Scoliosis electrical stimulation correction system and spinal electrical stimulation parameter information prediction method

Abstract

Embodiments of the present disclosure disclose a scoliosis electrical stimulation correction system and a method for predicting scoliosis electrical stimulation parameter information. A specific embodiment of the system includes a server, a target client, and a scoliosis electrical stimulation flexible support battery box, wherein: the server is configured to perform the following steps: receiving user scoliosis feature data; generating scoliosis detection data; generating scoliosis type information; inputting the scoliosis detection data, the scoliosis type information, and user basic information into a pre-trained scoliosis electrical stimulation parameter prediction model; the target client is configured to perform the following steps: generating scoliosis electrical stimulation adjustment parameter information and wearing alignment point configuration information; the scoliosis electrical stimulation flexible support battery box is configured to control the activation of the position-adjusted electrical stimulation electrode sheet group and the alignment point assembly in the scoliosis electrical stimulation flexible support. This embodiment reduces the consumption of computing resources and reduces the waiting time of the user waiting for the prediction result.

Description

TECHNICAL FIELD

[0001] Embodiments of the present disclosure relate to the technical field of computer technology, in particular to a scoliosis electrical stimulation correction system and a scoliosis electrical stimulation parameter information prediction method. BACKGROUND

[0002] With the development of wearable device miniaturization, flexible electrodes and low-power chips make long-term attached electrical stimulation possible, and scoliosis electrical stimulation flexible braces have been widely used in the correction of adolescent idiopathic scoliosis (AIS). The scoliosis electrical stimulation correction system is a system for generating scoliosis electrical stimulation parameter information to control the electrical stimulation output corresponding to the scoliosis electrical stimulation parameter information in the scoliosis electrical stimulation flexible brace, so as to achieve the effect of correcting the scoliosis. At present, when generating the scoliosis electrical stimulation parameter information, the commonly used method is to predict the scoliosis electrical stimulation parameter information by using finite element simulation.

[0003] However, when the above method is used to generate the scoliosis electrical stimulation parameter information, the following technical problems often exist:

[0004] Predicting the scoliosis electrical stimulation parameter information by using finite element simulation requires simulating the coupling effect of vertebral body, intervertebral disc, ligament, muscle and other tissues, and the calculation amount increases exponentially, resulting in large consumption of computing resources and long waiting time for the user to wait for the prediction result.

[0005] The above information disclosed in this BACKGROUND section is only for the purpose of enhancing the understanding of the background of the present inventive concepts, and therefore, it can contain information that is not prior art known to those of ordinary skill in the art. SUMMARY

[0006] The summary of the present disclosure is intended to introduce the concepts in a simplified form, which will be described in detail in the specific embodiments section. The summary of the present disclosure is not intended to identify key or essential features of the claimed technical solutions, nor is it intended to be used to limit the scope of the claimed technical solutions.

[0007] Some embodiments of the present disclosure propose a scoliosis electrical stimulation correction system and a scoliosis electrical stimulation parameter information prediction method to solve one or more of the technical intentions mentioned in the background section.

[0008] In a first aspect, some embodiments of the present disclosure provide a scoliosis electrical stimulation correction system, comprising: a server, a target client, a spinal electrical stimulation flexible support battery box, wherein: the server is configured to perform the following steps: receiving user spinal feature data sent by the target client, wherein the user spinal feature data comprises user basic information, user spinal image data and user spinal dynamic data; constructing a three-dimensional spinal model based on the user spinal image data; generating scoliosis detection data based on the user spinal image data; generating scoliosis type information based on user spinal dynamic data; labeling the scoliosis detection data on the three-dimensional spinal model; inputting the scoliosis detection data, the scoliosis type information and the user basic information into a pre-trained spinal electrical stimulation parameter prediction model to obtain spinal electrical stimulation parameter information; transmitting the three-dimensional spinal model labeled with the scoliosis detection data, the scoliosis type information and the spinal electrical stimulation parameter information to the target client; the target client is configured to perform the following steps: in response to receiving the three-dimensional spinal model, the scoliosis type information and the spinal electrical stimulation parameter information sent by the server, displaying the three-dimensional spinal model, the scoliosis type information and the spinal electrical stimulation parameter information on a preset parameter adjustment and wearing alignment point configuration page; generating spinal electrical stimulation adjustment parameter information and wearing alignment point configuration information based on the interactive operation information of the target user and the preset parameter adjustment and wearing alignment point configuration page; sending the spinal electrical stimulation adjustment parameter information and the wearing alignment point configuration information to the spinal electrical stimulation flexible support battery box; the spinal electrical stimulation support battery box is configured to receive the spinal electrical stimulation adjustment parameter information and the wearing alignment point configuration information sent by the target client to control the start of the position-adjusted electrical stimulation electrode sheet group and the alignment point assembly in the spinal electrical stimulation flexible support.

[0009] In a second aspect, some embodiments of the present disclosure provide a spinal electrical stimulation parameter information prediction method, comprising: receiving user spinal feature data sent by the target client, wherein the user spinal feature data comprises user basic information, user spinal image data and user spinal dynamic data; constructing a three-dimensional spinal model based on the user spinal image data; generating scoliosis detection data based on the user spinal image data; generating scoliosis type information based on user spinal dynamic data; labeling the scoliosis detection data on the three-dimensional spinal model; inputting the scoliosis detection data, the scoliosis type information and the user basic information into a pre-trained spinal electrical stimulation parameter prediction model to obtain spinal electrical stimulation parameter information; transmitting the three-dimensional spinal model labeled with the scoliosis detection data, the scoliosis type information and the spinal electrical stimulation parameter information to the target client.

[0010] The above various embodiments of the present disclosure have the following beneficial effects: through the scoliosis electrical stimulation correction system of some embodiments of the present disclosure, the consumption of computing resources and the prediction time for users to wait for prediction results are reduced. Specifically, the reason for the large consumption of computing resources and the long prediction time for users to wait for prediction results is that predicting the spinal electrical stimulation parameter information by using finite element simulation requires simulating the coupling of multiple tissues such as vertebral bodies, intervertebral discs, ligaments, and muscles, and the calculation amount increases exponentially, resulting in large consumption of computing resources and long waiting time for users to wait for prediction results. Based on this, the scoliosis electrical stimulation correction system of some embodiments of the present disclosure includes a server, a target client, and a spinal electrical stimulation flexible support battery box, wherein: first, the server is configured to perform the following steps: first, receiving user spinal feature data sent by the target client, wherein the user spinal feature data includes user basic information, user spinal image data, and user spinal dynamic data. Second, based on the user spinal image data, a spinal three-dimensional model is constructed. Thus, the user's spinal three-dimensional model can be restored using the user's spinal image data. Based on the user spinal image data, spinal scoliosis detection data is generated. Thus, the spinal scoliosis detection data for predicting the spinal electrical stimulation parameter information can be obtained. Third, based on the user spinal dynamic data, spinal scoliosis type information is generated. Thus, the spinal scoliosis type information representing the type of spinal scoliosis can be generated. Third, the spinal scoliosis detection data is labeled on the spinal three-dimensional model. The spinal scoliosis detection data, the spinal scoliosis type information, and the user basic information are input into a pre-trained spinal electrical stimulation parameter prediction model to obtain spinal electrical stimulation parameter information. Thus, based on the spinal scoliosis detection data, the spinal scoliosis type information, and the user basic information, the spinal electrical stimulation parameter information can be predicted by the spinal electrical stimulation parameter prediction model, without simulating the coupling of multiple tissues such as vertebral bodies, intervertebral discs, ligaments, and muscles, avoiding the high complexity calculation of multiple tissue coupling in finite element simulation, and achieving exponential reduction in computing resource consumption. Fourth, the spinal three-dimensional model labeled with the spinal scoliosis detection data, the spinal scoliosis type information, and the spinal electrical stimulation parameter information are transmitted to the target client. Thus, the spinal three-dimensional model labeled with the spinal scoliosis detection data, the spinal scoliosis type information, and the spinal electrical stimulation parameter information can be transmitted to the target client, so that the user of the target client can intuitively understand the spinal scoliosis situation and obtain the spinal electrical stimulation parameter information for generating spinal electrical stimulation adjustment parameter information. The target client is configured to perform the following steps: first, in response to receiving the spinal three-dimensional model, the spinal scoliosis type information, and the spinal electrical stimulation parameter information sent by the server, displaying the spinal three-dimensional model, the spinal scoliosis type information, and the spinal electrical stimulation parameter information on a preset parameter adjustment and wearing alignment point configuration page.Thereby, the preset parameter adjustment and wearing alignment point configuration page for adjusting and configuring the spinal cord electrical stimulation parameter information and the wearing alignment point can be displayed. In the second step, the spinal cord electrical stimulation adjustment parameter information and the wearing alignment point configuration information are generated based on the interaction operation information of the target user with the preset parameter adjustment and wearing alignment point configuration page. Thereby, the target client can visualize the spinal cord three-dimensional model, the scoliosis type information and the spinal cord electrical stimulation parameter information, and generate the spinal cord electrical stimulation adjustment parameter information and the wearing alignment point configuration information for controlling the activation of the electrical stimulation electrode sheet group and the alignment point assembly. In the third step, the spinal cord electrical stimulation adjustment parameter information and the wearing alignment point configuration information are sent to the spinal cord electrical stimulation flexible support battery box. The spinal cord electrical stimulation flexible support battery box is configured to receive the spinal cord electrical stimulation adjustment parameter information and the wearing alignment point configuration information sent by the target client to control the activation of the electrical stimulation electrode sheet group and the alignment point assembly in the spinal cord electrical stimulation flexible support. Thereby, the activation of the electrical stimulation electrode sheet group and the alignment point assembly in the spinal cord electrical stimulation flexible support can be controlled according to the spinal cord electrical stimulation adjustment parameter information and the wearing alignment point configuration information to release electrical stimulation in the scoliosis area, so as to achieve the effect of scoliosis correction. In addition, before the activation of the electrical stimulation electrode sheet group and the alignment point assembly in the spinal cord electrical stimulation flexible support to release electrical stimulation in the scoliosis area, the spinal cord electrical stimulation parameter information is predicted by the server included in the scoliosis electrical stimulation correction system based on the scoliosis detection data, the scoliosis type information and the user basic information by using the spinal cord electrical stimulation parameter prediction model, without simulating the multi-tissue coupling effect of the vertebral body, intervertebral disc, ligament and muscle, avoiding the high complexity calculation of multi-tissue coupling in finite element simulation, realizing the exponential reduction of the consumption of computing resources, reducing the consumption of computing resources and reducing the waiting time of the user waiting for the prediction result. BRIEF DESCRIPTION OF DRAWINGS

[0011] The above and other features, advantages, and aspects of embodiments of the present disclosure will become more apparent by describing in detail exemplary embodiments thereof with reference to the attached drawings in which:

[0012] Figure 1 is an architecture diagram of an exemplary system of the scoliosis electrical stimulation correction system according to the present disclosure;

[0013] Figure 2 is a flowchart of some embodiments of the spinal cord electrical stimulation parameter information prediction method according to the present disclosure. DETAILED DESCRIPTION

[0014] Embodiments of the present disclosure will be described below in greater detail with reference to the accompanying drawings. While certain embodiments of the present disclosure are shown in the drawings, it is understood that the present disclosure can be embodied in various forms and should not be interpreted as being limited to the embodiments set forth herein. Rather, these embodiments are provided so that the present disclosure will be thoroughly and completely understood. It should be understood that the drawings and embodiments of the present disclosure are only for illustrative purposes and should not be construed as limiting the scope of protection of the present disclosure.

[0015] In addition, it should be further noted that only parts related to the present application are shown in the drawings for ease of description. The embodiments in the present disclosure and the features in the embodiments can be combined with each other without conflict.

[0016] It should be noted that the terms "first", "second", and the like mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not intended to limit the order or interdependence of the functions performed by these devices, modules or units.

[0017] It should be noted that the terms "one", "multiple" mentioned in the present disclosure are illustrative and not restrictive, and those skilled in the art should understand that unless otherwise explicitly stated in the context, it should be understood as "one or more".

[0018] The names of the messages or information exchanged between the devices in the embodiments of the present disclosure are only for illustrative purposes, and are not intended to limit the scope of the messages or information.

[0019] The present disclosure will be described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.

[0020] Figure 1 An exemplary system architecture 100 of a heterogeneous data exchange system to which some embodiments of the present disclosure can be applied is shown.

[0021] As shown in Figure 1 The system architecture 100 can include a server 102, a target client 101, and a spine electric stimulation flexible support battery box 103. The server 102 and the target client 101 can be connected by wired, wireless communication link or optical fiber cable, etc. The target client 101 and the spine electric stimulation flexible support battery box 103 can be connected by wired, wireless communication link or optical fiber cable, etc.

[0022] In some embodiments, the server 102 can be configured to perform the following steps:

[0023] First, receive user spine feature data sent by the target client. The user spine feature data includes user basic information, user spine image data, and user spine dynamic data. The user basic information can include at least one of age and gender. The user spine image data can include a full-length frontal image of the spine and at least one full-length lateral image of the spine. The full-length frontal image of the spine can be an AP image of the spine. Each of the at least one full-length lateral image of the spine can be a spine image parallel to the sagittal plane on the left side or the right side of the human body.

[0024] Second, construct a three-dimensional model of the spine based on the user spine image data. The three-dimensional model of the spine can be a digital three-dimensional model representing the spine of the human body.

[0025] In some optional implementations of some embodiments, the server 102 can construct a three-dimensional model of the spine based on the user spine image data by the following steps:

[0026] Based on the full-length frontal image of the spine and the at least one full-length lateral image of the spine included in the user spine image data, construct a three-dimensional model of the spine. In practice, the server can construct a three-dimensional model of the spine based on the full-length frontal image of the spine and the at least one full-length lateral image of the spine by a medical image three-dimensional reconstruction technology.

[0027] Third, generate spine scoliosis detection data based on the user spine image data.

[0028] In the process of using the technical solutions to solve the problems mentioned in the background, the following problems often occur:

[0029] When detecting and positioning the apex and the end vertebrae in the full-length frontal image of the spine included in the user spine image data to generate the spine scoliosis degree information, the user spine image data included in the full-length frontal image of the spine is detected by directly regressing the vertebral boundary box through the target detection network without considering the influence of the background image on the detection. Moreover, the boundary box output by the target detection network is a rectangular box information, which only contains the minimum circumscribed rectangular coordinates of the vertebral body. The four vertices of the boundary box are only the projection circumscribed points of the vertebral body in the image, which are different from the real corner points of the vertebral body. This leads to poor accuracy of the vertebral feature point position information containing the position information of the vertebral apex and the center point position information generated, and further leads to poor accuracy of the spine scoliosis detection data generated based on the vertebral feature point position information.

[0030] In the face of the above technical problems, the inventors decided to use the following solutions:

[0031] In some optional implementations of some embodiments, the server 102 can generate the scoliosis detection data based on the user spine image data by the following steps:

[0032] Firstly, a full-length frontal image of the spine included in the user spine image data is determined as a spine image to be detected.

[0033] Secondly, a spine region segmentation processing is performed on the spine image to be detected to obtain a spine region segmentation image. In practice, the server can perform the spine region segmentation processing on the spine image to be detected by using a pre-trained U-Net network to obtain the spine region segmentation image. The spine region segmentation image to be detected can be a spine image to be detected with background (e.g., an image where ribs and soft tissues are located) removed.

[0034] Thirdly, a vertebra detection and segmentation processing is performed on the spine region segmentation image to obtain each vertebra positioning information and each vertebra mask image. Each vertebra positioning information in the each vertebra positioning information corresponds to a corresponding vertebra mask image in the each vertebra mask image. The vertebra positioning information includes vertebra position information and a vertebra coding identifier. In practice, first, the server can detect each vertebra position information of each vertebra in the spine region segmentation image by using a pre-trained object detection model (e.g., a yolov8 object detection model, a Faster R-CNN object detection model, etc.). The vertebra position information can be each vertex coordinate of a vertebra bounding box. The vertebra position information can be represented by a left upper corner vertex coordinate of a vertebra bounding box, a right upper corner vertex coordinate of a vertebra bounding box, a left lower corner vertex coordinate of a vertebra bounding box, and a right lower corner vertex coordinate of a vertebra bounding box. Then, the server can code each vertebra from top to bottom according to the position corresponding to the vertebra position information to obtain a vertebra coding identifier (e.g., T1, T2, etc.). Next, for each vertebra position information in the each vertebra position information, the server can determine the vertebra position information and the vertebra coding identifier corresponding thereto as the vertebra positioning information to obtain the each vertebra positioning information. Then, the server can input the spine region segmentation image into a pre-trained Mask R-CNN model to obtain each vertebra mask image of each vertebra in the spine region segmentation image. Each vertebra mask image in the each vertebra mask image can be a pixel-level binary segmentation image of a spine vertebra. It should be noted that the vertebra mask image is completely consistent in size with the spine region segmentation image.

[0035] In the fourth step, based on the above-mentioned each vertebral mask image, each vertebral feature point position information of each vertebral body corresponding to the above-mentioned each vertebral mask image is generated. Each vertebral feature point position information in the above-mentioned each vertebral feature point position information includes vertebral apex position information and center point position information. In practice, for each vertebral mask image in the above-mentioned each vertebral mask image, the server can perform a bitwise AND operation on the vertebral mask image and the above-mentioned spinal region segmentation image to obtain a vertebral body image. Then, the server can perform corner point detection on the vertebral body image by Harris corner point detection to obtain the vertebral apex position information. The vertebral apex position information can be represented by the coordinates of each corner point of the vertebral body. Then, the server can take the average of each coordinate included in the vertebral apex position information as the center point position information. As an example, the detected vertebral apex position information can be {(35, 25), (65, 25), (65, 75), (35, 75)}, and the center point position information can be (50, 50). Alternatively, the server can perform contour detection processing on the vertebral body image to obtain vertebral contour information. The vertebral contour information includes each contour point coordinate. Then, the four contour point coordinates corresponding to the points with the minimum and maximum x coordinates and the minimum and maximum y coordinates can be determined as the vertebral apex position information. The average of each coordinate included in the vertebral apex position information is then taken as the center point position information.

[0036] In the fifth step, based on the above-mentioned each vertebral feature point position information, fitting spinal midline information is generated. In practice, the server can determine each center point position information included in the above-mentioned each vertebral feature point position information as a center point position information set. Then, by using a spline curve fitting technique, each vertebral feature point position information is fitted to obtain the fitting spinal midline information. The fitting spinal midline information can be represented by a function.

[0037] In the sixth step, based on the above-mentioned fitting spinal midline information, top vertebral point position information, upper end vertebral point position information, and lower end vertebral point position information are generated. In practice, the server can determine the coordinates of the curvature extreme point of the midline corresponding to the fitting spinal midline information as the top vertebral point position information. By traversing the midline from the top vertebral point upwards, a point where the curvature first significantly increases (e.g., curvature > 0.01 mm-1) from close to zero (e.g., curvature < 0.005 mm-1) is found, and the coordinates of the found point are determined as the upper end vertebral point position information. By traversing the midline from the top vertebral point downwards, a point where the curvature first significantly decreases (e.g., curvature > 0.01 mm-1) from close to zero (e.g., curvature < 0.005 mm-1) is found, and the coordinates of the found point are determined as the lower end vertebral point position information.

[0038] In the seventh step, based on the top vertebra point position information, the upper end vertebra point position information, the lower end vertebra point position information, and the vertebra positioning information, the apex cone positioning information, the end vertebra positioning information, and the scoliosis degree information are generated. In practice, first, the server can determine the vertebra position information included in each vertebra positioning information as the target vertebra position information. Then, for each target vertebra position information, in response to determining that the top vertebra point position information corresponds to a point within the region represented by the target vertebra position information, the target vertebra position information is determined as the top vertebra position information. The vertebra code corresponding to the top vertebra position information and the top vertebra position information are determined as the apex cone positioning information. In response to determining that the lower end vertebra point position information corresponds to a point within the region represented by the target vertebra position information, the target vertebra position information is determined as the lower end vertebra position information corresponding to the vertebra code, and the lower end vertebra position information is determined as the lower end vertebra positioning information. In response to determining that the upper end vertebra point position information corresponds to a point within the region represented by the target vertebra position information, the target vertebra position information is determined as the upper end vertebra position information corresponding to the vertebra code, and the upper end vertebra position information is determined as the upper end vertebra positioning information. The lower end vertebra positioning information and the lower end vertebra positioning information are determined as the end vertebra positioning information. In practice, the server can determine the difference between the top left corner vertex coordinates of the vertebra boundary box included in the upper end vertebra point position information and the top right corner vertex coordinates of the vertebra boundary box as the upper endplate vector. The difference between the bottom left corner vertex coordinates of the vertebra boundary box included in the lower end vertebra point position information and the bottom right corner vertex coordinates of the vertebra boundary box is determined as the lower endplate vector. Then, the upper endplate vector and the lower endplate vector are input into the Cobb angle calculation formula based on vector dot product to obtain the Cobb angle as the scoliosis degree information.

[0039] In the eighth step, the apex cone positioning information, the end vertebra positioning information, and the scoliosis degree information are determined as the scoliosis detection data.

[0040] The technical scheme and related content thereof serve as one inventive point of the embodiments of the present disclosure, and solve the technical problem of poor accuracy of generated vertebra feature point position information and poor accuracy of scoliosis detection data. Factors leading to poor accuracy of vertebra feature point position information and poor accuracy of scoliosis detection data are often as follows: when detecting the top cone and end vertebra in the full-length frontal image of the user's spine image data and generating scoliosis degree information, the target detection network directly regresses the vertebra bounding box to detect the full-length frontal image of the user's spine image data, without considering the influence of the background image on the detection, and the bounding box output by the target detection network is a rectangular box information, which only contains the minimum circumscribed rectangle coordinates of the vertebra, and the four vertices of the bounding box are only the projection circumscribed points of the vertebra in the image, which are different from the real corner points of the vertebra, resulting in poor accuracy of the generated vertebra feature point position information containing the vertebra top point position information and the center point position information, and further leading to poor accuracy of the scoliosis detection data generated based on the vertebra feature point position information. If the above factors are solved, the accuracy of the vertebra feature point position information and the scoliosis detection data can be improved. To achieve this effect, first, the full-length frontal image of the user's spine image data is determined as a to-be-detected spine image. Then, the to-be-detected spine image is subjected to spine region segmentation processing to obtain a to-be-detected spine region segmentation image. In this way, the spine region can be separated from the background, reducing the interference of the background (irrelevant tissue), and only retaining the to-be-detected spine region segmentation image of the key information such as the vertebra, intervertebral space, and pedicle. Subsequently, the spine region segmentation image is subjected to vertebra detection and segmentation processing to obtain each vertebra positioning information and each vertebra mask image, wherein each vertebra positioning information in the each vertebra positioning information corresponds to a corresponding vertebra mask image in the each vertebra mask image, and the vertebra positioning information includes vertebra position information and a vertebra coding identifier. In this way, each vertebra mask image used to generate the top vertebra point position information, the upper end vertebra point position information, and the lower end vertebra point position information can be obtained. Then, based on the each vertebra mask image, each vertebra feature point position information corresponding to the each vertebra mask image is generated, wherein each vertebra feature point position information in the each vertebra feature point position information includes vertebra top point position information and center point position information. In this way, the vertebra mask image can capture the actual geometric shape (such as trapezoidal, irregular quadrilateral) of the spinal vertebra, and the vertebra feature point position information generated based on the vertebra mask image can directly reflect the actual boundary of the spinal vertebra in the image, thereby making the accuracy of the generated each vertebra feature point position information higher. Based on the each vertebra feature point position information, fitting spine midline information is generated. Subsequently, based on the fitting spine midline information, the top vertebra point position information, the upper end vertebra point position information, and the lower end vertebra point position information are generated.After that, based on the above apex vertebra position information, the above upper end vertebra position information, the above lower end vertebra position information, and the above each vertebra positioning information, apex cone positioning information, end vertebra positioning information, and scoliosis degree information are generated. Then, the above apex cone positioning information, the above end vertebra positioning information, and the above scoliosis degree information are determined as scoliosis detection data. In this way, the scoliosis detection data with higher accuracy can be generated on the basis of the each vertebra feature point position information with higher accuracy through the above steps.

[0041] Fourthly, based on the user spine dynamic data, scoliosis type information is generated.

[0042] In some optional implementations of some embodiments, the server 102 can generate the scoliosis type information based on the user spine dynamic data through the following steps:

[0043] Firstly, the user spine dynamic data is determined as to-be-detected user spine dynamic data. The to-be-detected user spine dynamic data includes each dynamic spine image in each posture and spine motion state information. The spine motion state information can represent the maximum activity angle of the spine in each direction (forward bending, backward stretching, lateral bending, etc.). As an example, the spine motion state information can be “forward bending maximum activity angle: 60°, backward stretching maximum activity angle: 20°, left lateral bending maximum activity angle: 30°, right lateral bending maximum activity angle: 30°”.

[0044] Secondly, based on each spine image, each scoliosis degree information in each posture is generated. In practice, for each spine image in the above each spine image, the server can perform cobb angle detection on the spine image through a cobb angle measurement method based on an AiteekEngine algorithm model, and obtain the cobb angle corresponding to the spine image as the scoliosis degree information.

[0045] Thirdly, each scoliosis degree information in each posture is determined as each dynamic scoliosis degree information. Each dynamic scoliosis degree information in the above each dynamic scoliosis degree information corresponds to a corresponding one of the above each spine image. The above each posture can be standing, forward bending, right lateral bending, and left lateral bending.

[0046] In the fourth step, the dynamic scoliosis degree information and the spinal motion state information are input into a pre-trained scoliosis type prediction model to obtain scoliosis type information. The scoliosis type prediction model can be a pre-trained ResNet-LSTM network taking the dynamic scoliosis degree information and the spinal motion state information as input information and taking the scoliosis type information as output information. The scoliosis type information can represent the type of scoliosis (e.g., congenital scoliosis, postural scoliosis).

[0047] In the fifth step, the scoliosis detection data is labeled on the three-dimensional spinal model. In practice, the server can label the scoliosis detection data on the three-dimensional spinal model through a pre-set three-dimensional visualization rendering engine (e.g., Unity3D engine).

[0048] In the sixth step, the scoliosis detection data, the scoliosis type information, and the user basic information are input into a pre-trained spinal electrical stimulation parameter prediction model to obtain spinal electrical stimulation parameter information. The spinal electrical stimulation parameter prediction model includes an input layer, a feature extraction layer, an embedding layer, a feature fusion layer, a spinal electrical stimulation parameter prediction layer, and an output layer.

[0049] In some optional implementations of some embodiments, the server 102 can input the scoliosis detection data, the scoliosis type information, and the user basic information into a pre-trained spinal electrical stimulation parameter prediction model to obtain spinal electrical stimulation parameter information by the following steps:

[0050] In the first step, the scoliosis detection data, the scoliosis type information, and the user basic information are input into the input layer of the intent recognition model to obtain scoliosis detection data feature information corresponding to the scoliosis detection data, scoliosis type feature information corresponding to the scoliosis type information, and user basic feature information corresponding to the user basic information. The input layer can convert the input data into a vector. The scoliosis detection data feature information can be a vector representing the scoliosis detection data. The scoliosis type feature information can be a vector representing the scoliosis type information. The user basic feature information can be a vector representing the user basic information.

[0051] In the second step, the scoliosis detection data feature information and the user basic feature information are determined as the feature information to be extracted.

[0052] In the third step, the feature information to be extracted is input into the feature extraction layer to obtain extracted feature information. The feature extraction layer can be a convolutional layer that extracts and converts the feature information to be extracted. The extracted feature information can be a feature vector representing the scoliosis detection data feature information and the user basic feature information.

[0053] In the fourth step, the scoliosis type feature information is input into the embedding layer to obtain scoliosis type coding feature information. The embedding layer can be a neural network layer that takes the scoliosis type information (e.g., "congenital scoliosis" or "postural scoliosis") as input and outputs discrete classification variables (e.g., 0 for congenital scoliosis and 1 for postural scoliosis). The scoliosis type coding feature information can represent the classification variables of the scoliosis type feature information.

[0054] In the fifth step, the extracted feature information and the scoliosis type coding feature information are input into the feature fusion layer to obtain scoliosis feature fusion information. The feature fusion layer can be a concatenation layer or an attention mechanism layer that fuses the extracted feature information and the scoliosis type coding feature information (e.g., concatenation, weighted summation, attention mechanism, etc.). The scoliosis feature fusion information can be a fused vector.

[0055] In the sixth step, the scoliosis feature fusion information is input into the spinal cord electrical stimulation parameter prediction layer to obtain initial spinal cord electrical stimulation parameter prediction information. The spinal cord electrical stimulation parameter prediction layer can be a fully connected layer that predicts the initial spinal cord electrical stimulation parameter prediction information based on the scoliosis feature fusion information. The initial spinal cord electrical stimulation parameter prediction information can be a vector representing the first concave side electrical stimulation prediction information, the second concave side electrical stimulation prediction information, the first convex side electrical stimulation prediction information, and the second convex side electrical stimulation prediction information. The first concave side electrical stimulation prediction information can represent the parameter information for electrical stimulation below the upper end vertebra above the apex cone concave side. The second concave side electrical stimulation prediction information can represent the parameter information for electrical stimulation above the lower end vertebra below the apex cone concave side. The first convex side electrical stimulation prediction information can be the parameter information for electrical stimulation below the upper end vertebra above the apex cone convex side. The second convex side electrical stimulation prediction information can be the parameter information for electrical stimulation above the lower end vertebra below the apex cone convex side. The parameter information can include current intensity, frequency, and pulse width.

[0056] In the seventh step, the initial spinal cord electrical stimulation parameter prediction information is input into the output layer to obtain spinal cord electrical stimulation parameter information. The output layer can be a layer that converts the initial spinal cord electrical stimulation parameter prediction information (e.g., a floating-point number array with a length of 256) from an abstract feature vector into specific electrical stimulation parameters (e.g., current intensity, frequency, and pulse width).

[0057] Seventh, the spine three-dimensional model labeled with scoliosis detection data, the scoliosis type information and the spinal electrical stimulation parameter information are transmitted to the target client.

[0058] In some optional implementations of some embodiments, the server 102 can transmit the spine three-dimensional model labeled with scoliosis detection data, the scoliosis type information and the spinal electrical stimulation parameter information to the target client by the following steps:

[0059] First, the spine three-dimensional model labeled with scoliosis detection data, the scoliosis type information and the spinal electrical stimulation parameter information are determined as to-be-transmitted data.

[0060] Second, the to-be-transmitted data is compressed to obtain compressed to-be-transmitted data. In practice, the execution subject can use a lossless compression algorithm (for example, an octree encoding compression algorithm) to compress the to-be-transmitted data to obtain the compressed to-be-transmitted data.

[0061] Third, the compressed to-be-transmitted data is sent to the target client. The target client can be a target terminal (for example, a doctor terminal, that is, an electronic device or a software platform used by a doctor in the diagnosis and treatment process).

[0062] In some embodiments, the target client 101 can be configured to perform the following steps:

[0063] First, in response to receiving the spine three-dimensional model, the scoliosis type information and the spinal electrical stimulation parameter information sent by the server, the spine three-dimensional model, the scoliosis type information and the spinal electrical stimulation parameter information are displayed on a preset parameter adjustment and wearing alignment point configuration page. In practice, the target client can display the spine three-dimensional model in a first preset display box in the preset parameter adjustment and wearing alignment point configuration page, and can display the scoliosis type information and the spinal electrical stimulation parameter information in a second preset display box in the preset parameter adjustment and wearing alignment point configuration page. The preset parameter adjustment and wearing alignment point configuration page can be a preset page for the target user (for example, a doctor) to adjust and configure the displayed spinal electrical stimulation parameter information and the wearing alignment point.

[0064] Second, based on the interaction operation information of the target user and the preset parameter adjustment and wearing alignment point configuration page, the spinal electrical stimulation adjustment parameter information and the wearing alignment point configuration information are generated.

[0065] In some optional implementations of some embodiments, the target client 101 can generate the spinal cord electrical stimulation adjustment parameter information and the wearing alignment point configuration information based on the interaction operation information of the target user with the preset parameter adjustment and wearing alignment point configuration page by the following steps:

[0066] In response to detecting the modification operation on the spinal cord electrical stimulation parameter information displayed in the preset parameter adjustment and wearing alignment point configuration page, the modified spinal cord electrical stimulation parameter information is determined as the spinal cord electrical stimulation adjustment parameter information. The spinal cord electrical stimulation adjustment parameter information includes the apex electrical stimulation information, the first concave side electrical stimulation information, the second concave side electrical stimulation information, the first convex side electrical stimulation information, the second convex side electrical stimulation information, and the preset parameter adjustment and wearing alignment point configuration page includes the alignment point configuration interactive control. The apex electrical stimulation information can be the electrical stimulation parameter for electrical stimulation on the apex after the modification operation. The first concave side electrical stimulation information can be the electrical stimulation parameter for electrical stimulation on the upper end vertebra below the apex concave side above after the modification operation. The second concave side electrical stimulation information can be the electrical stimulation parameter for electrical stimulation on the lower end vertebra above the apex concave side below after the modification operation. The first convex side electrical stimulation information can be the electrical stimulation parameter for electrical stimulation on the upper end vertebra below the apex convex side above after the modification operation. The second convex side electrical stimulation information can be the electrical stimulation parameter for electrical stimulation on the lower end vertebra above the apex convex side below after the modification operation.

[0067] In response to detecting the interaction operation on the alignment point configuration interactive control, the information corresponding to the alignment point configuration interactive control is determined as the wearing alignment point configuration information. The alignment point configuration interactive control can be a light selection control or an alignment point current parameter information input box. The wearing alignment point configuration information can be information for controlling the activation of the alignment point component. The alignment point component can be a preset alignment point (e.g., a preset point on the navel eye part of the spinal cord electrical stimulation flexible support, etc.). The alignment point component can be an electrical sheet or a prompt light that releases alignment prompts.

[0068] In some embodiments, the spinal cord electrostimulation flexible support battery box 103 described above can be configured to receive the spinal cord electrostimulation adjustment parameter information and the wearing alignment point configuration information sent by the target client to control the activation of the position-adjusted electrostimulation electrode sheet group and the alignment point assembly in the spinal cord electrostimulation flexible support. Among them, the electrostimulation electrode sheet group includes the top vertebra point stimulation electrode, the top cone concave side upper end vertebra below electrode sheet, the top cone concave side lower end vertebra above electrode sheet, the top cone convex side upper end vertebra below electrode sheet and the top cone convex side lower end vertebra above electrode sheet. The top vertebra point stimulation electrode can be an electrode sheet installed at the corresponding top vertebra point in the spinal cord electrostimulation flexible support. The top cone concave side upper end vertebra below electrode sheet can be an electrode sheet installed at the corresponding top cone concave side upper end vertebra below position in the spinal cord electrostimulation flexible support. The top cone concave side lower end vertebra above electrode sheet can be an electrode sheet installed at the corresponding top cone concave side lower end vertebra above position in the spinal cord electrostimulation flexible support. The top cone convex side upper end vertebra below electrode sheet can be an electrode sheet installed at the corresponding top cone convex side upper end vertebra below position in the spinal cord electrostimulation flexible support. The top cone convex side lower end vertebra above electrode sheet can be an electrode sheet installed at the corresponding top cone convex side lower end vertebra above position in the spinal cord electrostimulation flexible support.

[0069] In some optional implementations of some embodiments, the spinal cord electrostimulation flexible support battery box 103 described above can receive the spinal cord electrostimulation adjustment parameter information and the wearing alignment point configuration information sent by the target client to control the activation of the position-adjusted electrostimulation electrode sheet group and the alignment point assembly in the spinal cord electrostimulation flexible support by the following steps:

[0070] First, in response to receiving the opening information sent by the preset control device, the following steps are performed:

[0071] First sub-step, based on the wearing alignment point configuration information, control the activation of the alignment point assembly to guide the wearer of the spinal cord electrostimulation flexible support to complete the wearing alignment of the spinal cord electrostimulation flexible support. The preset control device can be a remote controller for controlling the opening and closing of the spinal cord electrostimulation flexible support battery box. The spinal cord electrostimulation flexible support battery box is integrated with a microcontroller or a dedicated chip inside.

[0072] Second sub-step, based on the top cone electrostimulation information included in the spinal cord electrostimulation adjustment parameter information, control the top vertebra point stimulation electrode to release the current corresponding to the top cone electrostimulation information to prompt the area where the lateral bending part of the wearer of the spinal cord electrostimulation flexible support is located.

[0073] Third sub-step, based on the first concave side electrostimulation information included in the spinal cord electrostimulation adjustment parameter information, control the top cone concave side upper end vertebra below electrode sheet to release the current corresponding to the first concave side electrostimulation information to exercise the concave side paraspinal muscle below the top cone concave side upper end vertebra below electrode sheet.

[0074] A fourth sub-step, based on the second concave side electrical stimulation information included in the spinal cord electrical stimulation adjustment parameter information, controlling the electrical sheet above the lower end vertebra below the convex side of the top cone to release the electrical current corresponding to the second concave side electrical stimulation information, so as to exercise the concave side paraspinal muscle below the electrical sheet. Wherein, the first concave side electrical stimulation information and the second concave side electrical stimulation information can represent high frequency (such as 40-60Hz) electrical current (electrical stimulation) that can cause complete muscle tetanic contraction.

[0075] A fifth sub-step, based on the first convex side electrical stimulation information included in the spinal cord electrical stimulation adjustment parameter information, controlling the electrical sheet below the upper end vertebra above the convex side of the top cone to release the electrical current corresponding to the first convex side electrical stimulation information, so as to relax the convex side paraspinal muscle below the electrical sheet.

[0076] A sixth sub-step, based on the second convex side electrical stimulation information included in the spinal cord electrical stimulation adjustment parameter information, controlling the electrical sheet above the lower end vertebra below the convex side of the top cone to release the electrical current corresponding to the first convex side electrical stimulation information, so as to relax the convex side paraspinal muscle below the electrical sheet. Wherein, the first convex side electrical stimulation information and the second convex side electrical stimulation information can represent low frequency pulse current (such as 1-5Hz) that can cause single muscle contraction.

[0077] The technical solution and related content are an invention point of the embodiment of the present disclosure, which solves the technical problem of poor experience of the user of the spinal cord electric stimulation flexible support. The factors leading to poor experience of the user of the spinal cord electric stimulation flexible support are usually as follows: when the user wears the spinal cord electric stimulation flexible support for scoliosis correction, the top vertebra point stimulation electrode is not controlled to release the current for prompting the area where the scoliosis part of the wearer of the spinal cord electric stimulation flexible support is located, and the user of the spinal cord electric stimulation flexible support is not clear about the scoliosis part. At the same time, the start of the alignment point assembly is not controlled, and the correction force action point is prone to deviation when the spinal cord electric stimulation flexible support is worn, resulting in poor correction effect, and further leading to poor experience of the user of the spinal cord electric stimulation flexible support. If the above factors are solved, the effect of improving the experience of the user of the spinal cord electric stimulation flexible support can be achieved. In order to achieve this effect, first, in response to detecting the start information sent by the preset control device, the following steps are performed: in a first sub-step, based on the wearing alignment point configuration information, the start of the alignment point assembly is controlled to guide the wearer of the spinal cord electric stimulation flexible support to complete the wearing alignment of the spinal cord electric stimulation flexible support. Thus, the start of the alignment point assembly can be controlled to guide the wearer of the spinal cord electric stimulation flexible support to complete the wearing alignment of the spinal cord electric stimulation flexible support. In a second sub-step, based on the top cone electric stimulation information included in the spinal cord electric stimulation adjustment parameter information, the top vertebra point stimulation electrode is controlled to release the current corresponding to the top cone electric stimulation information to prompt the area where the scoliosis part of the wearer of the spinal cord electric stimulation flexible support is located. In a third sub-step, based on the first concave side electric stimulation information included in the spinal cord electric stimulation adjustment parameter information, the top cone concave side upper end vertebra below electric sheet is controlled to release the current corresponding to the first concave side electric stimulation information to exercise the concave side paraspinal muscle below the top cone concave side upper end vertebra below electric sheet. In a fourth sub-step, based on the second concave side electric stimulation information included in the spinal cord electric stimulation adjustment parameter information, the top cone concave side lower end vertebra above electric sheet is controlled to release the current corresponding to the second concave side electric stimulation information to exercise the concave side paraspinal muscle below the top cone concave side lower end vertebra above electric sheet. In a fifth sub-step, based on the first convex side electric stimulation information included in the spinal cord electric stimulation adjustment parameter information, the top cone convex side upper end vertebra below electric sheet is controlled to release the current corresponding to the first convex side electric stimulation information to relax the convex side paraspinal muscle below the top cone convex side upper end vertebra below electric sheet. In a sixth sub-step, based on the second convex side electric stimulation information included in the spinal cord electric stimulation adjustment parameter information, the top cone convex side lower end vertebra above electric sheet is controlled to release the current corresponding to the first convex side electric stimulation information to relax the convex side paraspinal muscle below the top cone convex side lower end vertebra above electric sheet. Also, when the user wears the spinal cord electric stimulation flexible support for scoliosis correction, the start of the alignment point assembly is controlled to guide the wearer of the spinal cord electric stimulation flexible support to complete the wearing alignment of the spinal cord electric stimulation flexible support, which reduces the possibility of deviation of the correction force action point, and further improves the correction effect.And control the top vertebra point stimulation electrode to release the current corresponding to the top cone electric stimulation information, to prompt the area where the wearer of the spine electric stimulation flexible support is located, and improve the experience of the user of the spine electric stimulation flexible support.

[0078] Figure 2 The flow 200 of some embodiments of a spine electric stimulation parameter information prediction method of a server included in the spine scoliosis electric stimulation correction system according to the present disclosure is shown. The spine electric stimulation parameter information prediction method includes the following steps:

[0079] Step 201, receiving user spine feature data sent by a target client.

[0080] In some embodiments, the execution subject of the spine electric stimulation parameter information prediction method (for example, the server included in the spine scoliosis electric stimulation correction system) can receive the user spine feature data sent by the target client. Wherein, the user spine feature data includes user basic information, user spine image data and user spine dynamic data.

[0081] Step 202, constructing a spine three-dimensional model based on the user spine image data.

[0082] In some embodiments, the execution subject can generate spine scoliosis detection data based on the user spine image data.

[0083] Step 203, generating spine scoliosis detection data based on the user spine image data.

[0084] In some embodiments, the execution subject can generate spine scoliosis detection data based on the user spine image data.

[0085] Step 204, generating spine scoliosis type information based on the user spine dynamic data.

[0086] In some embodiments, the execution subject can generate spine scoliosis type information based on the user spine dynamic data.

[0087] Step 205, labeling the spine scoliosis detection data on the spine three-dimensional model.

[0088] In some embodiments, the execution subject can label the spine scoliosis detection data on the spine three-dimensional model.

[0089] Step 206, inputting the spine scoliosis detection data, the spine scoliosis type information and the user basic information into a pre-trained spine electric stimulation parameter prediction model to obtain spine electric stimulation parameter information.

[0090] In some embodiments, the execution subject can input the above-mentioned scoliosis detection data, the above-mentioned scoliosis type information and the above-mentioned user basic information into a pre-trained scoliosis electrical stimulation parameter prediction model to obtain scoliosis electrical stimulation parameter information.

[0091] Step 207, transmitting the spine three-dimensional model labeled with scoliosis detection data, scoliosis type information and scoliosis electrical stimulation parameter information to the target client.

[0092] In some embodiments, the execution subject can transmit the spine three-dimensional model labeled with scoliosis detection data, the above-mentioned scoliosis type information and the above-mentioned scoliosis electrical stimulation parameter information to the target client.

[0093] The above various embodiments of the present disclosure have the following beneficial effects: through the scoliosis electrical stimulation correction system of some embodiments of the present disclosure, the consumption of computing resources and the prediction time for users to wait for prediction results are reduced. Specifically, the reason for the large consumption of computing resources and the long prediction time for users to wait for prediction results is that predicting the spinal electrical stimulation parameter information by using finite element simulation requires simulating the coupling of multiple tissues such as vertebral bodies, intervertebral discs, ligaments, and muscles, and the calculation amount increases exponentially, resulting in large consumption of computing resources and long waiting time for users to wait for prediction results. Based on this, the scoliosis electrical stimulation correction system of some embodiments of the present disclosure includes a server, a target client, and a spinal electrical stimulation flexible support battery box, wherein: first, the server is configured to perform the following steps: first, receiving user spinal feature data sent by the target client, wherein the user spinal feature data includes user basic information, user spinal image data, and user spinal dynamic data. Second, based on the user spinal image data, a spinal three-dimensional model is constructed. Thus, the user's spinal three-dimensional model can be restored using the user's spinal image data. Based on the user spinal image data, spinal scoliosis detection data is generated. Thus, the spinal scoliosis detection data for predicting the spinal electrical stimulation parameter information can be obtained. Third, based on the user spinal dynamic data, spinal scoliosis type information is generated. Thus, the spinal scoliosis type information representing the type of spinal scoliosis can be generated. Third, the spinal scoliosis detection data is labeled on the spinal three-dimensional model. The spinal scoliosis detection data, the spinal scoliosis type information, and the user basic information are input into a pre-trained spinal electrical stimulation parameter prediction model to obtain spinal electrical stimulation parameter information. Thus, based on the spinal scoliosis detection data, the spinal scoliosis type information, and the user basic information, the spinal electrical stimulation parameter information can be predicted by the spinal electrical stimulation parameter prediction model, without simulating the coupling of multiple tissues such as vertebral bodies, intervertebral discs, ligaments, and muscles, avoiding the high complexity calculation of multiple tissue coupling in finite element simulation, and achieving exponential reduction in computing resource consumption. Fourth, the spinal three-dimensional model labeled with the spinal scoliosis detection data, the spinal scoliosis type information, and the spinal electrical stimulation parameter information are transmitted to the target client. Thus, the spinal three-dimensional model labeled with the spinal scoliosis detection data, the spinal scoliosis type information, and the spinal electrical stimulation parameter information can be transmitted to the target client, so that the user of the target client can intuitively understand the spinal scoliosis situation and obtain the spinal electrical stimulation parameter information for generating spinal electrical stimulation adjustment parameter information. The target client is configured to perform the following steps: first, in response to receiving the spinal three-dimensional model, the spinal scoliosis type information, and the spinal electrical stimulation parameter information sent by the server, displaying the spinal three-dimensional model, the spinal scoliosis type information, and the spinal electrical stimulation parameter information on a preset parameter adjustment and wearing alignment point configuration page.Thus, the preset parameter adjustment and wearing alignment point configuration page for adjusting and configuring the spinal cord electrical stimulation parameter information and the wearing alignment point can be displayed. In the second step, the spinal cord electrical stimulation adjustment parameter information and the wearing alignment point configuration information are generated based on the interaction operation information of the target user with the preset parameter adjustment and wearing alignment point configuration page. Thus, the target client can visualize the spinal cord three-dimensional model, the scoliosis type information, and the spinal cord electrical stimulation parameter information, and generate the spinal cord electrical stimulation adjustment parameter information and the wearing alignment point configuration information for controlling the activation of the electrical stimulation electrode sheet group and the alignment point assembly. In the third step, the spinal cord electrical stimulation adjustment parameter information and the wearing alignment point configuration information are sent to the spinal cord electrical stimulation flexible support battery box. The spinal cord electrical stimulation flexible support battery box is configured to receive the spinal cord electrical stimulation adjustment parameter information and the wearing alignment point configuration information sent by the target client to control the activation of the electrical stimulation electrode sheet group and the alignment point assembly in the spinal cord electrical stimulation flexible support. Thus, the activation of the electrical stimulation electrode sheet group and the alignment point assembly in the spinal cord electrical stimulation flexible support can be controlled according to the spinal cord electrical stimulation adjustment parameter information and the wearing alignment point configuration information to release electrical stimulation in the scoliosis area, so as to achieve the effect of scoliosis correction. In addition, before the activation of the electrical stimulation electrode sheet group and the alignment point assembly in the spinal cord electrical stimulation flexible support to release electrical stimulation in the scoliosis area, the spinal cord electrical stimulation parameter information is predicted by the spinal cord electrical stimulation parameter prediction model based on the scoliosis detection data, the scoliosis type information, and the user basic information through the server included in the spinal cord electrical stimulation correction system, so as to avoid the high complexity calculation of multi-tissue coupling in finite element simulation, realize the exponential reduction of the consumption of computing resources, reduce the consumption of computing resources, and reduce the waiting time of the user waiting for the prediction result.

[0094] The above description is only some of the preferred embodiments of the present disclosure and an explanation of the principles of the technology applied. Those skilled in the art should understand that the scope of the application involved in the embodiments of the present disclosure is not limited to the technical solutions formed by the specific combination of technical features, and should also cover other technical solutions formed by any combination of technical features or their equivalent features without departing from the inventive concept. For example, the technical solutions formed by replacing the features with the technical features disclosed in the embodiments of the present disclosure (but not limited to) having similar functions.

Claims

1. A scoliosis electrical stimulation correction system, comprising: Server, target client, spinal electrocautery flexible support battery box, among which: The server is configured to perform the following steps: Receive user spinal feature data sent by the target client, wherein the user spinal feature data includes user basic information, user spinal imaging data and user spinal dynamic data; Based on the user's spinal imaging data, a three-dimensional model of the spine is constructed; Based on the user's spinal imaging data, scoliosis detection data is generated; Based on the user's dynamic spinal data, generate information on the type of scoliosis. The scoliosis detection data is annotated on the three-dimensional model of the spine; The scoliosis detection data, the scoliosis type information, and the user basic information are input into a pre-trained spinal electrical stimulation parameter prediction model to obtain spinal electrical stimulation parameter information. The three-dimensional model of the spine labeled with scoliosis detection data, the scoliosis type information, and the spinal electrical stimulation parameter information are transmitted to the target client. The target client is configured to perform the following steps: In response to receiving the three-dimensional model of the spine, the scoliosis type information, and the spinal electrical stimulation parameter information sent by the server, the three-dimensional model of the spine, the scoliosis type information, and the spinal electrical stimulation parameter information are displayed on the preset parameter adjustment and wear alignment point configuration page; Based on the interaction information between the target user and the preset parameter adjustment and wear alignment point configuration page, spinal electrical stimulation adjustment parameter information and wear alignment point configuration information are generated. The spinal electrical stimulation adjustment parameter information and the wear alignment point configuration information are sent to the spinal electrical stimulation flexible support battery box. The spinal electrostimulation support battery box is configured to receive spinal electrical stimulation adjustment parameter information and wear alignment point configuration information sent by the target client, so as to control the activation of the position-adjusted electrical stimulation electrode pad group and alignment point assembly in the spinal electrostimulation flexible support.

2. The scoliosis electrical stimulation correction system according to claim 1, wherein, The user's spinal image data includes a full-length frontal image of the spine and at least one full-length lateral image of the spine. The server is further configured to construct a three-dimensional model of the spine based on the user's spinal image data through the following steps: A three-dimensional model of the spine is constructed based on the user's spinal imaging data, including a full-length frontal image of the spine and at least one full-length lateral image of the spine.

3. The scoliosis electrical stimulation correction system according to claim 1, wherein, The server is further configured to transmit a three-dimensional spinal model annotated with scoliosis detection data, the scoliosis type information, and the spinal electrical stimulation parameter information to the target client through the following steps: The three-dimensional spinal model labeled with scoliosis detection data, the scoliosis type information, and the spinal electrical stimulation parameter information are identified as the data to be transmitted. The data to be transmitted is compressed to obtain compressed data to be transmitted; The compressed data to be transmitted is sent to the target client.

4. The scoliosis electrical stimulation correction system according to claim 1, wherein, The server is further configured to generate scoliosis type information based on the user's dynamic spinal data through the following steps: The user's spinal dynamic data is determined as the user's spinal dynamic data to be detected, wherein the user's spinal dynamic data to be detected includes dynamic spinal images and spinal motion state information under various postures. Based on the aforementioned spinal images, information on the degree of scoliosis of each spine in each posture is generated. Each spinal scoliosis degree information under each posture is determined as a dynamic spinal scoliosis degree information, wherein each dynamic spinal scoliosis degree information corresponds to a corresponding spinal image in each spinal image. The dynamic scoliosis degree information and the spinal motion state information are input into a pre-trained scoliosis type prediction model to obtain scoliosis type information.

5. The scoliosis electrical stimulation correction system according to claim 1, wherein, The spinal electrical stimulation parameter prediction model includes an input layer, a feature extraction layer, an embedding layer, a feature fusion layer, a spinal electrical stimulation parameter prediction layer, and an output layer. The server is further configured to input the scoliosis detection data, the scoliosis type information, and the user's basic information into the pre-trained spinal electrical stimulation parameter prediction model through the following steps to obtain spinal electrical stimulation parameter information: The scoliosis detection data, the scoliosis type information, and the user basic information are input into the input layer of the spinal electrical stimulation parameter prediction model to obtain scoliosis detection data feature information corresponding to the scoliosis detection data, scoliosis type feature information corresponding to the scoliosis type information, and user basic feature information corresponding to the user basic information. The scoliosis detection data feature information and the user basic feature information are identified as the feature information to be extracted; The feature information to be extracted is input into the feature extraction layer to obtain the extracted feature information; The scoliosis type feature information is input into the embedding layer to obtain the scoliosis type encoding feature information; The extracted feature information and the scoliosis type encoding feature information are input into the feature fusion layer to obtain scoliosis feature fusion information; The lateral curvature feature fusion information is input into the spinal electrical stimulation parameter prediction layer to obtain the initial spinal electrical stimulation parameter prediction information. The initial spinal electrical stimulation parameter prediction information is input into the output layer to obtain the spinal electrical stimulation parameter information.

6. The scoliosis electrical stimulation correction system according to claim 1, wherein, The target client is further configured to generate spinal electrical stimulation adjustment parameter information and wear alignment point configuration information based on the interaction information between the target user and the preset parameter adjustment and wear alignment point configuration page through the following steps: In response to detecting a modification operation of the spinal electrical stimulation parameter information displayed on the preset parameter adjustment and wear alignment point configuration page, the modified spinal electrical stimulation parameter information is determined as the spinal electrical stimulation adjustment parameter information. The spinal electrical stimulation adjustment parameter information includes apical condyle electrical stimulation information, first concave side electrical stimulation information, second concave side electrical stimulation information, first convex side electrical stimulation information, and second convex side electrical stimulation information. The preset parameter adjustment and wear alignment point configuration page includes an alignment point configuration interactive control. In response to detecting an interactive operation performed on the alignment point configuration interactive control, the information corresponding to the alignment point configuration interactive control is determined as the wear alignment point configuration information.

Images