A method and system for scoliosis end vertebra identification

Abstract

The present application relates to the technical field of vertebral body recognition, and provides a scoliosis end vertebra recognition method, which comprises the following steps: S1: inputting original image data into a vertebral body recognition division model to obtain labeled image data after vertebral body recognition; S2: calculating the vertebral body angles between all the vertebral bodies of thoracic vertebrae and lumbar vertebrae in the labeled image data, taking the maximum value in the vertebral body angles as a first Cobb angle, and taking the two vertebral bodies corresponding to the first Cobb angle as a first Cobb angle upper end vertebra and a first Cobb angle lower end vertebra; S3: judging the overall shape of the spine part, calculating a second Cobb angle and a third Cobb angle according to the overall shape, and obtaining end vertebrae corresponding to the second Cobb angle and the third Cobb angle. Compared with the traditional manual labeling method, the measurement speed is fast, the recognition accuracy is high, it is more suitable for large-scale data processing, and the operation process of the user is simplified.

Description

Technical Field

[0001] This invention relates to the technical field of vertebral body recognition, and more particularly to a method and system for recognizing vertebrae at the end of a scoliosis. Background Technology

[0002] With the rapid development of computer and medical technologies, the application of deep learning in image recognition and clinical parameter calculation is one of the current research hotspots in the field of artificial intelligence. Automated recognition methods using deep learning technology have been applied to measurement and data processing.

[0003] Scoliosis, a common spinal condition, is a hot topic in spinal stability research. On a standing X-ray, the vertebra with the largest cephalic and caudal tilt angles in scoliosis, known as the end vertebra (EV), is defined by drawing straight lines along the superior endplate of the upper vertebra and the inferior endplate of the lower vertebra. The angle between these two lines, or the angle between their perpendicular lines, is the Cobb angle. This angle is a numerical value used to assess and quantify the degree of scoliosis and has guiding value for clinical treatment.

[0004] In clinical trials and clinical treatment evaluations, the Cobb angle is an indispensable and crucial parameter. In actual measurement, orthopedic surgeons often need to review numerous radiographic images; vertebral identification and angle measurement are time-consuming, and diagnostic quality is affected by subjectivity and effort. Intelligent vertebral identification and angle measurement systems using artificial intelligence technology can help doctors save measurement and calculation time. During image review, doctors only need to confirm the recommendations provided by the AI ​​system, and the system can automatically measure and calculate the required data. Therefore, the use of automatic identification of vertebral segment types and location numbers, and automatic angle measurement, is crucial in clinical practice and clinical research.

[0005] Measuring the Cobb angle requires first determining the position of the end vertebrae. Traditional methods require clinicians to manually select the upper and lower vertebrae of the scoliosis and then perform manual measurements. This involves a huge workload during the processing of large amounts of data, and the accuracy of the measurements cannot be controlled, resulting in high data heterogeneity. Furthermore, since manual measurement requires obtaining standard standing anteroposterior radiographs, it places higher demands on radiologists.

[0006] Currently, intelligent end vertebrae recognition algorithms based on 3D images of the spine have emerged. These algorithms have high recognition efficiency and high computational accuracy. However, during use, the subject needs to undergo a CT scan first. After the CT images are used to reconstruct the spine in 3D, the spine model is registered with the X-ray images to identify the end vertebrae. The process is complex, the registration operation is labor-intensive, and the subject receives a high radiation dose. Summary of the Invention

[0007] To address the aforementioned problems, the present invention aims to provide a method and system for identifying vertebrae in scoliosis. Based on an artificial intelligence recognition system, it automatically identifies standing X-ray images and locates the coordinates of vertebrae in the scoliosis image. Compared to traditional manual annotation methods, the vertebrae identification method based on an artificial intelligence image recognition system offers faster measurement speed, higher recognition accuracy, and is more suitable for large-scale data processing. Compared to vertebrae identification methods based on 3D models, this invention uses standing X-ray images as the identification object, eliminating the need for CT scanning for 3D modeling of the subject, reducing radiation exposure dose, and simplifying the user's operation process.

[0008] The above-mentioned objective of this invention is achieved through the following technical solutions:

[0009] A method for identifying vertebrae in scoliosis includes the following steps:

[0010] S1: Input the raw image data of the spine into the vertebral body recognition and segmentation model to obtain the labeled image data after vertebral body recognition;

[0011] S2: In the labeled image data, calculate the vertebral angles between each pair of vertebrae of the thoracic and lumbar vertebrae, and take the maximum value of the vertebral angles as the first Cobb angle. The two vertebrae corresponding to the first Cobb angle are the upper vertebra and the lower vertebra of the first Cobb angle.

[0012] S3: Determine the overall shape of the spinal region, calculate the second Cobb angle and the third Cobb angle based on the overall shape, and obtain the end vertebrae corresponding to the second Cobb angle and the third Cobb angle.

[0013] Further, in step S1, the vertebral body recognition and segmentation model is constructed as follows:

[0014] Collect a large amount of image data of the spine as a training image data set;

[0015] The image data in the training image dataset undergoes preprocessing including window width, window level, contrast enhancement, and image size standardization.

[0016] The image data in the training image dataset is labeled with features including cone position, cone order, and cone corner points to construct a spine corner point recognition model.

[0017] The standardized spine model is input into the spine corner point recognition model for training, and the vertebral body recognition and segmentation model is constructed.

[0018] Further, in step S2, the vertebral angles between each pair of vertebrae of the thoracic and lumbar vertebrae are calculated, specifically as follows:

[0019] Calculate the center points on both sides of all the vertebral bodies of the thoracic and lumbar vertebrae to form a center point vector;

[0020] Calculate the angles between each pair of the center point vectors of all the vertebrae of the thoracic and lumbar vertebrae.

[0021] Furthermore, in step S3, the overall shape of the spinal region is determined, specifically as follows:

[0022] Calculate the slope of the center point vectors of the thoracic vertebra T1 and the lumbar vertebra L5 respectively;

[0023] When the slopes of the center point vectors of the thoracic vertebra T1 and the lumbar vertebra L5 are positive and negative respectively, the overall shape is not S-shaped; otherwise, the overall shape is S-shaped.

[0024] Further, in step S3, the second Cobb angle and the third Cobb angle are calculated based on the overall shape, and the end vertices corresponding to the second Cobb angle and the third Cobb angle are obtained, specifically as follows:

[0025] Let the first sequential terminal vertebra V1, the second sequential terminal vertebra V2, the third sequential terminal vertebra V3, and the fourth sequential terminal vertebra V4 be the four terminal vertebrae of the vertebral region from top to bottom;

[0026] When the overall shape is not S-shaped:

[0027] Taking the thoracic vertebra T1 as the first sequential end vertebra V1 and the vertebra above the first Cobb angle as the second sequential end vertebra V2, the angle between the first sequential end vertebra V1 and the second sequential end vertebra V2 is calculated as the second Cobb angle; taking the lumbar vertebra L5 as the fourth sequential end vertebra V4 and the vertebra below the first Cobb angle as the third sequential end vertebra V3, the angle between the fourth sequential end vertebra V4 and the third sequential end vertebra V3 is calculated as the third Cobb angle;

[0028] When the overall shape is S-shaped:

[0029] Determine the positions of the upper and lower vertebrae of the first Cobb angle in the labeled image data;

[0030] If the vertebrae at the upper end of the first Cobb angle and the vertebrae at the lower end of the first Cobb angle are in the upper half of the labeled image data, then for all vertebrae between the thoracic vertebra T1 and the vertebrae at the upper end of the first Cobb angle, the vertebral angles between them and the vertebrae at the upper end of the first Cobb angle are calculated, and the maximum value of the vertebral angles is taken as the second Cobb angle, the upper vertebrae corresponding to the second Cobb angle is taken as the first sequential end vertebra V1, and the upper vertebrae at the first Cobb angle is taken as the second sequential end vertebra V2; for all vertebrae between the lumbar vertebra L5 and the vertebrae at the lower end of the first Cobb angle, the vertebral angles between them and the vertebrae at the lower end of the first Cobb angle are calculated, and the maximum value of the vertebral angles is taken as the third Cobb angle, the lower vertebrae at the first Cobb angle is taken as the third sequential end vertebra V3, and the lower vertebrae corresponding to the third Cobb angle is taken as the fourth sequential end vertebra V4;

[0031] If the upper and lower vertebrae of the first Cobb angle are located in the lower half of the labeled image data, then for all vertebrae between the thoracic vertebra T1 and the upper vertebra of the first Cobb angle, the vertebral angles between them and the upper vertebra of the first Cobb angle are calculated. The maximum value of these vertebral angles is taken as the second Cobb angle, the upper vertebra corresponding to the second Cobb angle is taken as the second sequential end vertebra V2, and the upper vertebra of the first Cobb angle is taken as the third sequential end vertebra V3. For all vertebrae between the thoracic vertebra T1 and the second sequential end vertebra V2, the vertebral angles between them and the second sequential end vertebra V2 are calculated. The maximum value of these vertebral angles is taken as the third Cobb angle, the upper vertebra corresponding to the third Cobb angle is taken as the first sequential end vertebra V1, and the lower vertebra of the first Cobb angle is taken as the fourth sequential end vertebra V4.

[0032] Further, the positions of the upper and lower vertebrae of the first Cobb angle in the labeled image data are determined as follows:

[0033] Let the center coordinates of the two corner points on the lower side of the upper vertebra of the first Cobb angle and the lower vertebra of the first Cobb angle be P1 and P2, respectively.

[0034] Let the center coordinates of the two lower corner points of the thoracic vertebra T1 and the lumbar vertebra L5 be P3 and P4, respectively.

[0035] When (P1+P2)<1.2*(P3+P4), the upper vertebra of the first Cobb angle and the lower vertebra of the first Cobb angle are in the upper half of the labeled image data;

[0036] When (P1+P2)>=1.2*(P3+P4), the upper vertebra of the first Cobb angle and the lower vertebra of the first Cobb angle are in the lower half of the labeled image data.

[0037] Furthermore, the method for identifying vertebrae at the end of a scoliosis vertebra also includes:

[0038] S4: Calculate the apical vertebra or intervertebral disc between the terminal vertebrae sequentially, specifically:

[0039] A straight line connecting the midpoints of the first sequential end vertebra V1, the second sequential end vertebra V2, the third sequential end vertebra V3, and the fourth sequential end vertebra V4 from top to bottom is formed between two of the end vertebrae.

[0040] The distances between the vertebral body and intervertebral disc between the first sequential end vertebra V1 and the second sequential end vertebra V2, the second sequential end vertebra V2 and the third sequential end vertebra V3, and the third sequential end vertebra V3 and the fourth sequential end vertebra V4 are calculated sequentially.

[0041] Between the two end vertebrae, the vertebra or intervertebral disc that is longest in the straight line is the apical vertebra or apical intervertebral disc corresponding to the two end vertebrae.

[0042] Furthermore, the training image data set includes anteroposterior and lateral image data of the spinal region; the first Cobb angle, the second Cobb angle, and the third Cobb angle are calculated from the original anteroposterior and lateral image data, respectively, and the end vertebrae corresponding to the first Cobb angle, the second Cobb angle, and the third Cobb angle are obtained to perform stereoscopic recognition of scoliosis of the spinal region.

[0043] A scoliosis vertebral identification system for performing the scoliosis vertebral identification method as described above includes:

[0044] The image data annotation module is used to input the raw image data of the spine into the vertebral body recognition and segmentation model to obtain annotated image data after vertebral body recognition.

[0045] The first Cobb angle calculation module is used to calculate the vertebral angle between each pair of vertebrae in the annotated image data, and take the maximum value of the vertebral angle as the first Cobb angle, and the two vertebrae corresponding to the first Cobb angle as the upper vertebra of the first Cobb angle and the lower vertebra of the first Cobb angle.

[0046] The second and third Cobb angle calculation module is used to determine the overall shape of the spinal region, calculate the second and third Cobb angles based on the overall shape, and obtain the end vertebrae corresponding to the second and third Cobb angles.

[0047] The apical vertebra calculation module is used to sequentially calculate the apical vertebra or apical intervertebral disc between the terminal vertebrae.

[0048] A computer device includes a memory and one or more processors, the memory storing computer code that, when executed by the one or more processors, causes the one or more processors to perform the method described above.

[0049] A computer-readable storage medium storing computer code that, when executed, performs the method described above.

[0050] Compared with the prior art, the present invention has at least one of the following beneficial effects:

[0051] (1) A method for identifying vertebrae in scoliosis is provided, comprising the following steps: S1: Inputting the original image data of the spine into a vertebral body identification and segmentation model to obtain labeled image data after vertebral body identification; S2: Calculating the vertebral body angles between all pairs of vertebrae in the labeled image data for the thoracic and lumbar vertebrae, taking the maximum value of the vertebral body angles as the first Cobb angle, and taking the two vertebrae corresponding to the first Cobb angle as the upper and lower vertebrae of the first Cobb angle; S3: Judging the overall shape of the spine, calculating the second and third Cobb angles based on the overall shape, and obtaining the vertebrae corresponding to the second and third Cobb angles. Compared with the traditional manual labeling method, the vertebrae identification method based on the artificial intelligence image recognition system has a faster measurement speed, higher recognition accuracy, and is more suitable for large-scale data processing. Compared with the vertebrae identification method based on three-dimensional models, this invention uses standing X-ray films as the identification object, eliminating the need for CT scanning for three-dimensional modeling of the subject, reducing radiation exposure dose, and simplifying the user's operation process.

[0052] (2) By setting anteroposterior and lateral image data of the spine in the training image dataset; calculating the first Cobb angle, the second Cobb angle, and the third Cobb angle for the original anteroposterior and lateral image data respectively, and obtaining the end vertebrae corresponding to the first Cobb angle, the second Cobb angle, and the third Cobb angle. The above technical solution realizes the technology of three-dimensional coordinate positioning based on two-dimensional images, combining the advantages of CT spatial coordinate establishment and low X-ray radiation dose.

[0053] (3) Based on the intelligent vertebral body recognition system, the vector angle between the vertebral bodies is calculated to construct the vector angle of the vertebral body to determine the position of the top vertebra of the scoliosis spine, thereby realizing the identification of the scoliosis type and the end vertebra. Attached Figure Description

[0054] Figure 1 This is an overall flowchart of the spinal scoliosis end vertebra identification method of the present invention;

[0055] Figure 2 This is a schematic diagram of the vertebral body corner markings of the present invention;

[0056] Figure 3 This is a schematic diagram of the center point vector of the vertebral body calculated according to the present invention;

[0057] Figure 4 This is a schematic diagram showing the vertebrae at the upper end of the first Cobb angle and the vertebrae at the lower end of the first Cobb angle calculated according to the present invention;

[0058] Figure 5 This is an overall structural diagram of the spinal scoliosis end vertebra recognition system of the present invention. Detailed Implementation

[0059] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0060] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this specification means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0061] This invention relates to an end-vertebra identification algorithm based on an artificial intelligence vertebral body recognition system, which automatically identifies and locates the coordinates of end vertebrae in scoliosis images based on standing X-ray films. Compared to traditional manual annotation methods, the end-vertebra identification method based on an artificial intelligence image recognition system offers faster measurement speed, higher recognition accuracy, and is more suitable for large-scale data processing. Compared to end-vertebra identification methods based on 3D models, this invention uses standing X-ray films as the identification object, eliminating the need for CT scans for 3D modeling of the subject, reducing radiation exposure dose, and simplifying the user's operation process. The use of a dual-plane whole-body X-ray system enables 3D coordinate positioning based on 2D images, combining the advantages of CT spatial coordinate establishment and low X-ray radiation dose.

[0062] First Embodiment

[0063] like Figure 1 As shown, this embodiment provides a method for identifying vertebrae at the end of a scoliosis, including the following steps:

[0064] S1: Input the original image data of the spine into the vertebral body recognition and segmentation model to obtain the labeled image data after vertebral body recognition.

[0065] Specifically, before implementing the vertebral body identification method for scoliosis of the present invention, it is first necessary to construct a vertebral body identification and segmentation model. The steps for constructing the vertebral body identification and segmentation model are as follows:

[0066] (1) A large amount of image data of the vertebral region is collected as a training image dataset and preprocessed. In order to realize the subsequent three-dimensional coordinate positioning technology based on two-dimensional images, frontal and lateral dual-plane image data can be collected here. Among the collected standard image data in DI COM format, dual-plane image data with good quality and belonging to the vertebral region are selected for composite vertebral body recognition.

[0067] (2) The image data in the training image dataset is preprocessed, including window width (pixel value coverage range), window position (pixel value distribution center), contrast enhancement, and image size standardization, to obtain the data to be processed.

[0068] (3) Label the image data in the training image dataset, including cone position, cone order, and cone corner points, to construct a spine corner point recognition model. The labeling of the image data can be done manually by experienced experts or through other automated methods (such as...). Figure 2 ).

[0069] (4) The standardized spine model is input into the spine corner point recognition model for training to construct the vertebral body recognition and segmentation model. Data augmentation is performed by random left and right flipping, expansion, and photometric distortion (brightness, contrast, noise).

[0070] S2: In the labeled image data, calculate the vertebral angles between each pair of vertebrae of the thoracic and lumbar vertebrae, and take the maximum value of the vertebral angles as the first Cobb angle. The two vertebrae corresponding to the first Cobb angle are taken as the upper vertebra and the lower vertebra of the first Cobb angle.

[0071] The human spine, from top to bottom, includes 7 cervical vertebrae (C1-C7), 12 thoracic vertebrae (T1-T12), and 5 lumbar vertebrae (L1-L5). This invention identifies all thoracic and lumbar vertebrae when identifying the end vertebrae in scoliosis.

[0072] Specifically, by calculating the center points on both sides of all the vertebral bodies of the thoracic vertebrae (T1-T12) and the lumbar vertebrae (L1-L5), a center point vector is formed (e.g., Figure 3 Then, the angles between the center point vectors of all the vertebrae of the thoracic and lumbar vertebrae are calculated, and the maximum value of the vertebral body angles is taken as the first Cobb angle. The two vertebrae corresponding to the first Cobb angle are taken as the upper end vertebra and the lower end vertebra of the first Cobb angle (e.g., Figure 4 As shown, the two long black lines mark the two end vertebrae.

[0073] S3: Determine the overall shape of the spinal region, calculate the second Cobb angle and the third Cobb angle based on the overall shape, and obtain the end vertebrae corresponding to the second Cobb angle and the third Cobb angle.

[0074] In this step, the overall shape of the spine needs to be determined first. Specifically, the slopes of the center point vectors of the thoracic vertebra T1 and the lumbar vertebra L5 are calculated respectively. When the slopes of the center point vectors of the thoracic vertebra T1 and the lumbar vertebra L5 are one positive and one negative, the overall shape is not S-shaped; otherwise, the overall shape is S-shaped.

[0075] After obtaining the overall shape, the second and third Cobb angles are calculated based on the overall shape. Specifically, this includes the following cases:

[0076] Let the first sequential terminal vertebra V1, the second sequential terminal vertebra V2, the third sequential terminal vertebra V3, and the fourth sequential terminal vertebra V4 be the four terminal vertebrae of the vertebral region from top to bottom;

[0077] (1) When the overall shape is not S-shaped:

[0078] The thoracic vertebra T1 is designated as the first sequential end vertebra V1, and the vertebra above the first Cobb angle is designated as the second sequential end vertebra V2.

[0079] Calculate the angle between the first sequential end vertebra V1 and the second sequential end vertebra V2, and use it as the second Cobb angle;

[0080] Taking the lumbar vertebra L5 as the fourth-order end vertebra V4 and the vertebra below the first Cobb angle as the third-order end vertebra V3, the angle between the fourth-order end vertebra V4 and the third-order end vertebra V3 is calculated as the third Cobb angle.

[0081] (2) When the overall shape is S-shaped:

[0082] First, it is necessary to determine the positions of the upper and lower vertebrae of the first Cobb angle in the labeled image data. The specific method is as follows:

[0083] Let the center coordinates of the two corner points on the lower side of the first Cobb angle's upper and lower ends be P1 and P2, respectively; let the center coordinates of the two corner points on the lower sides of the thoracic vertebra T1 and the lumbar vertebra L5 be P3 and P4, respectively. For the corner points, approximate the vertebral body as a rectangle, numbering them clockwise from the upper left corner (1,2,3,4), with the center coordinates being the coordinates of the centers of corner points 3 and 4.

[0084] When (P1+P2)<1.2*(P3+P4), the upper vertebra of the first Cobb angle and the lower vertebra of the first Cobb angle are in the upper half of the labeled image data; when (P1+P2)>=1.2*(P3+P4), the upper vertebra of the first Cobb angle and the lower vertebra of the first Cobb angle are in the lower half of the labeled image data.

[0085] (21) If the upper vertebra of the first Cobb angle and the lower vertebra of the first Cobb angle are in the upper half of the labeled image data, then calculate the vertebral angle between the upper vertebra of the first Cobb angle and all the vertebrae between the thoracic vertebra T1 and the vertebrae between the first Cobb angle, and take the maximum value of the vertebral angle as the second Cobb angle, take the upper vertebra of the second Cobb angle as the first sequential end vertebra V1, and take the upper vertebra of the first Cobb angle as the second sequential end vertebra V2; calculate the vertebral angle between the lower vertebra of the first Cobb angle and all the vertebrae between the lumbar vertebra L5 and the vertebrae of the first Cobb angle, and take the maximum value of the vertebral angle as the third Cobb angle, take the lower vertebra of the first Cobb angle as the third sequential end vertebra V3, and take the lower vertebra of the third Cobb angle as the fourth sequential end vertebra V4.

[0086] (22) If the upper vertebra of the first Cobb angle and the lower vertebra of the first Cobb angle are in the lower half of the labeled image data, then calculate the vertebral angle between the upper vertebra of the first Cobb angle and all the vertebrae between the thoracic vertebra T1 and the upper vertebra of the first Cobb angle, and take the maximum value of the vertebral angle as the second Cobb angle, take the upper vertebra of the second Cobb angle as the second sequential end vertebra V2, and take the upper vertebra of the first Cobb angle as the third sequential end vertebra V3; calculate the vertebral angle between the second sequential end vertebra V2 and all the vertebrae between the thoracic vertebra T1 and the second sequential end vertebra V2, take the maximum value of the vertebral angle as the third Cobb angle, take the upper vertebra of the third Cobb angle as the first sequential end vertebra V1, and take the lower vertebra of the first Cobb angle as the fourth sequential end vertebra V4.

[0087] Furthermore, the method for identifying vertebrae at the end of a scoliosis vertebra also includes:

[0088] S4: Calculate the apical vertebra or intervertebral disc between the two end vertebrae in sequence, specifically: connect the midpoints of the first sequential end vertebra V1, the second sequential end vertebra V2, the third sequential end vertebra V3, and the fourth sequential end vertebra V4 from top to bottom to form a straight line connecting the two end vertebrae; calculate the distances of the vertebral body and intervertebral disc between the two end vertebrae V1 and V2, V2 and V3, and V3 and V4 from the straight line in sequence; the vertebral body or intervertebral disc with the longest distance from the straight line between the two end vertebrae is the apical vertebra or intervertebral disc corresponding to the current two end vertebrae.

[0089] Furthermore, the scoliosis end vertebra identification method of the present invention further includes: the training image data set includes anteroposterior and lateral image data of the vertebral region; calculating the first Cobb angle, the second Cobb angle, and the third Cobb angle for the original anteroposterior and lateral image data respectively, and obtaining the end vertebrae corresponding to the first Cobb angle, the second Cobb angle, and the third Cobb angle, and performing stereoscopic identification of scoliosis in the vertebral region.

[0090] Furthermore, the present invention displays the position of the end cone through an interactive interface and automatically generates measurement auxiliary lines to demonstrate the automatic measurement of the Cobb angle; users can modify the measurement position and measurement angle by adjusting the auxiliary lines on the interactive interface.

[0091] Second Embodiment

[0092] like Figure 5As shown, this embodiment provides a scoliosis vertebral identification system for performing the scoliosis vertebral identification method as described in the first embodiment, comprising:

[0093] Image data annotation module 1 is used to input the raw image data of the spine into the vertebral body recognition and segmentation model to obtain annotated image data after vertebral body recognition;

[0094] The first Cobb angle calculation module 2 is used to calculate the vertebral angle between each pair of all vertebrae of the thoracic and lumbar vertebrae in the labeled image data, and take the maximum value of the vertebral angle as the first Cobb angle, and the two vertebrae corresponding to the first Cobb angle as the upper end vertebra and the lower end vertebra of the first Cobb angle.

[0095] The second and third Cobb angle calculation module 3 is used to determine the overall shape of the spinal region, calculate the second and third Cobb angles based on the overall shape, and obtain the end vertebrae corresponding to the second and third Cobb angles.

[0096] The apical vertebra calculation module 4 is used to sequentially calculate the apical vertebra or apical intervertebral disc between the terminal vertebrae.

[0097] A computer-readable storage medium stores computer code that, when executed, performs the methods described above. Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. This program can be stored in a computer-readable storage medium, which may include: read-only memory (ROM), random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0098] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

[0099] 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.

[0100] It should be noted that the above embodiments can be freely combined as needed. The above description is only a preferred embodiment of the present invention. It should be pointed out that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for identifying vertebrae at the end of a scoliosis vertebra, characterized in that, Includes the following steps: S1: Input the raw image data of the spine into the vertebral body recognition and segmentation model to obtain the labeled image data after vertebral body recognition; S2: In the labeled image data, calculate the vertebral angles between each pair of vertebrae of the thoracic and lumbar vertebrae, and take the maximum value of the vertebral angles as the first Cobb angle. The two vertebrae corresponding to the first Cobb angle are the upper vertebra and the lower vertebra of the first Cobb angle. In step S2, the vertebral angles between each pair of vertebrae of the thoracic and lumbar vertebrae are calculated. Specifically, the center points on the left and right sides of each vertebra of the thoracic and lumbar vertebrae are calculated to form center point vectors; the angles between each pair of center point vectors of each vertebra of the thoracic and lumbar vertebrae are calculated respectively. S3: Determine the overall shape of the spinal region, calculate the second Cobb angle and the third Cobb angle based on the overall shape, and obtain the end vertebrae corresponding to the second Cobb angle and the third Cobb angle; In step S3, the overall shape of the spinal region is determined by: calculating the slope of the center point vectors of the thoracic vertebra T1 and the lumbar vertebra L5 respectively; when the slopes of the center point vectors of the thoracic vertebra T1 and the lumbar vertebra L5 are one positive and one negative, the overall shape is not S-shaped; otherwise, the overall shape is S-shaped.

2. The method for identifying vertebrae in scoliosis according to claim 1, characterized in that, In step S1, the vertebral body recognition and segmentation model is constructed as follows: Collect a large amount of image data of the spine as a training image data set; The image data in the training image dataset undergoes preprocessing including window width, window level, contrast enhancement, and image size standardization. The image data in the training image dataset is labeled with features including cone position, cone order, and cone corner points to construct a spine corner point recognition model. The standardized spine model is input into the spine corner point recognition model for training, and the vertebral body recognition and segmentation model is constructed.

3. The method for identifying vertebrae at the end of a scoliosis vertebra according to claim 1, characterized in that, In step S3, the second Cobb angle and the third Cobb angle are calculated based on the overall shape, and the end vertices corresponding to the second Cobb angle and the third Cobb angle are obtained, specifically: Let the first sequential terminal vertebra V1, the second sequential terminal vertebra V2, the third sequential terminal vertebra V3, and the fourth sequential terminal vertebra V4 be the four terminal vertebrae of the vertebral region from top to bottom; When the overall shape is not S-shaped: Taking the thoracic vertebra T1 as the first sequential end vertebra V1 and the vertebra above the first Cobb angle as the second sequential end vertebra V2, the angle between the first sequential end vertebra V1 and the second sequential end vertebra V2 is calculated as the second Cobb angle; taking the lumbar vertebra L5 as the fourth sequential end vertebra V4 and the vertebra below the first Cobb angle as the third sequential end vertebra V3, the angle between the fourth sequential end vertebra V4 and the third sequential end vertebra V3 is calculated as the third Cobb angle; When the overall shape is S-shaped: Determine the positions of the upper and lower vertebrae of the first Cobb angle in the labeled image data; If the vertebrae at the upper end of the first Cobb angle and the vertebrae at the lower end of the first Cobb angle are in the upper half of the labeled image data, then for all vertebrae between the thoracic vertebra T1 and the vertebrae at the upper end of the first Cobb angle, the vertebral angles between them and the vertebrae at the upper end of the first Cobb angle are calculated, and the maximum value of the vertebral angles is taken as the second Cobb angle, the upper vertebrae corresponding to the second Cobb angle is taken as the first sequential end vertebra V1, and the upper vertebrae at the first Cobb angle is taken as the second sequential end vertebra V2; for all vertebrae between the lumbar vertebra L5 and the vertebrae at the lower end of the first Cobb angle, the vertebral angles between them and the vertebrae at the lower end of the first Cobb angle are calculated, and the maximum value of the vertebral angles is taken as the third Cobb angle, the lower vertebrae at the first Cobb angle is taken as the third sequential end vertebra V3, and the lower vertebrae corresponding to the third Cobb angle is taken as the fourth sequential end vertebra V4; If the upper and lower vertebrae of the first Cobb angle are located in the lower half of the labeled image data, then for all vertebrae between the thoracic vertebra T1 and the upper vertebra of the first Cobb angle, the vertebral angles between them and the upper vertebra of the first Cobb angle are calculated. The maximum value of these vertebral angles is taken as the second Cobb angle, the upper vertebra corresponding to the second Cobb angle is taken as the second sequential end vertebra V2, and the upper vertebra of the first Cobb angle is taken as the third sequential end vertebra V3. For all vertebrae between the thoracic vertebra T1 and the second sequential end vertebra V2, the vertebral angles between them and the second sequential end vertebra V2 are calculated. The maximum value of these vertebral angles is taken as the third Cobb angle, the upper vertebra corresponding to the third Cobb angle is taken as the first sequential end vertebra V1, and the lower vertebra of the first Cobb angle is taken as the fourth sequential end vertebra V4.

4. The method for identifying vertebrae in scoliosis according to claim 3, characterized in that, The positions of the upper and lower vertebrae of the first Cobb angle in the labeled image data are determined as follows: Let the center coordinates of the two corner points on the lower side of the first Cobb angle's upper vertebra and lower vertebra be P1 and P2, respectively. Let the center coordinates of the two lower corner points of the thoracic vertebra T1 and the lumbar vertebra L5 be P3 and P4, respectively. When (P1+P2)<1.2*(P3+P4), the upper vertebra of the first Cobb angle and the lower vertebra of the first Cobb angle are in the upper half of the labeled image data; When (P1+P2)>=1.2*(P3+P4), the upper vertebra of the first Cobb angle and the lower vertebra of the first Cobb angle are in the lower half of the labeled image data.

5. The method for identifying vertebrae in scoliosis according to claim 3, characterized in that, Also includes: S4: Calculate the apical vertebra or intervertebral disc between the terminal vertebrae sequentially, specifically: A straight line connecting the midpoints of the first sequential end vertebra V1, the second sequential end vertebra V2, the third sequential end vertebra V3, and the fourth sequential end vertebra V4 from top to bottom is formed between two of the end vertebrae. The distances between the vertebral body and intervertebral disc between the first sequential end vertebra V1 and the second sequential end vertebra V2, the second sequential end vertebra V2 and the third sequential end vertebra V3, and the third sequential end vertebra V3 and the fourth sequential end vertebra V4 are calculated sequentially. Between the two end vertebrae, the vertebra or intervertebral disc that is longest in the straight line is the apical vertebra or apical intervertebral disc corresponding to the two end vertebrae.

6. The method for identifying vertebrae in scoliosis according to claim 2, characterized in that, Also includes: The training image dataset includes anteroposterior and lateral images of the spine. The first Cobb angle, the second Cobb angle, and the third Cobb angle are calculated from the original image data in the anterior and lateral views, respectively, and the end vertebrae corresponding to the first Cobb angle, the second Cobb angle, and the third Cobb angle are obtained to perform stereoscopic recognition of scoliosis in the vertebral region.

7. A scoliosis vertebral identification system for performing the scoliosis vertebral identification method as described in any one of claims 1-6, characterized in that, include: The image data annotation module is used to input the raw image data of the spine into the vertebral body recognition and segmentation model to obtain annotated image data after vertebral body recognition. The first Cobb angle calculation module is used to calculate the vertebral angle between each pair of all vertebrae of the thoracic and lumbar vertebrae in the labeled image data, and take the maximum value of the vertebral angle as the first Cobb angle, and the two vertebrae corresponding to the first Cobb angle as the upper end vertebra and the lower end vertebra of the first Cobb angle. The second and third Cobb angle calculation module is used to determine the overall shape of the spinal region, calculate the second and third Cobb angles based on the overall shape, and obtain the end vertebrae corresponding to the second and third Cobb angles. The apical vertebra calculation module is used to sequentially calculate the apical vertebra or apical intervertebral disc between the terminal vertebrae.

8. A computer-readable storage medium storing computer code, wherein when the computer code is executed, the method of any one of claims 1 to 6 is performed.

Images