Method and device for evaluating health degree of cervical vertebra

Abstract

The invention relates to the technical field of cervical vertebra diagnosis, in particular to a cervical vertebra health degree evaluation method and device. The method comprises the steps of obtaining an ultrasonic image and an MRI image of the cervical vertebra of a user, performing feature extraction on the ultrasonic image and the MRI image to obtain ultrasonic features and MRI features, performing feature fusion on the ultrasonic features and the MRI features to obtain a feature matrix of the cervical vertebra health degree, inputting the feature matrix into a trained cervical vertebra health degree evaluation model, and determining the cervical vertebra health degree; wherein the trained cervical vertebra health degree evaluation model is obtained by taking the feature matrix of the known cervical vertebra health degree and the cervical vertebra health degree corresponding to the feature matrix of the known cervical vertebra health degree as a sample training classifier. Therefore, the cervical vertebra health degree of the user can be accurately evaluated.

Description

Technical Field

[0001] The present invention relates to the technical field of cervical vertebra diagnosis, and in particular to a method and device for evaluating the health of the cervical vertebra. Background Art

[0002] In the current era of rapid information development, while my country enjoys the convenience it brings, it also faces a prominent public health issue: a high incidence of cervical spondylosis among young people. The widespread use of information technology has transformed people's learning, living, and working habits. The frequent use of electronic devices, such as smartphones, has led people of all ages to maintain a bowed head posture for extended periods, with this condition being particularly prevalent among young people. Prolonged bowing significantly increases cervical spine load. This abnormal stress can easily lead to cervical muscle strain, which in turn can trigger cervical degeneration. Over time, this can reduce cervical stability and function, ultimately leading to cervical spondylosis. The incidence of cervical musculoskeletal injuries is also increasing among young people, who are prone to the unhealthy habit of bowing their heads. Clarifying the pathogenic mechanism of bowing in cervical musculoskeletal degeneration is crucial for the prediction and intervention of cervical spondylosis disorders. Therefore, intelligent, precise, and scientific multi-faceted exploration of the clinical symptoms, signs, and pathological evaluation mechanisms of long-term bowing in the head can help guide young people in developing healthy habits, prevent chronic musculoskeletal injuries, and enable early prediction and targeted intervention for cervical degeneration, ultimately reducing the incidence of cervical spondylosis in my country and improving national health.

[0003] Based on this, the present invention proposes a method and device for evaluating the health of the cervical spine to solve the above technical problems. Summary of the Invention

[0004] The present invention describes a method and device for evaluating the health of the cervical spine, which can accurately evaluate the health of the user's cervical spine.

[0005] According to a first aspect, the present invention provides a method for evaluating cervical vertebra health, the method comprising:

[0006] Obtaining ultrasound images and MRI images of the user's cervical spine;

[0007] performing feature extraction on the ultrasound image and the MRI image respectively to obtain ultrasound features and MRI features;

[0008] Fusing the ultrasound features and the MRI features to obtain a feature matrix of cervical vertebra health;

[0009] Inputting the feature matrix into a trained cervical vertebra health evaluation model to determine the cervical vertebra health;

[0010] The trained cervical vertebra health evaluation model is obtained by training a classifier using a feature matrix of known cervical vertebra health and the cervical vertebra health corresponding to the feature matrix of known cervical vertebra health as samples.

[0011] According to a second aspect, the present invention provides a device for evaluating cervical vertebra health, comprising:

[0012] an acquisition unit configured to acquire an ultrasound image and an MRI image of a user's cervical spine;

[0013] a first data processing unit, configured to perform feature extraction on the ultrasound image and the MRI image respectively to obtain ultrasound features and MRI features;

[0014] a second data processing unit configured to perform feature fusion on the ultrasound features and the MRI features to obtain a feature matrix of cervical vertebra health;

[0015] a third data processing unit, configured to input the feature matrix into a trained cervical vertebra health evaluation model to determine cervical vertebra health;

[0016] The trained cervical vertebra health evaluation model is obtained by training a classifier using a feature matrix of known cervical vertebra health and the cervical vertebra health corresponding to the feature matrix of known cervical vertebra health as samples.

[0017] In a third aspect, an embodiment of this specification further provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method described in any embodiment of this specification is implemented.

[0018] In a fourth aspect, an embodiment of this specification further provides a computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to execute the method described in any embodiment of this specification.

[0019] The cervical spine health assessment method and device provided by the present invention utilizes ultrasound imaging and magnetic resonance imaging (MRI) technologies to simultaneously capture medical imaging data of the user's cervical spine. Computer vision algorithms are used to perform feature extraction on both the ultrasound and MRI images: texture analysis, edge detection, and morphological feature extraction are performed on the ultrasound images to obtain ultrasound feature vectors; multimodal grayscale features, spatial structure features, and tissue contrast features are extracted from the MRI images to obtain an MRI feature matrix. Subsequently, a deep feature fusion algorithm is used to integrate the ultrasound and MRI features across modalities. Through feature alignment, weighted fusion, and dimensionality reduction, a feature matrix for cervical spine health is constructed. This feature matrix is then input into a trained cervical spine health assessment model. This model, based on a supervised learning framework, uses a large number of annotated feature matrices of known cervical spine health values and their corresponding clinically diagnosed cervical spine health values as training samples. The model is constructed by optimizing classifier parameters (e.g., support vector machines, convolutional neural networks, or other classification models) to accurately determine the user's cervical spine health status. This technical solution significantly improves the accuracy and reliability of cervical spine health assessment through multimodal imaging data fusion and intelligent model analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0021] Figure 1 A schematic flow chart of a method for evaluating cervical vertebra health according to one embodiment is shown;

[0022] Figure 2 shows a schematic block diagram of a device for evaluating cervical vertebra health according to one embodiment;

[0023] Figure 3 A schematic diagram showing the position of the fifth vertebra of the cervical spine according to one embodiment;

[0024] Figure 4 A schematic structural diagram of a device for measuring cervical vertebrae data in a head-down posture according to one embodiment is shown;

[0025] Figure 5 A schematic diagram of an MRI image according to one embodiment is shown.

[0026] 1- First sliding scale;

[0027] 11-first sliding member;

[0028] 12-first sliding main scale;

[0029] 2- Second sliding scale;

[0030] 21- second sliding member;

[0031] 22- Second sliding main scale;

[0032] 3-Head limiting component. DETAILED DESCRIPTION

[0033] The solution provided by the present invention is described below with reference to the accompanying drawings.

[0034] Figure 1 A flow chart of a method for evaluating cervical spine health according to one embodiment is shown. It is understood that the method can be executed by any device, equipment, platform, or equipment cluster with computing and processing capabilities. Figure 1 As shown, the method includes:

[0035] Step 100: Acquire an ultrasound image and an MRI image of the user's cervical spine;

[0036] Step 102: extract features from the ultrasound image and the MRI image to obtain ultrasound features and MRI features;

[0037] Step 104: Fusing the ultrasound features and the MRI features to obtain a feature matrix of cervical spine health;

[0038] Step 106: Input the feature matrix into the trained cervical vertebra health evaluation model to determine the cervical vertebra health; wherein the trained cervical vertebra health evaluation model is obtained by training a classifier using the feature matrix of known cervical vertebra health and the cervical vertebra health corresponding to the feature matrix of known cervical vertebra health as samples.

[0039] In this embodiment, ultrasound imaging and magnetic resonance imaging (MRI) technologies are used to simultaneously capture medical imaging data of the user's cervical spine. Computer vision algorithms are used to perform feature extraction on both the ultrasound and MRI images: texture analysis, edge detection, and morphological feature extraction are performed on the ultrasound images to obtain ultrasound feature vectors; multimodal grayscale features, spatial structure features, and tissue contrast features are extracted from the MRI images to obtain an MRI feature matrix. Subsequently, a deep feature fusion algorithm is used to integrate the ultrasound and MRI features across modalities. Through feature alignment, weighted fusion, and dimensionality reduction, a feature matrix for cervical spine health is constructed. This feature matrix is then input into a trained cervical spine health assessment model. This model, based on a supervised learning framework, uses a large number of annotated feature matrices of known cervical spine health values and their corresponding clinically determined cervical spine health values as training samples. The model is constructed by optimizing classifier parameters (e.g., support vector machines, convolutional neural networks, or other classification models). Through model inference and calculation, the user's cervical spine health status is ultimately accurately determined. This technical solution significantly improves the accuracy and reliability of cervical spine health assessment through multimodal imaging data fusion and intelligent model analysis.

[0040] In one embodiment of the present invention, MRI features are extracted through the following steps:

[0041] Segmenting the MRI image to obtain a plurality of MRI sub-images; wherein the MRI sub-images include an image of the intervertebral region, an image of the articular process region, an image of the prevertebral muscle region, an image of the deep layer of the posterior cervical muscle group, and an image of the superficial layer of the posterior cervical muscle group;

[0042] Extract features from multiple MRI sub-images to obtain multiple MRI sub-features;

[0043] Multiple MRI sub-features are fused to obtain MRI features.

[0044] In this embodiment, first, an advanced and adaptive image segmentation algorithm is used to perform a segmentation operation on the MRI image. This process fully considers the anatomical structural characteristics of the cervical spine and the grayscale, texture and other information of the image, and uses methods such as threshold segmentation, region-based segmentation, edge-based segmentation or deep learning segmentation models to accurately segment the complete MRI image into multiple MRI sub-images with clear anatomical significance. These sub-images cover the image of the intervertebral area, which can clearly present the shape, height and signal characteristics of the intervertebral space and the intervertebral disc; the image of the articular process area, which can show the structure, space and bone condition of the articular process joint; the image of the prevertebral muscle area, which is used to observe the shape, size and signal intensity of the prevertebral muscle; the image of the deep area of the posterior cervical muscle group, which can reflect the state of the deep posterior cervical muscle group; and the image of the superficial area of the posterior cervical muscle group, which can reflect the characteristics of the superficial posterior cervical muscle group. Then, detailed feature extraction is performed on each MRI sub-image. According to the characteristics of the area represented by the sub-image, a variety of feature extraction methods are used. For example, for morphological features, the area, perimeter, and circularity of tissue within the sub-image are calculated; for grayscale features, statistics such as mean, variance, and skewness are extracted; and for texture features, texture information is extracted using methods such as gray-level co-occurrence matrices and local binary patterns. Through these methods, multiple MRI sub-features reflecting the tissue characteristics of the region are extracted from each MRI sub-image. Finally, a suitable feature fusion strategy is employed to fuse these multiple MRI sub-features. Linear fusion methods based on weighted summation can be used, assigning different weights to each sub-feature based on its importance in assessing cervical spine health. Alternatively, nonlinear fusion methods based on machine learning, such as principal component analysis (PCA) and independent component analysis (ICA), can be employed to project multiple sub-features into a low-dimensional space, thereby obtaining a comprehensive MRI feature that fully reflects the overall characteristics of the cervical spine. The MRI features obtained in this manner can provide richer and more accurate information for subsequent cervical spine health assessment.

[0045] like Figure 5As shown in this embodiment, the red area in the figure represents the image of the corresponding intervertebral area. The core structure of the intervertebral area is the intervertebral disc, which is located in a key position between the vertebrae in the image. Image characteristics: On MRI and other images, the signal of the nucleus pulposus of a normal intervertebral disc is relatively high, while the signal of the annulus fibrosus is slightly lower. If degeneration or herniation occurs, the signal and morphology will change. The green area represents the image of the corresponding articular process area. The articular process joints are composed of the upper and lower articular processes of adjacent vertebrae and are located on both sides of the posterior vertebral body in the cervical spine anatomy. The yellow area represents the image of the prevertebral muscle area. The prevertebral muscles include the longus colli and longus capitis muscles and are located in front of the cervical vertebrae. The blue area represents the image of the deep layer of the posterior cervical muscles. The deep layer of the posterior cervical muscles, such as the semispinalis capitis and semispinalis cervicis, plays an important role in maintaining cervical stability and movement. Image characteristics: MRI can better display the muscle morphology and signal condition. If the muscle is strained or damaged, the signal will show abnormal changes. The functional status of the muscle can be assessed by observing its morphology and signal. The purple area in the figure represents the image of the superficial layer of the posterior cervical muscles. Muscles like the trapezius belong to the superficial muscles of the posterior neck group. They are located superficially and cover a large area.

[0046] In this embodiment, the impact of relevant MRI features on cervical spine health can be determined by analyzing MRI images and using the Pfirrmann grading scale for intervertebral disc degeneration shown in Table 1. As shown in Table 1, cervical spine health is inversely proportional to the Pfirrmann grading scale for intervertebral disc degeneration; that is, the higher the grading scale, the worse the cervical spine health.

[0047] Table 1

[0048]

[0049]

[0050] In one embodiment of the present invention, the ultrasound image is acquired through the following steps:

[0051] Obtaining the position of the fifth vertebra of the user's cervical spine; wherein the position of the fifth vertebra is the reference midpoint for measuring the ultrasound image;

[0052] Based on the position of the fifth vertebra, determine the horizontal parallel line on the surface of the neck skin; wherein the horizontal parallel line is the positioning reference line for ultrasound image detection;

[0053] An ultrasound image is acquired based on the position of the fifth vertebra and a transverse horizontal parallel line; wherein the ultrasound image includes the trapezius muscle area, the splenius capitis muscle area, the semispinalis capitis muscle area, the semispinalis cervicis muscle area, and the multifidus muscle area.

[0054] In this embodiment, first, the anatomical position of the user's fifth cervical vertebra (C5) is accurately identified and marked by palpation combined with imaging-assisted positioning technology (such as X-rays, MRI pre-scans, etc.). Figure 5As shown in the figure, the C5 vertebrae, a key anatomical landmark (red dot in the image), is established as the spatial reference center for subsequent ultrasound image measurements, providing a stable three-dimensional positioning reference for subsequent operations. Secondly, based on ergonomic and anatomical principles, a precise horizontal parallel line is drawn transversely on the surface of the neck skin using a spirit level or a high-precision positioning device, using the C5 vertebrae center as a reference. This transverse horizontal parallel line serves as the core positioning reference line for ultrasound image detection, ensuring the consistency and repeatability of the ultrasound probe scanning path and effectively reducing measurement errors. Finally, the long axis of the ultrasound probe is precisely aligned with the C5 vertebrae center, and the long axis of the probe is kept parallel to the transverse horizontal parallel line. Ensure that the probe is in close contact with the skin surface and there are no air bubbles interfering, and multi-directional and multi-angle ultrasound scans are performed along the predetermined scanning path. Based on this positioning system, complete ultrasound image data of the neck musculature, including the trapezius, splenius capitis, semispinalis capitis, semispinalis cervicis, and multifidus muscles, is acquired, providing high-resolution imaging for subsequent muscle morphology, structure, and function analysis.

[0055] In this embodiment, the detection positions and functions of the trapezius muscle area, the splenius capitis muscle area, the semispinalis capitis muscle area, the semispinalis cervicis muscle area, and the multifidus muscle area each have their own characteristics, as shown in Table 2 below:

[0056] Table 2

[0057]

[0058] In one embodiment of the present invention, after inputting the feature matrix into the trained cervical vertebra health evaluation model to determine the cervical vertebra health, the method further includes:

[0059] Obtain longitudinal and transverse cervical spine data of the user from an upright posture to a head-down posture;

[0060] Determine the correction factor of cervical spine health based on longitudinal cervical spine data and transverse cervical spine data;

[0061] The cervical vertebra health is corrected according to the correction coefficient to obtain the corrected cervical vertebra health.

[0062] In this embodiment, considering that the assessment of cervical spine health relies on data collected by electronic medical devices, there is inevitably the possibility of errors introduced due to improper operation during manual operation, which will affect the accuracy of the cervical spine health assessment. Therefore, in order to ensure that the cervical spine health assessment results can more accurately reflect the user's actual cervical spine condition, the inventors creatively thought of correcting the initially obtained cervical spine health with the help of a correction coefficient. The specific operating steps are as follows: Obtain relevant data of the cervical spine when the user changes from an upright posture to a bowed posture. The device can record the longitudinal and transverse displacement changes of the cervical spine respectively, thereby obtaining longitudinal cervical spine data and transverse cervical spine data. The longitudinal cervical spine data can reflect the morphological changes of the cervical spine in the vertical direction, such as the degree of vertebral flexion, changes in the height of the intervertebral space, etc.; the transverse cervical spine data reflects the horizontal movement of the cervical spine, such as the scoliosis amplitude and rotation angle of the cervical spine.

[0063] Based on the longitudinal and transverse cervical spine data, a correction factor is determined: the initially determined cervical spine health is multiplied by the calculated correction factor to correct the cervical spine health, ultimately obtaining the corrected cervical spine health. This correction method effectively reduces the impact of human error on the assessment results, allowing the corrected cervical spine health to more accurately reflect the user's actual cervical spine health level, providing a more reliable basis for subsequent medical diagnosis and rehabilitation treatment planning.

[0064] In one embodiment of the present invention, the correction coefficient is determined by the following formula:

[0065]

[0066] Where K is the correction coefficient, δ is the preset posture change coefficient, α is the preset longitudinal cervical data adjustment parameter, and w L is the preset first weight coefficient, s L is the preset second weight coefficient, ΔL is the transverse cervical vertebra data, β is the preset transverse cervical vertebra data adjustment parameter, and w T is the preset third weight coefficient, s T is the preset fourth weight coefficient, ΔT is the longitudinal cervical vertebra data, and γ is a constant greater than zero.

[0067] In this embodiment, K is a correction coefficient used to quantify the health of the cervical spine. The larger the K value, the better the health of the cervical spine when the head is lowered relative to the upright posture; the smaller the K value, the worse the health of the cervical spine. δ is a preset posture change coefficient, which is adjusted according to the degree of change from the upright posture to the lowered posture. When the degree of change is large, δ is appropriately increased to highlight the impact of the change on the health of the cervical spine; when the degree of change is small, δ is close to 1. α is a preset longitudinal cervical data adjustment parameter. The preset longitudinal cervical data adjustment parameter is a constant greater than 0 and is used to adjust the overall influence of the longitudinal data in the formula. It can be set based on clinical experience or data analysis. w L is the preset first weight coefficient, which is greater than 0, reflecting the importance of the change in longitudinal cervical data in assessing cervical health. The larger the weight, the greater the impact of longitudinal data on cervical health assessment. L is the preset second weight coefficient. When the longitudinal change direction (such as the flexion direction of the cervical spine when bowing the head) is unfavorable to the health of the cervical spine (such as excessive flexion), the preset second weight coefficient is greater than 1; when the change direction is normal or favorable, the preset second weight coefficient is less than 1. ΔL is the transverse cervical data, and β is the preset transverse cervical data adjustment parameter. The preset transverse cervical data adjustment parameter is a constant greater than 0 and is used to adjust the overall influence of the transverse data in the formula. It should be reasonably set according to actual conditions. T is the preset second weight coefficient. When it is greater than 0, it indicates the importance of the change in transverse cervical spine data in assessing cervical spine health. The larger the weight, the more critical the role of transverse data in cervical spine health assessment. T It is the preset fourth weight coefficient. When the lateral change direction (such as the possible scoliosis direction of the cervical spine when lowering the head) is unfavorable to the health of the cervical spine (such as excessive scoliosis), the preset fourth weight coefficient is greater than 1; when the change direction is normal or favorable, the preset fourth weight coefficient is less than 1. ΔT is the longitudinal cervical spine data, and γ is a constant greater than zero. Its function is to prevent the denominator from being zero, to ensure the stability and effectiveness of the formula, and at the same time, the value range of the correction coefficient K can be adjusted to make K more accurately reflect the health status of the cervical spine.

[0068] In one embodiment of the present invention, the longitudinal cervical vertebra data and the transverse cervical vertebra data are obtained by a device for measuring cervical vertebra data in a head-down posture:

[0069] The measuring device includes a first sliding scale 1, a second sliding scale 2 and a head limiting component 3. The first sliding scale 1 includes a first sliding component 11 and a first sliding main scale 12. The second sliding scale 2 includes a second sliding component 21 and a second sliding main scale 22.

[0070] The first sliding component 11 is connected to the head limiting component 3, the first sliding main scale 12 is connected to the second sliding component 21, and the first sliding scale 1 and the second sliding scale 2 are arranged in a vertical and orthogonal space;

[0071] When the user changes from an upright posture to a head-down posture, the first sliding component 11 can generate a lateral displacement along the sliding track of the first sliding main scale 12, so as to use the first sliding main scale 12 to measure the user's lateral cervical vertebrae data from the upright posture to the head-down posture; the second sliding component 21 can generate a longitudinal displacement along the sliding track of the second sliding main scale 22, so as to use the second sliding main scale 22 to measure the user's longitudinal cervical vertebrae data from the upright posture to the head-down posture;

[0072] The head limiting component 3 is used to limit the rotation of the user's head.

[0073] like Figure 4 As shown, in this embodiment, the measuring device includes a first sliding scale 1, a second sliding scale 2 and a head limiting component 3. The head limiting component 3 is firmly connected to the first sliding component 11 to ensure the accuracy of the measurement reference; the first sliding main scale 12 is nested with the second sliding component 21, and the first sliding scale 1 and the second sliding scale 2 are arranged in a vertical and orthogonal space to form a two-dimensional measurement coordinate system. When the user changes from an upright posture to a head-down posture, the first sliding component 11 will produce a lateral displacement along the slide rail. Its linear change from the initial position to the measurement end position can reflect the activity data of the cervical spine in the lateral direction; at the same time, the second sliding component 21 will produce a longitudinal displacement. The change in the distance between its initial position and the measurement end position can obtain the activity parameters of the cervical spine in the vertical direction. Through this dual-scale orthogonal collaborative measurement mechanism, the dynamic data changes of the cervical spine during the head-down movement can be captured in an all-round and high-precision manner. The head limiter 3 effectively limits the left and right rotation and pitch deviation of the head, providing a stable reference for cervical spine data measurement, avoiding measurement errors caused by unexpected head movement, and thus ensuring the accuracy and reliability of data acquired by the first sliding scale 1 and the second sliding scale 2. This application can accurately and efficiently complete comprehensive measurement of cervical spine data in the user's head-down posture, providing reliable data support for health assessment.

[0074] In one embodiment of the present invention, the cervical spine health evaluation model is a convolutional neural network model.

[0075] In this embodiment, the cervical spine health assessment model uses a convolutional neural network (CNN) model. As a powerful deep learning architecture, convolutional neural networks are particularly suitable for processing data with spatial structural characteristics such as medical images. Specifically, the convolutional neural network model uses a series of convolutional layers, pooling layers, and fully connected layers. The convolutional layer uses convolution kernels to slide on the data for convolution operations, which can effectively capture local features and patterns in the image, such as identifying specific tissue morphology and texture features in MRI images and ultrasound images of the cervical spine. The pooling layer is used to downsample the data, reduce the data dimension, reduce the amount of calculation, and prevent overfitting. After multiple layers of feature extraction and transformation, the fully connected layer integrates the previously extracted features and continuously adjusts the network parameters through the backpropagation algorithm and optimizer (such as stochastic gradient descent) based on the known cervical spine health feature matrix in the training sample and its corresponding actual cervical spine health label, so that the model can accurately classify or regress the input cervical spine health-related features, thereby determining the health level of the cervical spine. This cervical spine health assessment model based on convolutional neural networks, with its powerful feature learning capabilities and efficient processing capabilities of medical imaging data, can more accurately explore the potential relationship between cervical spine health status and imaging features compared to traditional methods, providing more accurate and reliable support for the assessment of cervical spine health.

[0076] The foregoing description describes specific embodiments of the present invention. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different from that described in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order shown or the sequential order to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0077] According to another embodiment, the present invention provides a device for evaluating the health of the cervical spine. Figure 2 A schematic block diagram of a device for evaluating cervical spine health according to one embodiment is shown. It is understood that the device can be implemented by any device, equipment, platform, or device cluster with computing and processing capabilities. Figure 2 As shown, the device includes: an acquisition unit 200, a first data processing unit 202, a second data processing unit 204 and a third data processing unit 206. The main functions of each component unit are as follows:

[0078] An acquisition unit 200 is configured to acquire an ultrasound image and an MRI image of the user's cervical spine;

[0079] A first data processing unit 202 is configured to extract features from the ultrasound image and the MRI image to obtain ultrasound features and MRI features;

[0080] The second data processing unit 204 is configured to perform feature fusion on the ultrasound features and the MRI features to obtain a feature matrix of cervical vertebra health;

[0081] The third data processing unit 206 is configured to input the feature matrix into the trained cervical vertebra health evaluation model to determine the cervical vertebra health;

[0082] The trained cervical vertebra health evaluation model is obtained by training a classifier using a feature matrix of known cervical vertebra health and the cervical vertebra health corresponding to the feature matrix of known cervical vertebra health as samples.

[0083] In one embodiment of the present invention, the MRI features are extracted by the following steps:

[0084] Segmenting the MRI image to obtain a plurality of MRI sub-images; wherein the MRI sub-images include an image of an intervertebral region, an image of an articular process region, an image of a prevertebral muscle region, an image of a deep region of the posterior cervical muscle group, and an image of a superficial region of the posterior cervical muscle group;

[0085] Extract features from multiple MRI sub-images to obtain multiple MRI sub-features;

[0086] Multiple MRI sub-features are fused to obtain MRI features.

[0087] In one embodiment of the present invention, the ultrasound image is acquired through the following steps:

[0088] Obtaining the position of the fifth vertebral segment of the user's cervical spine; wherein the position of the fifth vertebral segment is the reference midpoint of the ultrasound image;

[0089] Based on the position of the fifth vertebral segment, a horizontal parallel line on the surface of the neck skin is determined; wherein the horizontal parallel line is a positioning reference line for ultrasound image detection;

[0090] Based on the position of the fifth vertebral segment and the transverse horizontal parallel line, the ultrasound image is acquired; wherein the ultrasound image includes the trapezius muscle area, the splenius capitis muscle area, the semispinalis capitis muscle area, the semispinalis cervicis muscle area and the multifidus muscle area

[0091] In one embodiment of the present invention, after inputting the feature matrix into a trained cervical vertebra health evaluation model to determine the cervical vertebra health, the method further includes:

[0092] Obtain longitudinal and transverse cervical spine data of the user from an upright posture to a head-down posture;

[0093] determining a correction coefficient of cervical vertebra health based on the longitudinal cervical vertebra data and the transverse cervical vertebra data;

[0094] The cervical vertebra health is corrected according to the correction coefficient to obtain a corrected cervical vertebra health.

[0095] In one embodiment of the present invention, the correction coefficient is determined by the following formula:

[0096]

[0097] Where K is the correction coefficient, δ is the preset posture change coefficient, α is the preset longitudinal cervical data adjustment parameter, and w L is the preset first weight coefficient, s L is the preset second weight coefficient, ΔL is the transverse cervical vertebra data, β is the preset transverse cervical vertebra data adjustment parameter, and w T is the preset third weight coefficient, s T is a preset fourth weight coefficient, ΔT is the longitudinal cervical vertebra data, and γ is a constant greater than zero.

[0098] In one embodiment of the present invention, the longitudinal cervical vertebra data and the transverse cervical vertebra data are obtained by a device for measuring cervical vertebra data in a head-down posture:

[0099] The measuring device includes a first sliding scale 1, a second sliding scale 2 and a head limiting component 3. The first sliding scale 1 includes a first sliding component 11 and a first sliding main scale 12. The second sliding scale 2 includes a second sliding component 21 and a second sliding main scale 22.

[0100] The first sliding component 11 is connected to the head limiting component 3, the first sliding main scale 12 is connected to the second sliding component 21, and the first sliding scale 1 and the second sliding scale 2 are arranged in a vertical and orthogonal space;

[0101] When the user changes from an upright posture to a head-down posture, the first sliding component 11 can generate a lateral displacement along the sliding track of the first sliding main scale 12, so as to use the first sliding main scale 12 to measure the user's lateral cervical vertebrae data from the upright posture to the head-down posture; the second sliding component 21 can generate a longitudinal displacement along the sliding track of the second sliding main scale 22, so as to use the second sliding main scale 22 to measure the user's longitudinal cervical vertebrae data from the upright posture to the head-down posture;

[0102] The head limiting component 3 is used to limit the rotation of the user's head.

[0103] In one embodiment of the present invention, the cervical spine health evaluation model is a convolutional neural network model.

[0104] According to another embodiment, there is also provided a computer readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to execute a combination of Figure 1 The method described.

[0105] According to another embodiment, an electronic device is provided, comprising a memory and a processor, wherein the memory stores an executable code, and when the processor executes the executable code, the system realizes the combination of Figure 1 The method described.

[0106] The various embodiments of the present invention are described in a progressive manner. Similar portions between the various embodiments can be referred to in conjunction with each other. Each embodiment focuses on the differences between the other embodiments. In particular, the device embodiments are described briefly because they are generally similar to the method embodiments. For relevant portions, refer to the description of the method embodiments.

[0107] Those skilled in the art will appreciate that, in one or more of the above examples, the functions described herein may be implemented using hardware, software, firmware, or any combination thereof. When implemented using software, these functions may be stored in a computer-readable medium or transmitted as one or more instructions or codes on a computer-readable medium.

[0108] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent replacements, improvements, etc. made on the basis of the technical solution of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for evaluating the health of the cervical spine, characterized in that: The method comprises: Obtaining ultrasound images and MRI images of the user's cervical spine; performing feature extraction on the ultrasound image and the MRI image respectively to obtain ultrasound features and MRI features; Fusing the ultrasound features and the MRI features to obtain a feature matrix of cervical vertebra health; Inputting the feature matrix into a trained cervical vertebra health evaluation model to determine the cervical vertebra health; The trained cervical vertebra health evaluation model is obtained by training a classifier using a feature matrix of known cervical vertebra health and the cervical vertebra health corresponding to the feature matrix of known cervical vertebra health as samples.

2. The method according to claim 1, characterized in that The MRI features are extracted through the following steps: Segmenting the MRI image to obtain a plurality of MRI sub-images; wherein the MRI sub-images include an image of an intervertebral region, an image of an articular process region, an image of a prevertebral muscle region, an image of a deep region of the posterior cervical muscle group, and an image of a superficial region of the posterior cervical muscle group; Extract features from multiple MRI sub-images to obtain multiple MRI sub-features; Multiple MRI sub-features are fused to obtain MRI features.

3. The method according to claim 1, characterized in that The ultrasound image is obtained through the following steps: Obtaining the position of the fifth vertebral segment of the user's cervical spine; wherein the position of the fifth vertebral segment is the reference midpoint of the ultrasound image; Based on the position of the fifth vertebral segment, a horizontal parallel line on the surface of the neck skin is determined; wherein the horizontal parallel line is a positioning reference line for ultrasound image detection; The ultrasound image is acquired based on the position of the fifth vertebral segment and the transverse horizontal parallel line; wherein the ultrasound image includes the trapezius muscle area, the splenius capitis muscle area, the semispinalis capitis muscle area, the semispinalis cervicis muscle area, and the multifidus muscle area.

4. The method according to claim 1, wherein After inputting the feature matrix into the trained cervical vertebra health evaluation model to determine the cervical vertebra health, the method further includes: Obtain longitudinal and transverse cervical spine data of the user from an upright posture to a head-down posture; determining a correction coefficient of cervical vertebra health based on the longitudinal cervical vertebra data and the transverse cervical vertebra data; The cervical vertebra health is corrected according to the correction coefficient to obtain a corrected cervical vertebra health.

5. The method according to claim 4, characterized in that The correction factor is determined by the following formula: Where K is the correction coefficient, δ is the preset posture change coefficient, α is the preset longitudinal cervical data adjustment parameter, and w L is the preset first weight coefficient, s L is the preset second weight coefficient, ΔL is the transverse cervical vertebra data, β is the preset transverse cervical vertebra data adjustment parameter, and w T is the preset third weight coefficient, s T is a preset fourth weight coefficient, ΔT is the longitudinal cervical vertebra data, and γ is a constant greater than zero.

6. The method according to claim 5, characterized in that The longitudinal cervical vertebra data and the transverse cervical vertebra data are obtained by a device for measuring cervical vertebra data in a head-down posture: The measuring device comprises a first sliding scale (1), a second sliding scale (2) and a head limiting component (3); the first sliding scale (1) comprises a first sliding component (11) and a first main sliding scale (12); the second sliding scale (2) comprises a second sliding component (21) and a second main sliding scale (22); The first sliding component (11) is connected to the head limiting component (3), the first sliding main scale (12) is connected to the second sliding component (21), and the first sliding scale (1) and the second sliding scale (2) are arranged in a vertical and orthogonal spatial arrangement; When the user changes from an upright posture to a head-down posture, the first sliding component (11) can generate a lateral displacement along the sliding rail of the first sliding main scale (12), so as to measure the user's lateral cervical vertebra data from the upright posture to the head-down posture using the first sliding main scale (12); the second sliding component (21) can generate a longitudinal displacement along the sliding rail of the second sliding main scale (22), so as to measure the user's longitudinal cervical vertebra data from the upright posture to the head-down posture using the second sliding main scale (22); The head limiting component (3) is used to limit the rotation of the user's head.

7. The method according to claim 1, characterized in that The cervical spine health evaluation model is a convolutional neural network model.

8. A device for evaluating the health of the cervical spine, characterized in that: include: an acquisition unit configured to acquire an ultrasound image and an MRI image of a user's cervical spine; a first data processing unit, configured to perform feature extraction on the ultrasound image and the MRI image respectively to obtain ultrasound features and MRI features; a second data processing unit configured to perform feature fusion on the ultrasound features and the MRI features to obtain a feature matrix of cervical vertebra health; a third data processing unit, configured to input the feature matrix into a trained cervical vertebra health evaluation model to determine cervical vertebra health; The trained cervical vertebra health evaluation model is obtained by training a classifier using a feature matrix of known cervical vertebra health and the cervical vertebra health corresponding to the feature matrix of known cervical vertebra health as samples.

9. An electronic device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that A computer program is stored, and when the computer program is executed in a computer, the computer is caused to execute the method according to any one of claims 1 to 7.

Images