Automatic positioning method, device and equipment for spine anatomical surface, medium and program product

Abstract

The invention provides an automatic positioning method, device and equipment for a spine anatomical surface, a medium and a program product, and relates to the field of medical image processing. The method comprises the steps that based on a pre-trained spine segmentation model, spine segmentation masks of a CT image of a target spine are acquired, and segmentation masks corresponding to different spines in the spine segmentation masks are different; and for each section of spine in different sections of spines, based on a pre-trained spine key point identification model, obtaining key points of the section spines from the segmentation masks corresponding to the section spines, and further, based on the key points, determining spine anatomical surfaces of the section spines. According to the method and the device, each section of spine in the CT image can be automatically identified and segmented by using the pre-trained spine segmentation model, the key point of each section of spine is rapidly extracted from the segmentation mask through the key point identification model, and the anatomical surface of the spine is automatically determined based on the identified key point, so that the accuracy of the spine segmentation is improved. Therefore, the efficiency of positioning the anatomical surface of the spine is obviously improved.

Description

Technical Field

[0001] The present application relates to the field of medical image processing, and in particular to a method, device, equipment, medium and program product for automatically positioning spinal anatomical planes. Background Art

[0002] In modern medical imaging diagnosis, computed tomography (CT) technology has become an important tool for diagnosis and treatment planning of spinal diseases due to its high resolution and three-dimensional reconstruction capabilities. However, how to accurately obtain the anatomical surface of the spine from CT images remains a technical challenge.

[0003] In the related art, when processing CT images, it is usually necessary to perform complex cross-sectional projection operations, and rely on calculating the connected domain area, central moment and bounding box (Bbox) of the segmented image to determine key points, and further locate the anatomical surface based on the key points. However, there is a problem of low efficiency in locating the spinal anatomical surface. Summary of the invention

[0004] The embodiments of the present application provide a method, device, equipment, medium and program product for automatically positioning the anatomical plane of the spine, so as to improve the problem of low efficiency in positioning the anatomical plane of the spine in the related art.

[0005] In a first aspect, an embodiment of the present application provides a method for automatically positioning a spinal anatomical plane, comprising:

[0006] Based on the pre-trained spine segmentation model, a spine segmentation mask of the CT image of the target spine is obtained, wherein different segments of the spine have different corresponding segmentation masks;

[0007] For each vertebra in different vertebrae, based on the pre-trained vertebrae key point recognition model, the key points of the vertebrae are obtained from the segmentation mask corresponding to the vertebrae;

[0008] Based on the key points, the spinal anatomical planes of the vertebra are determined.

[0009] In a possible implementation, based on a pre-trained spinal key point recognition model, the key points of the segmented vertebrae are obtained from the segmentation mask corresponding to the segmented vertebrae, including: inputting the segmentation mask corresponding to the segmented vertebrae into the spinal key point recognition model; generating a set number of heat maps corresponding to the segmented vertebrae based on the segmentation mask corresponding to the segmented vertebrae; for each heat map in the set number of heat maps, among the pixels contained in the heat map, determining the pixel with the largest pixel value as a key point of the segmented vertebrae.

[0010] In a possible embodiment, the spinal anatomical planes include sagittal planes, transverse planes, coronal planes, parasagittal planes, and oblique sagittal planes, and the key points include the center point of the vertebral body, the center point of the posterior edge of the vertebral body, the center point of the junction between the spinous process and the lamina, the center point of the left pedicle, and the center point of the right pedicle. Based on the key points, the spinal anatomical planes of the segmental vertebra are determined, including: based on the center point of the vertebral body, the center point of the posterior edge of the vertebral body, and the center point of the junction between the spinous process and the lamina, the sagittal plane of the segmental vertebra is determined; in a section at right angles to the sagittal plane, the section through the vertebral The section through the center point of the body and the center point of the posterior edge of the vertebra is the cross-section of the segmental vertebra; in the section that is at right angles to both the sagittal plane and the cross-section, the section passing through the center point of the posterior edge of the vertebra is determined to be the coronal plane of the segmental vertebra; in the section parallel to the sagittal plane, the section passing through the center point of the left pedicle and the center point of the right pedicle is determined to be the parasagittal plane of the segmental vertebra; in the section that is at right angles to the cross-section and at a set angle to the coronal plane, the section passing through the center point of the left pedicle and the center point of the right pedicle is determined to be the oblique sagittal plane of the segmental vertebra.

[0011] In a possible embodiment, the automatic positioning method of the spinal anatomical plane also includes: determining a first key point and a second key point on the sagittal plane, the first key point being the midpoint from the center point of the vertebral body along the cross section to the front edge of the vertebral body, and the second key point being the midpoint from the center point of the vertebral body along the cross section to the rear edge of the vertebral body; based on the first key point and the second key point, the spinal anatomical plane is corrected to obtain a corrected spinal anatomical plane.

[0012] In a possible implementation, the spinal anatomical plane is corrected according to the first key point and the second key point to obtain a corrected spinal anatomical plane, including: determining a first straight line passing through the first key point and perpendicular to the cross section, and determining a second straight line passing through the second key point and perpendicular to the cross section; determining the midpoint of the first straight line segment inside the vertebral body as the third key point, and determining the midpoint of the second straight line segment inside the vertebral body as the fourth key point; determining the center point of the junction of the revised spinous process and the lamina according to the third key point and the fourth key point; determining the corrected spinal anatomical plane based on the center point of the junction of the spinous process and the lamina.

[0013] In a possible implementation, the revised center point of the junction between the spinous process and the lamina is determined based on the third key point and the fourth key point, including: connecting the third key point and the fourth key point to obtain a third straight line; determining the intersection point of the third straight line and the posterior edge of the vertebral body as the revised center point of the junction between the spinous process and the lamina.

[0014] In a possible embodiment, a revised spinal anatomical plane is determined based on the revised center point of the junction between the spinous process and the lamina, including: determining the first normal vector of the revised cross-section according to the revised center point of the junction between the spinous process and the lamina; cross-producting the second normal vector of the sagittal plane and the first normal vector to obtain the third normal vector of the revised coronal plane; and constructing the revised spinal anatomical plane according to the first normal vector, the second normal vector and the third normal vector.

[0015] In a second aspect, an embodiment of the present application provides a device for automatically positioning a spinal anatomical plane, comprising:

[0016] A spine segmentation module is used to obtain a spine segmentation mask of a CT image of a target spine based on a pre-trained spine segmentation model, wherein different segments of the spine have different segmentation masks corresponding to them.

[0017] A key point recognition module is used to obtain the key points of each vertebra in different vertebrae from the segmentation mask corresponding to the vertebrae based on a pre-trained vertebrae key point recognition model;

[0018] The processing module is used for determining the spinal anatomical plane of the vertebra based on the key points.

[0019] In a third aspect, the present application provides an electronic device, comprising: a processor, and a memory communicatively connected to the processor;

[0020] Memory for storing computer-executable instructions;

[0021] A processor, configured to execute computer-executable instructions stored in a memory to implement any one of the methods described in the first aspect.

[0022] In a fourth aspect, the present application provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed, they are used to implement any of the methods in the first aspect.

[0023] In a fifth aspect, the present application provides a computer program product, including a computer program, which implements any method in the first aspect when executed.

[0024] The method, apparatus, device, medium and program product for automatically positioning the anatomical plane of the spine provided in the present application obtain a vertebral segmentation mask of a CT image of a target spine based on a pre-trained vertebral segmentation model, wherein the segmentation masks corresponding to different vertebrae in the vertebral segmentation mask are different; for each vertebra in different vertebrae, based on a pre-trained vertebral key point recognition model, the key points of the vertebrae are obtained from the segmentation mask corresponding to the vertebrae; further, based on the key points, the anatomical plane of the vertebrae is determined. In this process, by utilizing the pre-trained spinal segmentation model, each vertebra in the CT image can be automatically identified and segmented, and through the key point recognition model, the key points of each vertebra can be quickly extracted from the segmentation mask, and the anatomical plane of the spine can be automatically determined based on the identified key points, without the need for tedious projection operations and extraction of image feature information, etc., reducing computing costs and shortening processing time, thereby improving the efficiency of spinal anatomical plane positioning; in addition, through automated deep learning technology, the errors caused by human operation are reduced, and even in the case of irregular spine shape or poor image quality, a high positioning accuracy can be maintained, ensuring the consistency and reliability of key point and anatomical plane positioning, thereby providing faster and more accurate image analysis for medical imaging diagnosis, thereby improving the efficiency of clinical diagnosis and treatment. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0026] Figure 1 A schematic diagram of an application scenario of the method for automatically locating spinal anatomical planes provided in an exemplary embodiment of the present application;

[0027] Figure 2 A schematic diagram of a flow chart of a method for automatically locating spinal anatomical planes provided for an exemplary embodiment of the present application;

[0028] Figure 3 A schematic diagram of spine segmentation provided for an exemplary embodiment of the present application;

[0029] Figure 4 A schematic diagram of spine key point identification provided for an exemplary embodiment of the present application;

[0030] Figure 5 A schematic diagram of the recognition results of key points of the spine provided for an exemplary embodiment of the present application;

[0031] Figure 6 A schematic diagram of spinal anatomical planes for determining a vertebra based on key points provided by an exemplary embodiment of the present application;

[0032] Figure 7A schematic diagram of modifying the spinal anatomy provided for an exemplary embodiment of the present application;

[0033] Figure 8 An example of three views of the modified spinal anatomical plane provided for an exemplary embodiment of the present application;

[0034] Fig. 9 Another schematic diagram of a flow chart of a method for automatically locating spinal anatomical planes provided by an exemplary embodiment of the present application;

[0035] Fig.10 A schematic structural diagram of a spinal anatomical surface automatic positioning device provided for an exemplary embodiment of the present application;

[0036] Fig.11 A schematic diagram of the structure of an electronic device provided for an exemplary embodiment of the present application.

[0037] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0038] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0039] The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, and need not be used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable in appropriate circumstances, so that the embodiments of the present application described herein can be implemented in a sequence other than those illustrated or described herein, for example. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, the process, system, product or equipment comprising a series of steps or units need not be limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, products or equipment.

[0040] 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, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0041] First, the terms involved in this application are explained:

[0042] Sagittal plane: A section that divides the spine into two symmetrical parts.

[0043] Transverse plane: A section that is at right angles to the sagittal plane and divides the vertebral body into two symmetrical upper and lower parts.

[0044] Coronal Plane: A plane that is at right angles to the sagittal and transverse planes and divides the spine into anterior and posterior parts.

[0045] Parasagittal Plane: A section that is close to and parallel to the sagittal plane.

[0046] Oblique Sagittal Plane: A section that is at a certain angle to both the transverse and coronal planes.

[0047] Considering that in spinal CT imaging diagnosis, accurate acquisition of spinal anatomical planes is crucial for disease diagnosis and treatment planning. However, traditional CT image processing technology can only obtain intervertebral disc anatomical planes, and the acquisition of spinal anatomical planes mainly relies on manual interception and positioning. This process is not only time-consuming and labor-intensive, but also due to the limitations of human operation, the efficiency of obtaining spinal anatomical planes through manual interception and positioning is low, and it is difficult to ensure the accuracy and consistency of anatomical plane positioning. With the increasing incidence of spinal diseases, the demand for efficient and accurate imaging diagnostic technology is becoming increasingly urgent.

[0048] In the related art, when processing CT images, it is usually necessary to perform complex cross-sectional projection operations, and rely on calculating the connected domain area, central moment, bounding box and other features of the segmented image to determine the key points, and further locate the anatomical surface based on the key points, which leads to the problem of low efficiency in locating the anatomical surface of the spine. In addition, this series of operations not only has high computational cost and long processing time, but also in the case of irregular spine shape or poor image quality, the accuracy of key point positioning is difficult to guarantee, thereby affecting the accuracy of subsequent image analysis and diagnosis.

[0049] In order to solve the above problems, the embodiment of the present application provides a solution for automatic positioning of spinal anatomical planes, which directly uses deep learning models to extract necessary anatomical information from CT images, and quickly and accurately realizes automatic positioning of anatomical planes. Since there is no need to perform tedious projection operations and extract image feature information, the computing cost is reduced and the processing time is shortened, thereby significantly improving the efficiency of spinal anatomical plane positioning. In addition, through automated deep learning technology, not only the errors caused by human operation are reduced, but also a high positioning accuracy can be maintained in the case of irregular spine shape or poor image quality, ensuring the consistency and reliability of key point and anatomical plane positioning, thereby providing faster and more accurate image analysis for medical imaging diagnosis, thereby improving the efficiency of clinical diagnosis and treatment.

[0050] Figure 1 Schematic diagram of an application scenario of the method for automatically locating the spinal anatomical plane provided by the exemplary embodiment of the present application. Figure 1 As shown, the application scenario includes a client 11 and a server 12, wherein the number of the client 11 can be at least one. In practical applications, when the server 12 detects the CT image of the target spine submitted by the client 11, it executes the automatic positioning method of the spinal anatomical plane provided in the present application to determine the spinal anatomical plane of each vertebra of the target spine.

[0051] It should be noted that the server 12 may also be replaced by a server cluster or other computing devices with certain computing power. The client 11 may be a mobile phone, a computer, a notebook or a personal digital assistant (PDA).

[0052] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0053] Figure 2 A schematic diagram of a flow chart of a method for automatically locating a spinal anatomical plane provided by an exemplary embodiment of the present application. Figure 2 As shown, the method for automatically locating the vertebral anatomical plane comprises the following steps:

[0054] S201. Based on a pre-trained spine segmentation model, a spine segmentation mask of a CT image of a target spine is obtained, wherein different segments of the spine segmentation mask correspond to different segmentation masks.

[0055] For example, a large number of annotated spinal CT image datasets are collected, where each CT image should clearly mark the location and name of each vertebra (such as lumbar vertebrae L1, L2 or thoracic vertebrae T12, etc.), and the dataset should cover a variety of spinal morphologies and image qualities to enhance the generalization ability of the model; for example, a 3D U-Net convolutional neural network is selected as the basic model. This model is suitable for processing complex three-dimensional medical image segmentation tasks with its encoding, decoding and jump connection network structure; the annotated spinal CT image dataset is input into the 3D U-Net model for training, and the recognition and segmentation capabilities of the spinal region are gradually improved by learning the data features in the CT images, and iterative optimization methods such as the cross entropy loss function and the Adam optimizer are used to adjust the model parameters to improve the segmentation accuracy. In addition, during the training process, data enhancement (such as rotation and scaling) can also be performed to improve the robustness of the model; the model performance is tested on an independent validation dataset, and the accuracy of the model is evaluated by comparing the segmentation results output by the model with the actual annotations. Further, based on the validation results, the model hyperparameters and network structure are adjusted to optimize the segmentation effect.

[0056] It can be understood that the end of the model training process is similar to that in the related art. The spinal segmentation model training is considered completed after reaching a set number of training times, or a trained spinal segmentation model is obtained when the loss value is less than a set threshold.

[0057] Correspondingly, a CT image of the target spine is obtained, and the pre-processed CT image is input into the trained spine segmentation model; the spine segmentation model automatically processes the input CT image of the target spine to generate a segmentation mask of the target spine, wherein the model output includes a segmentation mask and a corresponding label (e.g., lumbar vertebrae L1, L2 or thoracic vertebrae T12, etc.) for each vertebra, and the segmentation masks of different vertebrae are displayed in different colors or labels in the image. For example, Figure 3 A schematic diagram of spine segmentation provided for an exemplary embodiment of the present application. Figure 3 (a) is the CT image of the target spine. Figure 3 (b) in the figure is the corresponding spine segmentation mask, in which the segmentation masks corresponding to different spine sections have different colors, and different colors correspond to different labels.

[0058] Optionally, before inputting the CT image of the target spine into the pre-trained spine segmentation model, the CT image of the target spine may be pre-processed to adapt to the input requirements of the spine segmentation model, wherein the pre-processing operation includes but is not limited to normalization or resizing.

[0059] S202 , for each vertebra in different vertebrae, based on a pre-trained vertebrae key point recognition model, obtain the key points of the vertebrae from the segmentation mask corresponding to the vertebrae.

[0060] For example, collect and prepare a dataset of segmental vertebrae segmentation masks containing annotated key points, and ensure that the dataset covers the morphology and image quality of multiple segmental vertebrae segmentation masks to improve the generalization ability of the model; select a deep learning model suitable for three-dimensional medical image processing, such as 3D U-Net or V-Net, as the basic model. The selected basic model must have strong feature extraction capabilities and be suitable for identifying complex anatomical structures; use the annotated dataset to train the model, gradually improve the recognition ability of key points, and use loss functions (such as mean square error) and optimization algorithms (such as Adam optimizer) to adjust the model parameters to improve the positioning accuracy of key points. In addition, during the training process, data enhancement (such as rotation and scaling) can also be performed to improve the robustness of the model; further, test the model performance on an independent validation dataset, evaluate the accuracy of the model by comparing the key point positions output by the model with the actual annotations, and adjust the model hyperparameters and network structure based on the validation results to optimize the key point recognition effect.

[0061] It can be understood that the end of the model training process is similar to that in the related art. The spinal key point recognition model training is considered completed after reaching the set number of training times, or the trained spinal key point recognition model is obtained when the loss value is less than the set threshold.

[0062] Correspondingly, for each vertebra in different vertebrae, the segmentation mask corresponding to the vertebrae is input into the vertebral key point recognition model; further, the vertebral key point recognition model automatically processes the segmentation mask corresponding to the input vertebrae, identifies and outputs the key points of the vertebrae. Among them, the key points are used to describe the geometric features of the vertebral body, and different key points are displayed in the image with different marks to facilitate subsequent image analysis.

[0063] S203: Determine the spinal anatomical plane of the vertebra based on the key points.

[0064] For example, based on the geometric relationship between key points, the spatial position and direction of the spinal anatomical plane are calculated and determined; and by analyzing the relative position of the key points, the optimal cutting path of each anatomical plane is derived, thereby obtaining the spinal anatomical plane of each vertebra. The spinal anatomical plane can provide a comprehensive analysis perspective of the spinal structure, which has positive significance for the diagnosis and evaluation of medical images.

[0065] The automatic positioning method of the anatomical plane of the spine provided in the embodiment of the present application can automatically identify and segment each vertebral segment in the CT image by utilizing a pre-trained vertebral segmentation model. Through the key point recognition model, the key points of each vertebral segment can be quickly extracted from the segmentation mask, and the anatomical plane of the spine can be automatically determined based on the identified key points, without the need for tedious projection operations and extraction of image feature information, etc., thereby reducing computing costs and shortening processing time, thereby improving the efficiency of anatomical plane positioning of the spine; through automated deep learning technology, the errors caused by human operation are reduced, and even in the case of irregular spine shape or poor image quality, a high positioning accuracy can be maintained, ensuring the consistency and reliability of key point and anatomical plane positioning, thereby providing faster and more accurate image analysis for medical imaging diagnosis, thereby improving the efficiency of clinical diagnosis and treatment.

[0066] In some embodiments, based on a pre-trained spinal key point recognition model, the key points of the segmented vertebrae are obtained from the segmentation mask corresponding to the segmented vertebrae, including: inputting the segmentation mask corresponding to the segmented vertebrae into the spinal key point recognition model; generating a set number of heat maps corresponding to the segmented vertebrae based on the segmentation mask corresponding to the segmented vertebrae; for each heat map in the set number of heat maps, among the pixels contained in the heat map, determining the pixel with the largest pixel value as a key point of the segmented vertebrae.

[0067] For example, assume that the number is set to 5; correspondingly, the segmentation mask corresponding to the vertebral segment is input into the spine key point recognition model, and based on the segmentation mask corresponding to the vertebral segment, 5 heat maps corresponding to the vertebral segment are generated, and the 5 heat maps correspond to different areas of the vertebral segment respectively; for each of the 5 heat maps, among the pixels contained in the heat map, the pixel with the largest pixel value is determined as a key point of the vertebral segment; based on the 5 heat maps, the 5 key points of the vertebral segment can be determined. For example, Figure 4 The schematic diagram of spine key point recognition provided by the exemplary embodiment of the present application. Figure 4 As shown, among the pixels contained in a heat map, the pixel with the largest pixel value is determined as a key point of the spine.

[0068] It should be noted that the above-mentioned setting number of 5 is only an example, and the setting number can also be 10, 15 or 20, etc., and the value of the setting number is not limited here.

[0069] In the embodiment of the present application, by generating a set number of heat maps and using the probability distribution characteristics of the heat maps to identify the pixel with the largest pixel value in each heat map, the key points of the spine can be accurately located. This identification method can adapt to CT images of different shapes and qualities, and can maintain high positioning accuracy and robustness even in the case of irregular spine shapes or poor image quality, thereby improving the accuracy of key point identification. In addition, the entire process of key point identification is automatically completed by the model, reducing the need for manual intervention and effectively reducing human errors. Moreover, by using heat maps to locate key points, complex calculation steps are reduced, and processing time is significantly shortened, thereby improving the efficiency and consistency of spinal key point identification.

[0070] In some embodiments, the spinal anatomical planes include sagittal planes, transverse planes, coronal planes, parasagittal planes, and oblique sagittal planes, and the key points include the center point of the vertebral body, the center point of the posterior edge of the vertebral body, the center point of the junction between the spinous process and the lamina, the center point of the left pedicle, and the center point of the right pedicle. Based on the key points, the spinal anatomical planes of the segmental vertebra are determined, including: determining the sagittal plane of the segmental vertebra based on the center point of the vertebral body, the center point of the posterior edge of the vertebral body, and the center point of the junction between the spinous process and the lamina; determining, in a section at right angles to the sagittal plane, the center point passing through the center of the vertebral body; The section through the center point of the posterior edge of the vertebral body is the cross-section of the segmental vertebra; in the section that is at right angles to both the sagittal plane and the cross-section, the section passing through the center point of the posterior edge of the vertebral body is determined to be the coronal plane of the segmental vertebra; in the section parallel to the sagittal plane, the section passing through the center point of the left pedicle and the center point of the right pedicle is determined to be the parasagittal plane of the segmental vertebra; in the section that is at right angles to the cross-section and at a set angle to the coronal plane, the section passing through the center point of the left pedicle and the center point of the right pedicle is determined to be the oblique sagittal plane of the segmental vertebra.

[0071] For example, Figure 5 The following is a schematic diagram of the recognition results of the spine key points provided by the exemplary embodiment of the present application. Figure 5 As shown, Figure 5 (a) Figure 5 (b) and Figure 5 (c) in the figure shows the positions of key points from different perspectives, and distinguishes different key points by different symbols; among them, P1 represents the center point of the vertebral body, P2 represents the center point of the posterior edge of the vertebral body, P3 represents the center point of the junction of the spinous process and the lamina, P4 represents the center point of the left pedicle, and P5 represents the center point of the right pedicle.

[0072] For example, Figure 6 The schematic diagram of the spinal anatomical plane of the vertebra determined based on key points provided by the exemplary embodiment of the present application. Figure 6 As shown in (a) of FIG. 1 , the sagittal plane of the segmental vertebra can be determined by dividing the segmental vertebra into left-right symmetrical parts through the center point P1 of the vertebral body, the center point P2 of the posterior edge of the vertebral body, and the center point P3 of the junction of the spinous process and the lamina; Figure 6As shown in (b), in the section at right angles to the sagittal plane, the section passing through the center point P1 of the vertebral body and the center point P2 of the posterior edge of the vertebral body is determined as the cross section of the segmental vertebra, and the cross section divides the segmental vertebra into upper and lower parts; Figure 6 As shown in (c) in the figure, among the sections that are perpendicular to the sagittal and transverse planes, the section that passes through the center point P2 of the posterior edge of the vertebral body is determined to be the coronal plane of the segmental vertebra. The coronal plane divides the segmental vertebra into two parts, anterior and posterior. Figure 6 As shown in (d) in the section parallel to the sagittal plane, the section passing through the center point P4 of the left pedicle and the center point P5 of the right pedicle is determined as the parasagittal plane of the vertebra; Figure 6 As shown in (e), in the section that is at right angles to the cross section and at a set angle, such as 45°, to the coronal plane, the section passing through the center point of the left pedicle P4 and the center point of the right pedicle P5 is determined to be the oblique sagittal plane of the segmental vertebra.

[0073] It should be noted that the above-mentioned setting angle of 45° is only an example. In actual applications, the setting angle can be set to 40°, 45° or 50°, etc. as needed, and is not limited here.

[0074] The embodiments of the present application utilize key points to automatically locate the spinal anatomical planes, thereby avoiding the limitations of manual operation, and not only improving the efficiency of spinal anatomical plane positioning, but also improving the accuracy and robustness of spinal anatomical plane positioning; in addition, by locating the sagittal plane, transverse plane, coronal plane, parasagittal plane and oblique sagittal plane, a multi-angle and multi-dimensional observation perspective of the spine is provided, which helps to better present the structure of the spine; through efficient, accurate and comprehensive spinal anatomical plane positioning, important technical support can be provided for the clinical diagnosis, treatment and research of the spine.

[0075] Normally, the cross section obtained by automatic positioning of key points can evenly divide the vertebral body of the segmental vertebra into two symmetrical upper and lower parts. However, when there is a certain angle between the upper and lower end plate planes of the vertebral body, or there is a certain error in the identified key points, this cross section may not be able to achieve symmetrical segmentation. In order to improve the accuracy of the cross section so that it can accurately divide the vertebral body into two symmetrical upper and lower parts, it is also necessary to make corresponding corrections to the spinal anatomical surface according to the segmentation mask of the segmental vertebra. Therefore, in some embodiments, the automatic positioning method of the spinal anatomical surface also includes: determining a first key point and a second key point on the sagittal plane, the first key point being the midpoint from the center point of the vertebral body along the cross section to the front edge of the vertebral body, and the second key point being the midpoint from the center point of the vertebral body along the cross section to the rear edge of the vertebral body; according to the first key point and the second key point, the spinal anatomical surface is corrected to obtain a corrected spinal anatomical surface.

[0076] For example, Figure 7 The schematic diagram of the correction of the spinal anatomy provided by the exemplary embodiment of the present application. Figure 7As shown, the sagittal plane of the spine passing through the center point P1 of the vertebral body is determined as the correction reference plane, and two key points are determined on the correction reference plane, that is, the midpoint from the center point P1 of the vertebral body along the cross section to the front edge of the vertebral body is determined as the first key point P6, and the midpoint from the center point P1 of the vertebral body along the cross section to the rear edge of the vertebral body is determined as the second key point P7. Further, according to the first key point P6 and the second key point P7, the spinal anatomical plane is corrected, and the correction process includes: using the first key point P6 and the second key point P7, adjusting the positioning and direction of the spinal anatomical plane to ensure that it more accurately reflects the true anatomical structure of the segmental vertebra.

[0077] In the embodiment of the present application, the accuracy of automatic positioning of the spinal anatomical plane is improved by determining the first key point and the second key point on the sagittal plane and correcting the spinal anatomical plane based on the key points.

[0078] On the basis of the above embodiments, in some embodiments, the spinal anatomical plane is corrected according to the first key point and the second key point to obtain a corrected spinal anatomical plane, including: determining a first straight line passing through the first key point and perpendicular to the cross section, and determining a second straight line passing through the second key point and perpendicular to the cross section; determining the midpoint of the first straight line in the line segment inside the vertebral body as the third key point, and determining the midpoint of the second straight line in the line segment inside the vertebral body as the fourth key point; determining the center point of the junction of the revised spinous process and the lamina according to the third key point and the fourth key point; determining the corrected spinal anatomical plane based on the center point of the junction of the spinous process and the lamina.

[0079] For example, still refer to Figure 7 , extract the bisector of the upper and lower endplate planes, specifically, make a first straight line passing through the first key point P6 and perpendicular to the cross section, and make a second straight line passing through the second key point P7 and perpendicular to the cross section; determine the midpoint of the first straight line in the line segment inside the vertebral body as the third key point P6', and determine the midpoint of the second straight line in the line segment inside the vertebral body as the fourth key point P7'; accordingly, use the positional relationship between the third key point P6' and the fourth key point P7' to derive the geometric center point of the junction of the spinous process and the lamina to ensure that the center point can accurately reflect the anatomical characteristics of the segmental vertebra, thereby determining the revised center point P3' of the junction of the spinous process and the lamina; further, use the revised center point P3' of the junction of the spinous process and the lamina as the origin of the revised anatomical surface to determine the revised spinal anatomical surface.

[0080] In some embodiments, the revised center point of the junction between the spinous process and the lamina is determined based on the third key point and the fourth key point, including: connecting the third key point and the fourth key point to obtain a third straight line; determining the intersection point of the third straight line and the posterior edge of the vertebral body as the revised center point of the junction between the spinous process and the lamina.

[0081] For example, still refer to Figure 7, connect the third key point P6' and the fourth key point P7' to obtain a third straight line; determine the intersection of the third straight line and the posterior edge of the vertebral body as the corrected center point P3' of the junction of the spinous process and the lamina, that is, the corrected origin of the anatomical surface.

[0082] Furthermore, in some embodiments, a revised spinal anatomical plane is determined based on the revised center point of the junction between the spinous process and the lamina, including: determining the first normal vector of the revised cross-section according to the revised center point of the junction between the spinous process and the lamina; cross-producting the second normal vector of the sagittal plane and the first normal vector to obtain the third normal vector of the revised coronal plane; and constructing the revised spinal anatomical plane according to the first normal vector, the second normal vector and the third normal vector.

[0083] For example, based on the center point of the junction of the corrected spinous process and the lamina, the first normal vector of the corrected cross section is determined. The normal vector is usually perpendicular to the cross section of the segmental vertebra. For example, it can be accurately calculated through a mathematical model to ensure the accuracy of the first normal vector; the sagittal plane remains unchanged, and its normal vector is determined to be the second normal vector; and through a cross multiplication operation, the second normal vector of the sagittal plane is cross-multiplied with the first normal vector of the corrected cross section, and the result of the cross multiplication is the third normal vector of the corrected coronal plane, and the third normal vector is used to define the direction of the coronal plane and ensure that it is perpendicular to other anatomical planes; further, according to the first normal vector, the second normal vector and the third normal vector, three mutually perpendicular planes are determined, namely the corrected cross section, the sagittal plane and the corrected coronal plane, so as to construct a complete corrected three-dimensional spinal anatomical plane.

[0084] For example, Figure 8 An example of three views of the modified spinal anatomical plane provided by the exemplary embodiment of the present application. Among them, Figure 8 (a) is an example of a corrected cross section. Figure 8 (b) is an example of the corrected coronal plane. Figure 8 (c) is an example of sagittal plane; Figure 8 In (a), it can be observed that the sagittal plane accurately divides the spine into two symmetrical parts. Figure 8 In (b), it can be observed that the modified cross section divides the vertebra into two symmetrical parts. Figure 8 In (c), it can be observed that the corrected coronal plane divides the spine into two parts, anterior and posterior, and the upper and lower end plates of the spine are in a horizontal position; the corrected three-view spinal anatomical plane can present the spinal anatomical plane more accurately.

[0085] The embodiments of the present application further improve the accuracy of automatic positioning of the spinal anatomical planes by correcting the spinal anatomical planes, thereby ensuring the accurate positioning of the spinal anatomical planes in space, which helps to perform medical analysis or surgical planning more accurately; and accurate spinal anatomical plane positioning can also be used to calculate the screw implantation angle, and has positive significance for assisting doctors in performing spinal fusion surgery and improving the safety and effectiveness of spinal fusion surgery.

[0086] Fig. 9 Another flow chart of the method for automatically locating the spinal anatomical plane provided by the exemplary embodiment of the present application. Fig. 9 As shown, the method for automatically locating the vertebral anatomical plane comprises the following steps:

[0087] S901. Based on a pre-trained spine segmentation model, a spine segmentation mask of a CT image of a target spine is obtained, wherein different segments of the spine segmentation mask correspond to different segmentation masks.

[0088] For example, a CT image of the target spine is obtained and input into a trained spine segmentation model; the spine segmentation model automatically processes the input CT image to generate a segmentation mask of the target spine, wherein the model output includes a segmentation mask and a corresponding label (e.g., lumbar vertebrae L1, L2 or thoracic vertebrae T12, etc.) for each vertebra, and the segmentation masks of different vertebrae are displayed in the image with different colors or marks.

[0089] S902. For each vertebra in different vertebrae, based on a pre-trained vertebral key point recognition model, the key points of the vertebrae are obtained from the segmentation mask corresponding to the vertebrae. The key points include the center point of the vertebral body, the center point of the posterior edge of the vertebral body, the center point of the junction of the spinous process and the lamina, the center point of the left pedicle, and the center point of the right pedicle.

[0090] S903. Determine the spinal anatomical planes of the vertebra based on the key points. The spinal anatomical planes include the sagittal plane, the transverse plane, the coronal plane, the parasagittal plane, and the oblique sagittal plane.

[0091] For example, Figure 6 As shown in (a) of FIG. 1 , the sagittal plane of the segmental vertebra can be determined by dividing the segmental vertebra into left-right symmetrical parts through the center point P1 of the vertebral body, the center point P2 of the posterior edge of the vertebral body, and the center point P3 of the junction of the spinous process and the lamina; Figure 6 As shown in (b), in the section at right angles to the sagittal plane, the section passing through the center point P1 of the vertebral body and the center point P2 of the posterior edge of the vertebral body is determined as the cross section of the segmental vertebra, and the cross section divides the segmental vertebra into upper and lower parts; Figure 6 As shown in (c) in the figure, among the sections that are perpendicular to the sagittal and transverse planes, the section that passes through the center point P2 of the posterior edge of the vertebral body is determined to be the coronal plane of the segmental vertebra. The coronal plane divides the segmental vertebra into two parts, anterior and posterior. Figure 6As shown in (d) in the section parallel to the sagittal plane, the section passing through the center point P4 of the left pedicle and the center point P5 of the right pedicle is determined as the parasagittal plane of the vertebra; Figure 6 As shown in (e), in the section that is at right angles to the cross section and at a set angle, such as 45°, to the coronal plane, the section passing through the center point of the left pedicle P4 and the center point of the right pedicle P5 is determined to be the oblique sagittal plane of the segmental vertebra.

[0092] S904. Determine a first key point and a second key point on the sagittal plane, the first key point being the midpoint from the center point of the vertebral body along the cross section to the front edge of the vertebral body, and the second key point being the midpoint from the center point of the vertebral body along the cross section to the rear edge of the vertebral body.

[0093] For example, Figure 7 As shown, the sagittal plane of the spine passing through the vertebral center point P1 is determined as the correction reference plane, and two key points are determined on the correction reference plane, that is, the midpoint from the vertebral center point P1 along the cross section to the front edge of the vertebral body is determined as the first key point P6, and the midpoint from the vertebral center point P1 along the cross section to the rear edge of the vertebral body is determined as the second key point P7.

[0094] S905, determine a first straight line passing through the first key point and perpendicular to the cross section, and determine a second straight line passing through the second key point and perpendicular to the cross section; determine the midpoint of the first straight line in the line segment inside the vertebral body as the third key point, and determine the midpoint of the second straight line in the line segment inside the vertebral body as the fourth key point.

[0095] For example, refer to Figure 7 , draw a first straight line passing through the first key point P6 and perpendicular to the cross section, and draw a second straight line passing through the second key point P7 and perpendicular to the cross section; determine the midpoint of the first straight line segment inside the vertebral body as the third key point P6', and determine the midpoint of the second straight line segment inside the vertebral body as the fourth key point P7'.

[0096] S906, connecting the third key point and the fourth key point to obtain a third straight line; determining the intersection point of the third straight line and the posterior edge of the vertebral body as the corrected center point of the junction between the spinous process and the lamina.

[0097] Still refer to Figure 7 , connect the third key point P6' and the fourth key point P7' to obtain a third straight line; determine the intersection of the third straight line and the posterior edge of the vertebral body as the corrected center point P3' of the junction of the spinous process and the lamina, that is, the corrected origin of the anatomical surface.

[0098] S907. Determine a revised spinal anatomical plane based on the revised center point of the junction between the spinous process and the lamina.

[0099] Specifically, the first normal vector of the revised cross section is determined according to the center point of the intersection of the spinous process and the lamina; the second normal vector of the sagittal plane and the first normal vector are cross-multiplied to obtain the third normal vector of the revised coronal plane; and the revised spinal anatomical plane is constructed according to the first normal vector, the second normal vector and the third normal vector. For example, the three views of the revised spinal anatomical plane are as follows: Figure 8 shown.

[0100] In summary, this application has at least the following advantages:

[0101] 1. By utilizing the pre-trained spinal segmentation model, each vertebra in the CT image can be automatically identified and segmented. Through the key point recognition model, the key points of each vertebra can be quickly extracted from the segmentation mask, and the anatomical plane of the spine can be automatically determined based on the identified key points, without the need for tedious projection operations and extraction of image feature information, etc., reducing computing costs and shortening processing time, thereby improving the efficiency of spinal anatomical plane positioning; through automated deep learning technology, the errors caused by human operation are reduced, and even in the case of irregular spine shape or poor image quality, a high positioning accuracy can be maintained, ensuring the consistency and reliability of key point and anatomical plane positioning, thereby providing faster and more accurate image analysis for medical imaging diagnosis, thereby improving the efficiency of clinical diagnosis and treatment.

[0102] 2. By using key points to automatically locate the anatomical planes of the spine, the limitations of manual operation are avoided, which not only improves the efficiency of anatomical plane positioning of the spine, but also improves the accuracy and robustness of anatomical plane positioning of the spine; in addition, by locating the sagittal plane, transverse plane, coronal plane, parasagittal plane and oblique sagittal plane, a multi-angle and multi-dimensional observation perspective of the spine is provided, which helps to better present the structure of the spine; through efficient, accurate and comprehensive anatomical plane positioning of the spine, it can provide important technical support for the clinical diagnosis, treatment and research of the spine.

[0103] 3. By correcting the spinal anatomical plane, the accuracy of automatic positioning of the spinal anatomical plane is further improved, thereby ensuring the accurate positioning of the spinal anatomical plane in space, which helps to perform medical analysis or surgical planning more accurately; and accurate spinal anatomical plane positioning can also be used to calculate the screw implantation angle, and has positive significance for assisting doctors in performing spinal fusion surgery and improving the safety and effectiveness of spinal fusion surgery.

[0104] Fig.10 A schematic diagram of a structure of a spinal anatomical surface automatic positioning device provided by an exemplary embodiment of the present application. Fig.10 As shown, the automatic vertebral anatomical plane positioning device 100 includes a vertebral segmentation module 101, a key point recognition module 102 and a key point recognition module 103. Among them:

[0105] The spine segmentation module 101 is used to obtain a spine segmentation mask of a CT image of a target spine based on a pre-trained spine segmentation model, wherein different segments of the spine segmentation mask have different corresponding segmentation masks;

[0106] A key point recognition module 102 is used to obtain key points of each vertebra in different vertebrae from a segmentation mask corresponding to the vertebrae based on a pre-trained vertebrae key point recognition model;

[0107] The processing module 103 is used to determine the spinal anatomical plane of the segmental vertebra based on the key points.

[0108] In one possible implementation, the key point identification module 102 can be specifically used to: input the segmentation mask corresponding to the segmented spine into the spinal key point identification model; generate a set number of heat maps corresponding to the segmented spine based on the segmentation mask corresponding to the segmented spine; for each heat map in the set number of heat maps, determine the pixel with the largest pixel value among the pixels contained in the heat map as a key point of the segmented spine.

[0109] In a possible implementation, the spinal anatomical planes include sagittal planes, transverse planes, coronal planes, parasagittal planes, and oblique sagittal planes, and the key points include the center point of the vertebral body, the center point of the posterior edge of the vertebral body, the center point of the junction between the spinous process and the lamina, the center point of the left pedicle, and the center point of the right pedicle. The processing module 103 can be specifically used to: determine the sagittal plane of the vertebra based on the center point of the vertebral body, the center point of the posterior edge of the vertebral body, and the center point of the junction between the spinous process and the lamina; determine the center point of the vertebral body in the section at right angles to the sagittal plane. The section through the center point of the posterior edge of the vertebral body is the cross-section of the segmental vertebra; in the section that is at right angles to both the sagittal plane and the cross-section, the section passing through the center point of the posterior edge of the vertebral body is determined to be the coronal plane of the segmental vertebra; in the section parallel to the sagittal plane, the section passing through the center point of the left pedicle and the center point of the right pedicle is determined to be the parasagittal plane of the segmental vertebra; in the section that is at right angles to the cross-section and at a set angle to the coronal plane, the section passing through the center point of the left pedicle and the center point of the right pedicle is determined to be the oblique sagittal plane of the segmental vertebra.

[0110] In a possible implementation, the processing module 103 can be specifically used to: determine a first key point and a second key point on the sagittal plane, the first key point being the midpoint from the center point of the vertebra along the cross section to the front edge of the vertebra, and the second key point being the midpoint from the center point of the vertebra along the cross section to the rear edge of the vertebra; based on the first key point and the second key point, correct the spinal anatomical plane to obtain a corrected spinal anatomical plane.

[0111] In a possible implementation, the processing module 103 can also be used to: determine a first straight line passing through a first key point and perpendicular to the cross section, and determine a second straight line passing through a second key point and perpendicular to the cross section; determine the midpoint of the first straight line segment inside the vertebral body as a third key point, and determine the midpoint of the second straight line segment inside the vertebral body as a fourth key point; determine the center point of the junction of the revised spinous process and the lamina based on the third key point and the fourth key point; determine the revised spinal anatomical surface based on the center point of the junction of the revised spinous process and the lamina.

[0112] In a possible implementation, the processing module 103 may also be used to: connect the third key point and the fourth key point to obtain a third straight line; and determine the intersection point of the third straight line and the posterior edge of the vertebral body as the corrected center point of the junction between the spinous process and the lamina.

[0113] In a possible implementation, the processing module 103 can also be used to: determine the first normal vector of the revised cross-section based on the center point of the junction between the revised spinous process and the lamina; cross-product the second normal vector of the sagittal plane and the first normal vector to obtain the third normal vector of the revised coronal plane; construct a revised spinal anatomical plane based on the first normal vector, the second normal vector and the third normal vector.

[0114] The automatic positioning device for spinal anatomical planes provided in the embodiment of the present application can execute the technical solution shown in the above-mentioned embodiment of the automatic positioning method for spinal anatomical planes. Its implementation principle and beneficial effects are similar and will not be repeated here.

[0115] It should be noted that it should be understood that the division of the various modules of the above device is only a division of logical functions. In actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated. And these modules can all be implemented in the form of software called by processing elements; they can also be all implemented in the form of hardware; some modules can also be implemented in the form of software called by processing elements, and some modules can be implemented in the form of hardware. For example, the processing module can be a separately established processing element, or it can be integrated in a chip of the above device. In addition, it can also be stored in the memory of the above device in the form of program code, and called and executed by a processing element of the above device. The implementation of other modules is similar. In addition, these modules can be fully or partially integrated together, or they can be implemented independently. The processing element here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each module above can be completed by an integrated logic circuit of hardware in the processor element or instructions in the form of software.

[0116] For example, the above modules may be one or more integrated circuits configured to implement the above methods, such as one or more application specific integrated circuits (ASIC), or one or more microprocessors (DSP), or one or more field programmable gate arrays (FPGA). For another example, when a module is implemented in the form of a processing element scheduling program code, the processing element may be a general-purpose processor, such as a central processing unit (CPU) or other processor that can call program code. For another example, these modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0117] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loading and executing computer instructions on a computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website site, a computer, a server or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (Digital Subscriber Line, referred to as DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server, a data center, etc. that contains one or more available media integrated. Available media may be magnetic media (eg, floppy disks, hard disks, magnetic tapes), optical media (eg, Digital Video Discs (DVDs)), or semiconductor media (eg, solid state disks (SSDs)).

[0118] Fig.11 The schematic diagram of the structure of the electronic device provided by the exemplary embodiment of the present application. Fig.11 As shown, the electronic device 110 of this embodiment includes:

[0119] At least one processor 111; and a memory 112 communicatively connected to the at least one processor;

[0120] The memory 112 stores instructions that can be executed by the at least one processor 111, and the instructions are executed by the at least one processor 111 to enable the electronic device to execute the method described in any of the above embodiments.

[0121] Optionally, the memory 112 may be independent or integrated with the processor 111 .

[0122] The memory 112 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory.

[0123] The processor 111 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application. Specifically, when implementing the automatic positioning method of the spinal anatomical plane described in the aforementioned method embodiment, the electronic device may be, for example, an electronic device with processing functions such as a server.

[0124] Optionally, the electronic device may further include a communication interface 113. In a specific implementation, if the communication interface 113, the memory 112 and the processor 111 are implemented independently, the communication interface 113, the memory 112 and the processor 111 may be interconnected through a bus and communicate with each other. The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc., but it does not mean that there is only one bus or one type of bus.

[0125] Optionally, in a specific implementation, if the communication interface 113, the memory 112 and the processor 111 are integrated on a chip, the communication interface 113, the memory 112 and the processor 111 can communicate through an internal interface.

[0126] The implementation principle and technical effects of the electronic device provided in this embodiment can be found in the aforementioned embodiments and will not be described in detail here.

[0127] An embodiment of the present application also provides a computer-readable storage medium, in which computer execution instructions are stored. When the computer execution instructions are executed, they are used to implement the method steps in the above method embodiment. The specific implementation method and technical effect are similar and will not be repeated here.

[0128] The above-mentioned computer readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read only memory (EEPROM), erasable programmable read only memory (EPROM), programmable read only memory (PROM), read only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk. The readable storage medium can be any available medium that can be accessed by a general or special-purpose computer.

[0129] An exemplary readable storage medium is coupled to the processor so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can be located in a dedicated integrated circuit. Of course, the processor and the readable storage medium can also be present in the automatic positioning device for spinal anatomical planes as discrete components.

[0130] The embodiment of the present application also provides a computer program product, including a computer program. When the computer program is executed, the method steps in the above method embodiment are implemented. The specific implementation method and technical effect are similar and will not be repeated here.

[0131] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0132] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for automatically positioning a spinal anatomical plane, characterized in that: include: Based on a pre-trained spine segmentation model, a spine segmentation mask of a CT image of a target spine is obtained, wherein different segments of the spine have different segmentation masks corresponding to them; For each vertebra in different vertebrae, based on a pre-trained vertebrae key point recognition model, the key point of the vertebrae is obtained from the segmentation mask corresponding to the vertebrae; Based on the key points, the spinal anatomical plane of the vertebral segment is determined.

2. The method according to claim 1, characterized in that The step of obtaining the key points of the vertebral segment from the segmentation mask corresponding to the vertebral segment based on the pre-trained vertebral key point recognition model includes: Inputting the segmentation mask corresponding to the vertebral segment into the vertebral key point recognition model; Based on the segmentation mask corresponding to the spine section, generating a set number of heat maps corresponding to the spine section; For each heat map in the set number of heat maps, a pixel point with a maximum pixel value among the pixels included in the heat map is determined as a key point of the vertebral column.

3. The method according to claim 1 or 2, characterized in that: The spinal anatomical planes include sagittal plane, transverse plane, coronal plane, parasagittal plane and oblique sagittal plane, the key points include the center point of the vertebral body, the center point of the posterior edge of the vertebral body, the center point of the junction of the spinous process and the lamina, the center point of the left pedicle and the center point of the right pedicle, and the spinal anatomical plane of the segment of the vertebra is determined based on the key points, including: Determine the sagittal plane of the segmental vertebra based on the center point of the vertebral body, the center point of the posterior edge of the vertebral body, and the center point of the junction of the spinous process and the lamina; In the section at right angles to the sagittal plane, determine the section passing through the center point of the vertebral body and the center point of the posterior edge of the vertebral body as the cross section of the segmental vertebra; In the section planes that are at right angles to both the sagittal plane and the transverse plane, determine the section plane passing through the center point of the posterior edge of the vertebral body as the coronal plane of the segmental vertebra; In the section parallel to the sagittal plane, determine the section passing through the center point of the left pedicle and the center point of the right pedicle as the parasagittal plane of the segmental vertebra; In the section that is at a right angle to the cross section and at a set angle to the coronal plane, the section passing through the center point of the left pedicle and the center point of the right pedicle is determined to be the oblique sagittal plane of the segmental vertebra.

4. The method according to claim 3, characterized in that The method further comprises: Determine a first key point and a second key point on the sagittal plane, wherein the first key point is a midpoint from the center point of the vertebral body along the cross section to the front edge of the vertebral body, and the second key point is a midpoint from the center point of the vertebral body along the cross section to the rear edge of the vertebral body; The spinal anatomical plane is corrected according to the first key point and the second key point to obtain a corrected spinal anatomical plane.

5. The method according to claim 4, characterized in that The step of correcting the spinal anatomical plane according to the first key point and the second key point to obtain a corrected spinal anatomical plane comprises: Determine a first straight line passing through the first key point and perpendicular to the cross section, and determine a second straight line passing through the second key point and perpendicular to the cross section; Determine the midpoint of the first straight line in the line segment inside the vertebral body as the third key point, and determine the midpoint of the second straight line in the line segment inside the vertebral body as the fourth key point; According to the third key point and the fourth key point, determining the corrected center point of the junction between the spinous process and the lamina; The revised spinal anatomical plane is determined based on the revised center point of the junction between the spinous process and the lamina.

6. The method according to claim 5, characterized in that Determining the corrected center point of the junction between the spinous process and the lamina according to the third key point and the fourth key point includes: Connecting the third key point and the fourth key point to obtain a third straight line; The intersection point of the third straight line and the posterior edge of the vertebral body is determined as the center point of the corrected junction between the spinous process and the lamina.

7. The method according to claim 5, characterized in that Determining the revised spinal anatomical plane based on the revised center point of the junction between the spinous process and the lamina includes: Determining a first normal vector of a revised cross section according to the revised center point of the junction between the spinous process and the lamina; Cross-product the second normal vector of the sagittal plane and the first normal vector to obtain a corrected third normal vector of the coronal plane; The modified spinal anatomical plane is constructed according to the first normal vector, the second normal vector and the third normal vector.

8. A device for automatically positioning spinal anatomical planes, characterized in that: include: A spine segmentation module, used for obtaining a spine segmentation mask of a CT image of a target spine based on a pre-trained spine segmentation model, wherein different segments of the spine have different segmentation masks corresponding to them; A key point recognition module is used to obtain the key points of each vertebra in different vertebrae from the segmentation mask corresponding to the vertebrae based on a pre-trained vertebrae key point recognition model; The processing module is used to determine the spinal anatomical plane of the vertebral segment based on the key points.

9. An electronic device, characterized in that: include: A processor, and a memory communicatively connected to the processor; The memory is used to store computer-executable instructions; The processor is used to execute the computer-executable instructions to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 7 when executed.

11. A computer program product, comprising a computer program, characterized in that When the computer program is executed, the method according to any one of claims 1 to 7 is implemented.

Images