CN113284090A

CN113284090A - Scoliosis detection method and medical platform

Info

Publication number: CN113284090A
Application number: CN202110449154.9A
Authority: CN
Inventors: 赵志阳
Original assignee: Shanghai Jihe Medical Technology Co ltd
Current assignee: Shanghai Jihe Medical Technology Co ltd
Priority date: 2021-04-25
Filing date: 2021-04-25
Publication date: 2021-08-20
Anticipated expiration: 2041-04-25
Also published as: CN113284090B

Abstract

The application provides a method for detecting scoliosis, which comprises the following steps: s1, acquiring a spine X-ray image; s2, inputting the spine X-ray image into a deep learning identification model for identification so as to identify each spine block; s3, calculating a Cobb angle based on the identified spinal column blocks; and S4, if the Cobb angle is larger than a preset value, determining that the scoliosis is caused. The scheme of this application can detect the judgement to the scoliosis condition automatically fast, has improved the efficiency and the reliability of scoliosis screening.

Description

Scoliosis detection method and medical platform

Technical Field

The application relates to the technical field of medical detection, in particular to a scoliosis detection method and a medical platform.

Background

Scoliosis refers to a deformity resulting from a lateral curvature of one or more segments of the spine off the body midline in the coronal plane, usually with rotation of the spine and an increase or decrease in kyphosis or lordosis in the sagittal plane. The incidence rate of the scoliosis of children and teenagers in China is 2% -4%, and the disease is a chronic disease which is harmful to the physical and mental health of children and teenagers in China after two diseases of myopia and obesity. Scoliosis not only affects the physical appearance of children and teenagers, but also affects the mental health of the children and teenagers, and can press organs such as heart and lung of the children and further threaten life when the children and the teenagers are serious.

The prior art has relied primarily on the size of the Cobb angle, a criterion for measuring scoliosis, named after John Robert Cobb by the american plastic surgeon, to classify idiopathic scoliosis. The traditional method of measuring the Cobb angle is by a physician manually measuring on an X-ray film using a pencil and a protractor. Since this method is performed manually, there is a certain error, and its reliability within and between observers is also poor. In addition, for some large hospitals or physical examination institutions, collective scoliosis screening is often required, for example, students in middle and primary schools may be organized to perform detection collectively, and at this time, if the detection is performed by a manual method, it obviously takes a very long time, and the increasingly-improved medical requirements of people cannot be met.

Therefore, the prior art does not effectively solve the problem of rapid and accurate automatic detection of scoliosis, and a simple and efficient detection technology is urgently needed to solve the technical problem.

Disclosure of Invention

In order to solve the technical problems existing in the background technology, the application provides a method for detecting scoliosis and a medical platform.

A first aspect of the present application provides a method of scoliosis detection, the method comprising:

s1, acquiring a spine X-ray image;

s2, inputting the spine X-ray image into a deep learning identification model for identification, and segmenting each spine block;

s3, calculating a Cobb angle based on the identified spinal column blocks;

and S4, if the Cobb angle is larger than a preset value, determining that the scoliosis is caused.

Optionally, before inputting the spine X-ray image into the deep learning recognition model for recognition, a preprocessing step is further included, including:

inputting the spine X-ray image into a spine recognition model to recognize a spine outline, and further segmenting a spine region image in the spine X-ray image;

accordingly, in step S2, the spine region image is input to a deep learning recognition model for recognition.

Optionally, inputting the spine X-ray image into a deep learning recognition model for recognition, and segmenting each spine block, including:

inputting the spine region image into a deep learning identification model, identifying each angular point in the spine region image by the deep learning identification model, grouping the angular points with the same attribute, and constructing a boundary by taking each angular point in the group as a corner so as to finish marking each spine block; wherein the same attribute means that the openings of the corner points have the same orientation, and the corner points in the group at least include three.

Optionally, the calculating a Cobb angle based on the identified each spinal column block includes:

fitting a spine curve based on the equivalent centroid position of each spine block, drawing tangent lines of the spine curve at each equivalent centroid position, and calculating the slope of each tangent line;

analyzing the slope of each tangent line one by taking an end vertebral block as an end vertebra and taking the end vertebra as a starting end vertebra, taking a corresponding vertebral block as an end vertebra when the slope has a second trend change, and calculating to obtain a Cobb angle value based on the tangent lines of the starting end vertebra and the end vertebra;

and taking the largest of the plurality of calculated Cobb angle values as a final Cobb angle value.

Optionally, the method further comprises:

and S5, inputting the spine curve into a classification model to output the scoliosis type.

A second aspect of the application provides a computer storage medium having stored thereon a computer program which, when executed by a processor, performs a method as set forth in any one of the preceding claims.

A third aspect of the application provides an electronic device comprising a processor and a memory, said memory having stored thereon a computer program which, when executed by the processor, performs the method according to any of the above.

A fourth aspect of the present application provides a medical platform comprising a detection module, a processing module, and an output module;

the detection module is used for detecting scoliosis by adopting the method;

the processing module is used for determining a corresponding treatment scheme based on the detection result of the detection module;

the output module is used for outputting the treatment scheme.

Optionally, the outputting the treatment plan comprises:

outputting the treatment plan to at least one of a doctor end, a patient end, and a hospital end.

Optionally, the medical platform further includes at least one of a reservation module, an information module, a mall module, an online diagnosis and treatment module, and an activity and sharing module.

The invention has the beneficial effects that:

in the scheme of the application, after the X-ray image of the spine is shot by the X-ray equipment, the detection and identification equipment can automatically and quickly detect and judge the scoliosis condition, so that manual marking and Cobb angle calculation are replaced, and the efficiency and the reliability of scoliosis screening are improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are required to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained from the drawings without inventive effort.

Fig. 1 is a schematic flow chart of a spinal column lateral bending detection method disclosed in an embodiment of the present application;

FIG. 2 is a schematic structural diagram of a medical platform disclosed in an embodiment of the present application;

fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. The components of the embodiments of the present application, generally described and illustrated in the figures herein, can be arranged and designed in a wide variety of different configurations.

Thus, the following detailed description of the embodiments of the present application, presented in the accompanying drawings, is not intended to limit the scope of the claimed application, but is merely representative of selected embodiments of the application. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined and explained in subsequent figures.

In the description of the present application, it should be noted that if the terms "upper", "lower", "inside", "outside", etc. are used for indicating the orientation or positional relationship based on the orientation or positional relationship shown in the drawings or the orientation or positional relationship which the present invention product is usually put into use, it is only for convenience of describing the present application and simplifying the description, but it is not intended to indicate or imply that the referred device or element must have a specific orientation, be constructed in a specific orientation and be operated, and thus, should not be construed as limiting the present application.

Furthermore, the appearances of the terms "first," "second," and the like, if any, are used solely to distinguish one from another and are not to be construed as indicating or implying relative importance.

It should be noted that the features of the embodiments of the present application may be combined with each other without conflict.

Example one

The severity of scoliosis is mostly judged by the angle of scoliosis, and the Cobb angle has become a quantitative standard for doctors to diagnose or observe scoliosis symptoms. At present, when measuring the Cobb angle, imaging doctors mostly select the vertebra which is most seriously inclined towards the lateral concave side of the spine as an upper end cone and a lower end cone by hand, and then measure the included angle between the upper end cone and the lower end cone of the spine by using a protractor. Therefore, the accuracy of the Cobb angle measurement depends on the subjective experience of the imaging physician, and the research shows that the imaging physician can cause an error of up to 11.8 degrees when manually selecting the upper and lower vertebrae of the scoliosis to measure the Cobb angle of the scoliosis, which greatly influences the diagnosis and treatment of the patient with the scoliosis. Meanwhile, parents and schools pay more attention to scoliosis nowadays, so some hospitals and physical examination institutions generally need to receive collective scoliosis screening from schools, obviously, if doctors are still used for manually measuring the Cobb angle, the workload is huge, and the screening efficiency is expected to be very low.

Therefore, referring to fig. 1, the present embodiment provides a scoliosis detection method for diagnosing the above-mentioned defects of scoliosis by manually measuring the Cobb angle in the prior art. As shown in fig. 1, a method for detecting scoliosis according to an embodiment of the present application includes:

s1, acquiring a spine X-ray image;

s3, calculating a Cobb angle based on the identified spinal column blocks;

In the embodiment of the present application, the method of the present application can be independently packaged into a specific scoliosis detection device, and the device can receive the detected scoliosis X-ray image from the X-ray device of the imaging department, and of course, the method of the present application can also be packaged into a scoliosis detection function module and integrated into the X-ray device, so that the scoliosis determination result can be directly output on site, for example, the result is marked on an X-ray film, which is beneficial for doctors and patients to know the own spine condition. The spine X-ray image is identified through a deep learning identification model, so that each spine block in the X-ray image can be identified, then the Cobb angle can be calculated by using the upper vertebra and the lower vertebra in the spine block, and if the Cobb angle is larger than a preset value, the scoliosis condition of the detection object is indicated. For example, typically, when the Cobb angle value is less than 10 °, it indicates that the patient is not suffering from scoliosis; when the Cobb angle value is more than 10 degrees and less than 25 degrees, the patient is considered to have slight scoliosis, and the physical appliance is generally adopted for auxiliary treatment at this time. When the angle value is greater than 25 °, the patient is considered to have more severe scoliosis, and further diagnosis is required, and even a specific surgical plan is determined.

In the embodiment of the application, before the deep learning identification model identifies the spine block, a spine identification model is designed, which can pre-process the spine X-ray image to segment the spine from the background image, i.e. identify the spine region image (spine ROI). Therefore, when the deep learning identification model identifies the spine block, only the simplified spine region image only including the spine needs to be identified, so that the identification accuracy and the identification efficiency of the spine block can be greatly improved.

For the identification of the image of the spine region (spine ROI), the following method can be used:

s201, performing binarization processing on the spine X-ray image, and extracting a plurality of connected region image blocks from the X-ray image after binarization processing by adopting a connected domain algorithm;

s202, performing curve characteristic analysis on the plurality of connected region image blocks, and executing the step S203 if the extracted edge curve does not meet the smooth condition;

s203, segmenting the plurality of connected region image blocks, and performing curve characteristic analysis on each segmented connected region sub image block to obtain the curve trend of each connected region sub image block;

and S204, sequentially carrying out difference accumulation on the curve trend of each connected region sub image block, and if the difference accumulation value is greater than a threshold th1, determining that the connected region image block is a spine region image.

The spine has obvious differences with respect to other skeletal structures such as ribs, hip bones, sacrum and the like, and the main difference is that the spine has significantly different curve characteristics, and particularly, the spine is clearly distinguished for ribs with higher similarity. For example, the curvature of the ribs is substantially uniform and the curvature of the two side edges is substantially smooth, whereas the curvature of the spine is variable (especially for patients with scoliosis and the curvature of the spine is also highly dependent on standing), and the two side edges of the spine are not smooth and the non-smooth characteristic is regular. Obviously, the above significant differences can be used for fast identification of spine region images (spine ROI), i.e. using curve characteristics within connected region sub-image blocks to distinguish spine from ribs, etc. Specifically, the method comprises the following steps:

in step S202, for each binarized connected region image block, an entity image edge curve (i.e., a curve formed by pixel points at the boundary between black and white pixels) can be extracted, so that curves at two sides can be obtained for the spine and the ribs, if the curve characteristics at the two sides do not satisfy the smooth condition, the spine is preliminarily determined, otherwise, the ribs are determined. Of course, the smooth condition analysis determination may be performed based on only the curve of one side. The determination of the smoothness condition is conventional in the image recognition field, and is not limited in this application.

However, the determination is not reliable only depending on whether the curve characteristic satisfies the smoothness condition, because the interference of other bone images may cause the spine image to become "smooth" as a whole, so that the following further determination is required:

in step S203, the connected region image block satisfying the smooth condition is further divided into a plurality of connected region sub image blocks, and then curve characteristic analysis is performed on the connected region sub image blocks. For example, line fitting is performed on each pixel aggregation point in each connected region sub-image block (for example, a NURBS curve construction method may be adopted), after a "line" is constructed, a tangent line may be made to each point of the "line" in each connected region image sub-block, then the obtained tangent lines are subjected to an average processing (of course, other processing methods, such as a median processing, and the like, may also be adopted), so as to obtain a slope of each connected region image sub-block (the slope represents a curve trend thereof), the above steps may be repeated to obtain a slope sequence of all connected region image sub-blocks, and a curve trend change characteristic may be obtained based on the slope sequence. Meanwhile, the curve characteristic of the spine should not be unidirectional, that is, the curve has a certain trend change, especially for the spine with lateral curvature, the characteristic that the trend of the curve changes back and forth is more obvious, and the curve is obviously different from the skeleton structures such as ribs with only a single curve trend, based on the above characteristics, the embodiment of the present application accumulates the difference of each slope (the absolute value of the difference is accumulated), if the accumulated value of the difference is greater than the threshold th1, the trend change of the curve is obvious, the image block in the connected region can be determined as the spine, otherwise, the trend of the curve is smooth and uniform, and the image block should be determined as the ribs.

Of course, if the spine of the photographed person is healthy, the curve characteristics among the sub image blocks of the connected region may not be obvious or the curve trends in all the sub image blocks of the connected region are substantially the same, and for this case, whether the connected region image is vertically oriented may be further detected, and if so, the connected region image may be directly determined as the spine region image (spine ROI).

In addition, the number of slices for a connected region image block may be determined based on the number of spine blocks, for example, the slicing principle may be such that at least 3 spine blocks are included in each connected region sub-image block. For example, the number of vertebras of a child at birth is 32-33, and the adult spine is formed by connecting 26 vertebrae (7 cervical vertebrae, 12 thoracic vertebrae, 5 lumbar vertebrae, 1 sacral vertebrae (5 just born), 1 caudal vertebra), ligaments, joints and intervertebral discs. It can be seen that the number of slices in the present application varies dynamically with the age of the subject, but as described above, it is preferable to include at least 3 spine blocks in each connected component, which is more advantageous for obtaining reliable curve characteristics. Therefore, the accuracy of spine region image identification can be further improved, and the measurement of the subsequent Cobb angle is further ensured.

inputting a deep learning identification model into the spine region image, wherein the deep learning identification model identifies each angular point and direction in the spine region image;

grouping the angular points with the same attribute, and constructing a boundary by taking each angular point in the group as a corner so as to finish marking each spine block; wherein the same attribute means that the openings of the corner points have the same orientation, and the corner points in the group include at least two.

In the embodiment of the present application, deep learning is a generic term of a class of pattern analysis methods, and tasks using deep learning as a theoretical technique include image classification, object detection, entity Recognition, Optical Character Recognition (OCR), and the like. The method and the device utilize a deep learning identification model (for example, the deep learning identification model based on the CNN) to identify the angular points in the spine region image, then the angular points are grouped according to the orientation of each angular point, the angular points in the group are the real angular points of each spine block, and the boundaries of the spine blocks can be constructed based on the angular points, so that the spine blocks are segmented.

Wherein the "same orientation" means that different corner points point (i.e. opening direction) in the same direction (substantially the center position of the spine block), and at least two corner points in the grouping must be satisfied. When there are two corner points in the group, it needs to be satisfied that the two corner points satisfy a diagonal relationship, that is, the two corner points are on a diagonal; when the number of the corner points in the grouping is three or four, the boundary contour of the complete spine block can be directly determined, and the segmentation is realized. In the case of three corner points, the fourth corner point may not be identified due to adhesion or the like, but three corner points having the same orientation may also determine the contour boundary of the spine block. In addition, when grouping is specifically implemented, since the angular points on different spine blocks also have the same or similar orientation, in order to distinguish between them, it is also necessary to define that all the angular points of the orientation lines of all the angular points in the group are located within a predetermined circle, and for the diameter of the circle, the diameter is set based on the width of the spine region image, for example, 1/10 with the diameter being the width can be set.

In order to ensure the recognition accuracy of the deep learning recognition model, a training set with enough spine region images for training needs to be preset to train the deep learning recognition model, and the spine region images in the training set need to be manually marked in advance on the corner points of the spine blocks. Although manual labeling needs more manpower, the initial labeling is usually only needed once, the deep learning recognition model can improve the recognition accuracy through self-learning subsequently, for the recognition result, verification (or obvious recognition error) and labeling can also be performed manually, then the spine region images which are checked out of recognition error and are subjected to manual labeling are periodically combined into a wrong recognition case training set, the wrong recognition case training set is used for performing medium-term training on the deep learning recognition model to optimize the model parameters, so that the deep learning recognition model can adapt to various spine region images after a plurality of times of training, and meanwhile, the recognition accuracy can be ensured. Of course, the labeled spine X-ray image can be directly used for training.

In addition, due to the influence of the precision of the shooting equipment, the shooting angle, the irregular standing posture of the shot person and the like, the spine block can be adhered in the spine X-ray image, so the adhesion condition can also exist in the binarized X-ray image, and the accurate measurement of the Cobb angle can be influenced. In view of this problem, the method of the present application further includes the steps of:

calculating the area of each identified spinal block, and if the area of a certain spinal block is more than twice of the area of any other spinal block, determining that the spinal block is adhered;

detecting whether a cavity exists in the vertebral column block, and if so, dividing by taking the cavity as a boundary;

if not, the area of the most adjacent spine block with the area smaller than that of the spine block is equally divided.

Wherein, if the inner part of the vertebral column block is provided with a cavity, the conglutination is shown but the conglutination is not complete, and the vertebral column block can be directly cut by taking the cavity as a boundary; on the contrary, because the size change of the adjacent spine blocks is gradual, the difference of the area sizes is not obvious, and equal segmentation can be performed on the basis of the areas of the adjacent spine blocks with smaller areas. The adhesion problem has been better solved to above-mentioned scheme.

In the embodiment of the application, after the spinal block is identified, the spinal curve can be obtained through curve fitting, and then the starting vertebra and the terminal vertebra for calculating the Cobb angle can be determined through analyzing the change trend of the slope of the curve. Obviously, the starting and ending vertebrae are the two spinal blocks with the largest inclination degree of the concave or convex side in the lateral bend.

Furthermore, the scoliosis is often complicated, that is, there are many possible scoliosis, and therefore, after the starting terminal vertebra is determined, the subsequent slope analysis should be continued by using the ending terminal vertebra as the starting terminal vertebra, so as to identify the remaining scoliosis and calculate the corresponding Cobb angle. Finally, the largest one of the plurality of Cobb angle values is taken as the Cobb angle value for determining scoliosis.

Wherein, the calculation formula of the Cobb angle value is as follows:

wherein k1 is the slope of the spine curve at the starting end vertebra, and k2 is the slope of the spine curve at the ending end vertebra.

Optionally, the method further comprises:

In the embodiment of the present application, the scheme of the present application is not limited to automatically determining whether scoliosis occurs, and further includes classifying and identifying the specific type of scoliosis.

Scoliosis also includes virtually many types: type 1 is a breast curve, type 2 is an upper breast curve, type 3 is a lower breast curve, type 4 is a whole breast curve, type 5 is a waist curve, and type 6 is a waist curve plus a breast curve. Therefore, when the condition of scoliosis of the photographed person is judged, a scoliosis classification step is needed, namely, the scoliosis is determined to belong to which kind of scoliosis, so that the workload of a doctor can be further reduced, and the corresponding treatment scheme can be beneficially called for the doctor and the patient to refer to. In specific implementation, a classification model based on a Support Vector Machine (SVM) may be designed, the spine curve is input into the trained classification model based on the SVM, and the attribution relationship with the various scoliosis categories is output.

Example two

Referring to fig. 2, fig. 2 is a schematic structural diagram of a medical platform according to an embodiment of the present application. As shown in fig. 2, a medical platform according to an embodiment of the present application includes a detection module, a processing module, and an output module;

the detection module is used for detecting scoliosis by adopting the method in the first embodiment;

the output module is used for outputting the treatment scheme.

In the embodiment of the application, after the image department utilizes the X-ray image of the spine of the examiner shot by the X-ray examination device, the examination module can quickly identify whether the spine is curved or not, and can even output the specific type of the spine. Then, the processing module calls out an optimal treatment scheme based on the preset incidence relation and outputs the optimal treatment scheme, so that effective detection and treatment reference can be provided for doctors, diagnosis and treatment workload of the doctors can be greatly reduced, and meanwhile, for patients, the self detection condition can be obtained more timely.

Optionally, the outputting the treatment plan comprises:

In addition to the above listed outputs, other compliant outputs may be selected, such as a database of medical information at various national/provincial levels, etc., which facilitates the national macro-analysis of scoliosis conditions.

Optionally, the medical platform further includes at least one of a reservation module, an information module, a mall module, a user module, an online diagnosis and treatment module, and an activity and sharing module.

Wherein, the reservation module: the client is supported to make the reservation on line through the background server, and the self-service center and the cooperation unit are supported to access to accept the reservation. In addition, channel reservation account numbers can be provided for the self-operation center and the cooperation units, namely, the self-operation center and the cooperation units are supported to reserve for channel clients.

An information module: the client is provided with pathological knowledge, health knowledge (such as relevant public number, video number) regularly, and also provides dynamically updated physician information, company information, certification authority information, industry dynamic information, and the like.

A mall module: diagnosis and treatment products, household equipment, medical equipment, online/offline courses and the like are provided online, and information such as inventory, discount and the like can be operated in a background.

A user module: the system comprises the functions of member information application and query, diagnosis and treatment information query, order management, expense payment, logistics query, point management, case and report management and the like.

The online diagnosis and treatment module comprises: providing the functions of diagnosis and treatment encyclopedia, automatic filing, on-line inquiry and question answering and the like.

An activity and sharing module: and providing functions of pushing the preferential activity information and the like.

Certainly, besides the functional modules, a plurality of blank functional modules and corresponding program interfaces can be further arranged to meet the subsequent new business requirements of the clients, so that the medical platform has better expandability.

EXAMPLE III

Referring to fig. 3, fig. 3 is an electronic device disclosed in the embodiment of the present application, which is characterized in that: the apparatus comprises:

a memory storing executable program code;

a processor coupled with the memory;

the processor calls the executable program code stored in the memory to execute the method according to the first embodiment.

Example four

The embodiment of the application also discloses a computer storage medium, wherein a computer program is stored on the storage medium, and when the computer program is executed by a processor, the method according to the first embodiment is executed.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

The above description is only for the specific embodiments of the present application, but the scope of the present application is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present application should be covered within the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A scoliosis detection method is characterized in that: the method comprises the following steps:

s1, acquiring a spine X-ray image;

s3, calculating a Cobb angle based on the identified spinal column blocks;

2. The method of claim 1, wherein: before the spine X-ray image is input into a deep learning identification model for identification, the method further comprises a preprocessing step, and the preprocessing step comprises the following steps:

3. The method of claim 2, wherein: inputting the spine X-ray image into a deep learning identification model for identification, and segmenting each spine block, wherein the method comprises the following steps:

4. A method according to claim 1 or 3, characterized in that: calculating a Cobb angle based on the identified spinal column blocks, comprising:

5. The method of claim 4, wherein: the method further comprises the following steps:

6. A computer storage medium having a computer program stored thereon, the computer program, when executed by a processor, performing the method of any one of claims 1-5.

7. An electronic device comprising a processor and a memory, said memory having stored thereon a computer program which, when executed by the processor, performs the method of any of claims 1-5.

8. A medical platform comprises a detection module, a processing module and an output module;

the detection module is used for detecting scoliosis by adopting the method of any one of claims 1 to 5;

the output module is used for outputting the treatment scheme.

9. The platform of claim 8, wherein: the outputting the treatment plan comprises:

10. The platform of claim 8 or 9, wherein: the medical platform further comprises at least one of a reservation module, an information module, a mall module, an online diagnosis and treatment module and an activity and sharing module.