feet of the target object, using the foot-to-head direction as the projection principal axis, the cluster centers are sorted according to their coordinates on the projection principal axis to determine the vertebra where each cluster center is located, thus improving accuracy. Further, based on the coordinate values ​​of each cluster center on the projection principal axis, they are sorted from low to high to obtain the sorting result of the cluster centers. Cluster centers with smaller coordinate values ​​are located at the bottom, closer to the feet of the target object, while cluster centers with larger coordinate values ​​are located at the top, closer to the head of the target object. Specifically, if the vertebrae are lumbar vertebrae, then the lumbar vertebrae corresponding to each cluster center in the ranking results are the fifth lumbar vertebra, the fourth lumbar vertebra, the third lumbar vertebra, the second lumbar vertebra, and the first lumbar vertebra, respectively. That is, the cluster center ranked first corresponds to the fifth lumbar vertebra, the cluster center ranked second corresponds to the fourth lumbar vertebra, the cluster center ranked third corresponds to the third lumbar vertebra, the cluster center ranked fourth corresponds to the second lumbar vertebra, and the cluster center ranked fifth corresponds to the first lumbar vertebra.

[0106] Optionally, the console sorts the cluster centers from low to high according to their coordinate values ​​on the projection principal axis, thus obtaining the sorting results for each cluster center.

[0107] In this embodiment, by sorting the coordinates of the cluster centers on the projection principal axis, the cluster centers in different vertebral segments can be more clearly identified and distinguished, which helps to avoid confusing the cluster centers of adjacent vertebrae. This allows for accurate guidance of the puncture needle or anesthetic needle during puncture.

[0108] In one embodiment, determining the target center closest to the transverse process from the cluster centers based on the second coordinates of each transverse process in the ultrasound image of the target object and the coordinates of the cluster centers includes:

[0109] Second transverse process images of multiple transverse processes are segmented from ultrasound images of the vertebrae of the target object.

[0110] Determine the two-dimensional coordinates of each transverse process in the second transverse process image, and convert the two-dimensional coordinates into second coordinates in electromagnetic space.

[0111] The second coordinate is centered to obtain the centered second coordinate.

[0112] The centered second coordinate is projected onto the projection principal axis to obtain the fourth coordinate after projection of each transverse process.

[0113] Based on each fourth coordinate, the target center that is closest to each transverse process is determined from the cluster centers.

[0114] Methods for segmenting the second transverse process image from ultrasound images include, but are not limited to, thresholding, edge detection, region generation, and model-based segmentation. Segmentation can also be a combination of these methods. Thresholding segmentation divides pixels in the image into several parts by setting one or more thresholds, thus separating the second transverse process image from other tissues in the ultrasound image based on pixel values. Edge detection uses edge detection algorithms to identify the edge contours of the transverse process in the image, locating and segmenting the second transverse process image based on these contours. Model-based methods use a trained segmentation model to segment the second transverse process image from the ultrasound image.

[0115] Two-dimensional coordinates are obtained by scanning with an ultrasound probe. After scanning the transverse process with the ultrasound probe, the two-dimensional coordinates of the transverse process can be obtained first, and then the two-dimensional coordinates are converted into a three-dimensional second coordinate form through a transformation matrix. The transformation matrix is ​​the transformation matrix from the ultrasound plane to electromagnetic space.

[0116] The centralization process includes: determining the target data matrix composed of each second coordinate, calculating the average value of each column in the target data matrix to obtain the target mean vector, and calculating the difference between the target data matrix and the target mean vector to obtain the centralized second coordinates.

[0117] In some embodiments, the first coordinates can be directly centered to obtain centered first coordinates, and then projected onto the projection principal axis to obtain the fourth coordinates of each transverse process after projection. This can save time in obtaining the second coordinates and improve the efficiency of transverse process positioning.

[0118] Optionally, the console segments multiple transverse process images from the ultrasound image of the vertebral region of the target object. The console determines the two-dimensional coordinates of each transverse process in the second transverse process image and converts the two-dimensional coordinates into second coordinates in electromagnetic space. The console centers the second coordinates to obtain centered second coordinates. The console projects the centered second coordinates onto the projection principal axis to obtain the fourth coordinates of each transverse process after projection; the console calculates the distance between each fourth coordinate and each cluster center, and determines the target center closest to each transverse process from the cluster centers based on the calculation results.

[0119] In this embodiment, the target center closest to each transverse process is determined from the cluster centers previously obtained through cluster analysis based on each fourth coordinate. This method not only improves the accuracy of transverse process localization, but also effectively identifies and distinguishes different vertebral segments, thus maintaining high reliability even in complex anatomical environments.

[0120] In one embodiment, the transverse process localization method further includes:

[0121] The third transverse process image is segmented from the medical images of the vertebrae of the target object, and the vertebrae in which each transverse process is located are identified in the third transverse process image.

[0122] Transverse processes in the same vertebra are identified as a pair of transverse processes, based on images of the second and third transverse processes.

[0123] Based on the three-dimensional coordinate information of each transverse process in the midline, the ultrasound image and medical image are registered to obtain the registration matrix.

[0124] Based on the registration matrix, the cross-sectional image of the ultrasound image in the corresponding three-dimensional reconstruction model of the medical image is determined, so as to guide the puncture based on the cross-sectional image and the ultrasound image.

[0125] Medical imaging refers to static image data of the vertebral region of a target object at a specific moment. Medical imaging includes, but is not limited to, computed tomography (CT) imaging, magnetic resonance imaging (MRI), and positron emission tomography (PET). Methods for segmenting the third transverse process from medical images include, but are not limited to, thresholding, edge detection, region generation, and model segmentation. Furthermore, methods for segmenting the third transverse process from medical images can also be combinations of thresholding, edge detection, and model segmentation.

[0126] Because the detection range of an ultrasound probe is limited, ultrasound images will not show all the transverse processes of the target vertebrae. Medical imaging, however, involves a comprehensive scan of the target vertebrae, capturing all the transverse processes. Therefore, the third transverse process image will show more transverse processes than the first and second images. For example, the human body has five lumbar vertebrae, each with one transverse process. An ultrasound probe can only detect three consecutive lumbar vertebrae, while magnetic resonance imaging (MRI) can scan all five. Thus, the first transverse process image will show three transverse processes, and the second image will show five.

[0127] In some embodiments, since the third transverse process image may include each transverse process of the vertebral region of the target object, the transverse processes in the third transverse process image can be sorted according to their three-dimensional coordinate information, and the vertebrae where each transverse process is located in the third transverse process image can be determined based on the sorting result. Specifically, the transverse processes are sorted according to their coordinate values ​​on the target axis, and the vertebrae where each transverse process is located in the third transverse process image are determined according to the sorting result. The direction of the target axis is from the foot to the head of the target object.

[0128] In some embodiments, the vertebrae in which each transverse process in the third transverse process image is located can also be determined based on the morphological characteristics of the transverse process and / or adjacent structures. For example, in the lumbar region, the transverse process of the third lumbar vertebra is usually relatively long, being the longest transverse process among all lumbar vertebrae; therefore, the longest transverse process in the third transverse process image can be identified as the transverse process of the third lumbar vertebra. The transverse process of the fifth lumbar vertebra is usually relatively robust and extends laterally, forming the lumbosacral joint with the sacrum; therefore, the relatively robust transverse process extending laterally in the third transverse process image can be identified as the transverse process of the fifth lumbar vertebra. The transverse process of the fifth lumbar vertebra connects to the sacrum, forming the lumbosacral angle; therefore, the transverse process connected to the sacrum can also be identified as the transverse process of the fifth lumbar vertebra.

[0129] A transverse process pair refers to transverse processes belonging to the same vertebra in the second and third transverse process images. For example, if transverse process A in the second transverse process image is a transverse process in the first vertebra, and transverse process B in the third transverse process image is also a transverse process in the first vertebra, then transverse processes A and B are in the same vertebra and belong to a transverse process pair.

[0130] Registration algorithms include, but are not limited to, point cloud registration algorithms. A registration matrix can be used to convert 3D coordinate information in ultrasound images into 3D coordinate information in medical images. The inverse of the registration matrix can be used to convert 3D coordinate information in medical images into 3D coordinate information in ultrasound images, thus enabling the fusion of ultrasound images and medical images.

[0131] The three-dimensional reconstruction model corresponding to medical images refers to the three-dimensional reconstruction model constructed using multiple consecutive frames of medical images.

[0132] In some embodiments, the process of acquiring a 3D reconstruction model includes: preprocessing multiple frames of medical images to obtain preprocessed medical images; and using a 3D reconstruction algorithm to perform 3D reconstruction on the preprocessed medical images to obtain a 3D reconstruction model. The data preprocessing includes at least one of denoising, image registration, and standardization. Denoising refers to reducing noise in the medical images; standardization refers to adjusting the grayscale range in the medical images; and image registration refers to registering multiple frames of medical images to the same coordinate system. The 3D reconstruction methods include, but are not limited to, voxel interpolation and surface reconstruction methods.

[0133] The process of acquiring the cross-sectional image includes: based on the registration matrix, converting the coordinates of the ultrasound image into transformed coordinates in the three-dimensional reconstruction model, and based on the transformed coordinates, determining the cross-sectional image of the ultrasound image in the three-dimensional reconstruction model. Specifically, the image containing the transformed coordinates in the three-dimensional reconstruction model is determined as the cross-sectional image.

[0134] In some embodiments, the console can also fuse cross-sectional images and ultrasound images to obtain a fused image, which can be used for puncture guidance based on the fused image, cross-sectional image, and ultrasound image to improve puncture accuracy.

[0135] Optionally, the console uses a trained segmentation model to segment the third transverse process image from medical images of the vertebrae of the target object. Based on the 3D coordinates of each transverse process in the third transverse process image, the console sorts the transverse processes in the third transverse process image, and determines the vertebrae where each transverse process is located based on the sorting result. The console identifies the transverse processes in the same vertebra from the second and third transverse process images as transverse process pairs, and registers the ultrasound image and medical image based on the 3D coordinates of each transverse process in the pair, obtaining a registration matrix. Based on the registration matrix, the console converts the coordinates of the ultrasound image into transformed coordinates in the 3D reconstruction model. The console then searches for the cross-sectional image corresponding to the transformed coordinates in the 3D reconstruction model. The console performs puncture guidance based on the cross-sectional image and the ultrasound image.

[0136] In this embodiment, the ultrasound image and medical image are registered according to the three-dimensional coordinate information of each transverse process, resulting in a registration matrix. Based on the registration matrix, the cross-sectional image of the ultrasound image in the corresponding three-dimensional reconstruction model of the medical image is determined. This allows for high-precision spatial alignment based on the coordinate information of multiple transverse processes, reducing registration errors and making the registration matrix more accurate. Consequently, in scenarios such as puncture, the ultrasound image and the cross-sectional image determined based on the registration matrix can be used to correctly guide the puncture.

[0137] This application also provides an application scenario in which the transverse process localization method described above is applied. Specifically, the transverse process localization method is applied in this scenario as follows:

[0138] The console uses one or more segmentation methods to segment the first transverse process image from an ultrasound image of the vertebrae of the target object. The console then binarizes the segmented first transverse process image to obtain a binarized image. The console identifies the edge contours of each transverse process in the binarized image and determines the centroid of the edge contour using the average coordinates of all points within the edge contour. The console either assigns the centroid of the edge contour as the center of the transverse process of the target vertebra, or it determines the circumscribed shape containing the edge contour and assigns the center of the circumscribed shape as the center of the transverse process of the target vertebra.

[0139] The control console uses an ultrasound probe to scan the center of each transverse process of the target object from multiple directions, obtaining the two-dimensional coordinates of each transverse process center in each direction. The control console acquires the transformation matrix from the ultrasound plane to electromagnetic space, and uses the transformation matrix to convert the two-dimensional coordinates of each transverse process center in each direction into the first coordinates in electromagnetic space.

[0140] The console constructs a data matrix based on multiple first coordinates. X Calculate the average value of each column in the data matrix to obtain the mean vector. Calculate the data matrix XSubtract the mean vector The difference yields the centered matrix. X c The console calculates the covariance matrix of the centered matrix and performs singular value decomposition on the covariance matrix. C The covariance matrix is ​​obtained. C The eigenvalues ​​and their corresponding eigenvectors are given. The console projects the eigenvector corresponding to the largest eigenvalue as the principal axis. V 1. Project each first coordinate onto the projection principal axis. V 1. Obtain the third coordinate after projecting each first coordinate. The console is based on each third coordinate. X p Cluster analysis is performed to obtain cluster centers. The console sorts the cluster centers from low to high according to their coordinates on the projection principal axis, obtaining the sorting results of each cluster center. Based on the sorting results, the corresponding vertebra is determined.

[0141] The console segments multiple transverse process images from ultrasound images of the vertebrae of the target object. The console determines the two-dimensional coordinates of each transverse process in the second transverse process image and converts these coordinates into second coordinates in electromagnetic space. The console centers these second coordinates to obtain centered second coordinates. The console projects these centered second coordinates onto the projection principal axis to obtain the projected fourth coordinates of each transverse process. The console calculates the distance between each fourth coordinate and each cluster center, and based on the calculation results, determines the target center closest to each transverse process from the cluster centers.

[0142] The console identifies the vertebrae corresponding to each target center as the vertebrae containing the transverse processes. Using a trained segmentation model, the console segments the third transverse process image from the medical image of the target vertebrae. Based on the 3D coordinates of each transverse process in the third transverse process image, it sorts the transverse processes, determining the vertebrae containing each transverse process in the third transverse process image. The console identifies the transverse processes in the same vertebra from the second and third transverse process images as transverse process pairs. Based on the 3D coordinates of each transverse process in the pair, it registers the ultrasound image and the medical image to obtain a registration matrix. Based on the registration matrix, the console converts the coordinates of the ultrasound image into transformed coordinates in the 3D reconstruction model. The console then locates the cross-sectional image corresponding to the transformed coordinates in the 3D reconstruction model. Finally, the console guides the puncture based on the cross-sectional image and the ultrasound image.

[0143] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0144] Based on the same inventive concept, this application also provides a transverse process positioning device for implementing the transverse process positioning method described above. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more transverse process positioning device embodiments provided below can be found in the limitations of the transverse process positioning method above, and will not be repeated here.

[0145] In one embodiment, such as Figure 4 As shown, a transverse protrusion positioning device is provided, comprising:

[0146] The scanning module 402 is used to scan the transverse process center of the target object from multiple directions to obtain the first coordinates of each transverse process center in each direction.

[0147] Analysis module 404 is used to perform principal component analysis and cluster analysis based on each first coordinate to obtain the cluster center of each vertebra.

[0148] The center determination module 406 is used to determine the target center closest to the transverse process from the cluster center based on the second coordinates of each transverse process in the ultrasound image of the target object and the coordinates of the cluster center; both the first coordinate and the second coordinate are three-dimensional coordinates.

[0149] The positioning module 408 is used to determine the vertebrae corresponding to the target center as the vertebrae where the transverse process is located.

[0150] In one embodiment, the transverse process positioning device is further used to: determine the transverse process center of the target object; scan each transverse process center from multiple directions using an ultrasonic probe to obtain the two-dimensional coordinates of each transverse process center in each direction; and convert the two-dimensional coordinates of each transverse process center in each direction into a first coordinate in electromagnetic space.

[0151] In one embodiment, the transverse process localization device is further configured to: segment a first transverse process image of multiple transverse processes from an ultrasound image of the vertebral region of the target object; binarize the first transverse process image to obtain a binarized image, and identify the edge contour of each transverse process in the binarized image; determine the centroid of the edge contour as the center of the transverse process of the target object vertebra, or determine the circumscribed shape containing the edge contour, and determine the center of the circumscribed shape as the center of the transverse process of the target object vertebra.

[0152] In one embodiment, the transverse process positioning device is further configured to: construct a data matrix based on each first coordinate, and center the data matrix to obtain a centered matrix; calculate the covariance matrix of the centered matrix, and calculate the eigenvalues ​​of the covariance matrix and the eigenvectors corresponding to the eigenvalues; project each first coordinate onto the projection axis using the eigenvector corresponding to the largest eigenvalue as the projection axis to obtain the third coordinate after projection of each first coordinate; and perform cluster analysis based on each third coordinate to obtain the cluster center of each vertebra.

[0153] In one embodiment, the transverse process positioning device is further configured to: sort the cluster centers according to the coordinates of each cluster center on the projection principal axis; and determine the vertebrae corresponding to each cluster center based on the sorting results of each cluster center.

[0154] In one embodiment, the transverse process positioning device is further configured to: segment a second transverse process image from an ultrasound image of the vertebral region of the target object to obtain a plurality of transverse processes; determine the two-dimensional coordinates of each transverse process in the second transverse process image and convert the two-dimensional coordinates into second coordinates in electromagnetic space; center the second coordinates to obtain the centered second coordinates; project the centered second coordinates onto the projection principal axis to obtain the fourth coordinates of each transverse process after projection; and determine the target center closest to each transverse process from the cluster center based on each fourth coordinate.

[0155] In one embodiment, the transverse process localization device is further configured to: segment a third transverse process image from a medical image of the vertebral region of the target object, and determine the vertebrae in which each transverse process in the third transverse process image is located; identify the transverse processes in the same vertebra as a transverse process pair by the second and third transverse process images; register the ultrasound image and the medical image according to the three-dimensional coordinate information of each transverse process in the transverse process pair to obtain a registration matrix; and determine the cross-sectional image of the ultrasound image in the three-dimensional reconstruction model corresponding to the medical image based on the registration matrix, so as to perform puncture guidance based on the cross-sectional image and the ultrasound image.

[0156] Each module in the aforementioned transverse protrusion positioning device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0157] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 5 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database stores first coordinates, cluster centers, second coordinates, target centers, and the vertebrae where the transverse process is located. The network interface is used for communication with external terminals via a network connection. When executed by the processor, the computer program implements a transverse process localization method.

[0158] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0159] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0160] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0161] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0162] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0163] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0164] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0165] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for locating transverse processes, characterized in that, The method includes: The transverse process center of the target object is scanned from multiple directions to obtain the first coordinate of each transverse process center in each direction; Principal component analysis and cluster analysis were performed based on each of the first coordinates to obtain the cluster center of each vertebra; Based on the second coordinates of each transverse process in the ultrasound image of the target object and the coordinates of the cluster center, the target center closest to the transverse process is determined from the cluster center; both the first coordinates and the second coordinates are three-dimensional coordinates; The vertebra corresponding to the target center is identified as the vertebra where the transverse process is located; The step of scanning the transverse process centers of the target object from multiple directions to obtain the first coordinates of each transverse process center in each of the aforementioned directions includes: Determine the center of the transverse process of the target object; The transverse process center is scanned from multiple directions using an ultrasonic probe to obtain the two-dimensional coordinates of each transverse process center in each direction. The two-dimensional coordinates of each transverse process center in each direction are converted into the first coordinates in electromagnetic space; The process of performing principal component analysis and cluster analysis based on each of the first coordinates to obtain the cluster centers for each vertebra includes: Based on each of the first coordinates, a data matrix is ​​constructed, and the data matrix is ​​centered to obtain a centered matrix; Calculate the covariance matrix of the centered matrix, and calculate the eigenvalues ​​of the covariance matrix and the eigenvectors corresponding to the eigenvalues; Using the eigenvector corresponding to the largest eigenvalue as the projection axis, each first coordinate is projected onto the projection axis to obtain the third coordinate after projection of each first coordinate. Cluster analysis was performed based on the aforementioned third coordinates to obtain the cluster center for each vertebra.

2. The method according to claim 1, characterized in that, Determining the transverse process center of the target object includes: The first transverse process image of multiple transverse processes is segmented from the ultrasound image of the vertebral region of the target object; The first transverse process image is binarized to obtain a binarized image, and the edge contours of each transverse process in the binarized image are identified. The centroid of the edge contour is determined as the center of the transverse process of the target vertebra, or a circumscribed shape containing the edge contour is determined, and the center of the circumscribed shape is determined as the center of the transverse process of the target vertebra.

3. The method according to claim 1, characterized in that, The method further includes: The cluster centers are sorted according to their coordinates on the principal axis of the projection. Based on the sorting results of each cluster center, the vertebrae corresponding to each cluster center are determined.

4. The method according to claim 1, characterized in that, The step of determining the target center closest to the transverse process from the cluster centers based on the second coordinates of each transverse process in the ultrasound image of the target object and the coordinates of the cluster centers includes: Second transverse process images of multiple transverse processes are segmented from ultrasound images of the vertebral region of the target object; Determine the two-dimensional coordinates of each transverse process in the second transverse process image, and convert the two-dimensional coordinates into second coordinates in electromagnetic space; The second coordinate is centered to obtain the centered second coordinate; The centered second coordinate is projected onto the projection principal axis to obtain the fourth coordinate of each transverse protrusion after projection. Based on each of the fourth coordinates, the target center that is closest to each of the transverse processes is determined from the cluster centers.

5. The method according to claim 4, characterized in that, The method further includes: The third transverse process image is segmented from the medical image of the vertebral region of the target object, and the vertebrae in which each transverse process is located in the third transverse process image are determined. The transverse processes in the same vertebra are identified as a pair of transverse processes, based on the second and third transverse process images. Based on the three-dimensional coordinate information of each transverse process in the transverse process alignment, the ultrasound image and the medical image are registered to obtain a registration matrix; Based on the registration matrix, a cross-sectional image of the ultrasound image in the three-dimensional reconstruction model corresponding to the medical image is determined, so as to perform puncture guidance based on the cross-sectional image and the ultrasound image.

6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5.

Images