Cervical vertebra X-ray film segmentation identification method for obtaining precise bone fracture treatment scheme

Abstract

The invention relates to the technical field of image processing, in particular to a cervical vertebra X-ray film segmentation identification method for acquiring a bone fracture precise treatment scheme. The method comprises the following steps: acquiring a cervical vertebra X-ray film of a bone fracture patient, and obtaining an image representation value according to the gray difference and gray value between each pixel point and a neighborhood pixel point; obtaining an edge feature value according to the position distribution of the pixel points and the image representation value; according to the gray difference between each pixel point and the pixel points around the pixel point and the edge characteristic value, obtaining the edge law degree, and further screening characteristic pixel points; and determining the position importance degree of each pixel point by combining the difference of the edge law degrees of the feature pixel points and the adjacent pixel points thereof, the edge law degrees, the relative position distribution of the feature pixel points and the similar longitudinal edge lines and the image representation value of each pixel point, and further dividing the pixel points to obtain a target area. According to the invention, the accuracy of different area division results of the cervical vertebra X-ray film is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a cervical spine X-ray film segmentation and identification method for obtaining an accurate treatment plan for bone injuries. Background Art

[0002] Cervical spondylosis is a common spinal disease that occurs in people of different ages. In recent years, it has gradually shown a trend of younger age groups and has become a global health problem. At present, the common clinical diagnosis method for cervical spondylosis is X-ray imaging, which has the characteristics of wide application range, low cost, low radiation dose to the human body, and is suitable for routine physical examinations of most patients. This diagnostic method can better help patients check the cervical sequence, curvature, bone development of vertebral bodies and accessories, and whether there are morphological variations such as fractures, slippage, and dislocation.

[0003] X-rays are essentially two-dimensional images that compress the three-dimensional cervical spine structure into a flat display, which may cause overlapping of anatomical structures at different levels. The boundaries between the cervical vertebrae and the intervertebral spaces are often not clear enough. Existing methods directly segment cervical spine X-rays of patients with bone injuries, resulting in low segmentation accuracy, which affects the accuracy of subsequent lesion analysis results by doctors. Summary of the invention

[0004] In order to solve the problem that the existing methods have low accuracy of segmentation results when segmenting different areas of cervical spine X-ray films of patients with bone injuries, the purpose of the present invention is to provide a cervical spine X-ray film segmentation and identification method for obtaining an accurate treatment plan for bone injuries. The technical solution adopted is as follows: The present invention provides a cervical spine X-ray film segmentation and identification method for obtaining a precise treatment plan for bone injuries, the method comprising the following steps: Obtain cervical spine radiographs for patients with bone injuries; According to the grayscale difference between each pixel point and the neighboring pixel points in the cervical spine X-ray film and the grayscale value of each pixel point, the image performance value of each pixel point is obtained; according to the position distribution of each pixel point on each quasi-transverse edge line and the image performance value, the edge feature value of each pixel point on each quasi-transverse edge line is obtained; according to the grayscale difference between each pixel point on each quasi-transverse edge line and the surrounding pixel points and the edge feature value, the edge regularity of the corresponding pixel point is obtained; The characteristic pixel points are selected based on the edge regularity; the position importance of each pixel point is determined by combining the difference in edge regularity between each characteristic pixel point and its adjacent pixel points, the edge regularity of each characteristic pixel point, the relative position distribution of each characteristic pixel point and the quasi-longitudinal edge line, and the image performance value of each pixel point; wherein the quasi-horizontal edge line and the quasi-longitudinal edge line are determined according to the inclination of the non-closed edge line; All pixels are divided using the importance of all positions to obtain the target area.

[0005] Preferably, the image performance value of each pixel is obtained according to the grayscale difference between each pixel and the neighboring pixel in the cervical spine X-ray and the grayscale value of each pixel, including: According to the grayscale difference between the candidate pixel and all the pixels in its neighborhood and the grayscale value of the candidate pixel, an image representation value of the candidate pixel is obtained, and the grayscale difference between the candidate pixel and all the pixels in its neighborhood and the grayscale value of the candidate pixel are both positively correlated with the image representation value; The candidate pixel point is any pixel point in the cervical spine X-ray film.

[0006] Preferably, obtaining the quasi-transverse edge line and the quasi-longitudinal edge line includes: The absolute value of the average slope of all pixel points on each non-closed edge line in the cervical spine X-ray film of patients with bone injuries was calculated and recorded as the inclination index of each non-closed edge line; A non-closed edge line with a tilt index less than a preset slope threshold is determined as a quasi-transverse edge line, and a non-closed edge line with a tilt index greater than or equal to the preset slope threshold is determined as a quasi-longitudinal edge line.

[0007] Preferably, obtaining the edge feature value of each pixel point on each quasi-lateral edge line according to the position distribution of each pixel point on each quasi-lateral edge line and the image performance value comprises: For each quasi-longitudinal edge line, draw a perpendicular line through the midpoint of each quasi-longitudinal edge line; Calculate the difference between the minimum value of the distance between the pixel to be analyzed and all vertical lines and half the length of the horizontal edge line where the pixel to be analyzed is located, and record it as the length difference; According to the image representation value of the pixel to be analyzed and the length difference, an edge feature value of the pixel to be analyzed is obtained, wherein the image representation value of the pixel to be analyzed is positively correlated with the edge feature value, and the length difference is negatively correlated with the edge feature value; The pixel point to be analyzed is any pixel point on any type of horizontal edge line.

[0008] Preferably, obtaining the edge regularity of the corresponding pixel point according to the grayscale difference between each pixel point and its surrounding pixels on each quasi-lateral edge line and the edge feature value includes: Calculate the average value of the grayscale value difference between the pixel to be analyzed and its adjacent pixels on the straight line where the pixel to be analyzed is located, and record it as the average grayscale difference of the pixel to be analyzed; The edge regularity of the pixel to be analyzed is obtained according to the edge feature value of the pixel to be analyzed and the grayscale average difference, wherein the edge feature value is positively correlated with the edge regularity, and the grayscale average difference is negatively correlated with the edge regularity.

[0009] Preferably, the screening of characteristic pixel points based on the edge regularity degree includes: determining pixel points whose edge regularity degree is greater than a preset regularity threshold as characteristic pixel points.

[0010] Preferably, the determining the position importance of each pixel by combining the difference in edge regularity between each feature pixel and its adjacent pixels, the edge regularity of each feature pixel, the relative position distribution of each feature pixel and the quasi-vertical edge line, and the image performance value of each pixel, comprises: The line connecting the feature pixel to be evaluated and the nearest endpoint of the quasi-vertical edge line closest to it is recorded as the reference line segment of the feature pixel to be evaluated; obtaining the rotation angle of the quasi-horizontal edge line where the feature pixel to be evaluated is located, with the horizontal right direction as the starting direction and rotated counterclockwise to the quasi-horizontal edge line where the feature pixel to be evaluated is located; if the rotation angle is less than 90 degrees, the angle characteristic value of the feature pixel to be evaluated is set to -1; if the rotation angle is greater than or equal to 90 degrees, the angle characteristic value of the feature pixel to be evaluated is set to 1; Obtaining a position importance factor of the feature pixel to be evaluated according to the difference in edge regularity between the feature pixel to be evaluated and its adjacent pixel points, the edge regularity of the feature pixel to be evaluated and the angle characteristic value; the feature pixel to be evaluated is any feature pixel point; Set the position importance factor of pixels other than feature pixels to 0; The position importance factor of each pixel and the image performance value are combined to determine the position importance of each pixel.

[0011] Preferably, the determining the position importance of each pixel by combining the position importance factor and the image performance value of each pixel includes: The position importance factor of each pixel and its image performance value are summed to determine the position importance of each pixel.

[0012] Preferably, the method of dividing all pixels by the importance of all positions to obtain the target area includes: All pixels in the cervical spine X-ray films of patients with bone injuries are clustered based on their positional importance, and the target area is extracted based on the clustering results.

[0013] The present invention has at least the following beneficial effects: The present invention first analyzes the image performance characteristics of each pixel point according to the grayscale difference between each pixel point and the neighboring pixel points and the grayscale value of each pixel point in the cervical spine X-ray film, obtains the image performance value, and then comprehensively considers the distribution of the cervical spine, and evaluates the edge regularity of the pixel point according to the position distribution of each pixel point on each quasi-lateral edge line and the grayscale difference between each pixel point and the surrounding pixel points on each quasi-lateral edge line, thereby screening out a plurality of characteristic pixel points, and then obtains the distribution change of the pixel points around the intervertebral space according to the different change characteristics on the edge of each vertebral body of the cervical spine, thereby evaluating the importance of the positions of the pixel points at different positions, and then dividing all the pixel points into a plurality of different regions based on the position importance, so as to facilitate the extraction of the target region. The method provided by the present invention realizes the effective segmentation of the vertebral body region in the image, improves the situation where the edges between the intervertebral spaces are difficult to distinguish or unclear, and improves the accuracy of the results of dividing different regions, providing support for subsequent medical diagnosis and helping doctors to identify cervical lesions. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0015] Figure 1 A flowchart of a cervical spine X-ray segmentation and identification method for obtaining an accurate treatment plan for bone injuries provided by an embodiment of the present invention; Figure 2 A structural block diagram of a cervical spine X-ray segmentation and identification system for obtaining an accurate treatment plan for bone injuries provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0016] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the cervical spine X-ray film segmentation and identification method for obtaining a precise treatment plan for bone injuries proposed in the present invention is described in detail below in combination with the accompanying drawings and preferred embodiments.

[0017] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0018] The specific scheme of the cervical spine X-ray film segmentation and identification method for obtaining a precise treatment plan for bone injuries provided by the present invention is described in detail below with reference to the accompanying drawings.

[0019] Example of a cervical spine X-ray segmentation and identification method for obtaining a precise treatment plan for bone injuries: The specific scenario targeted by this embodiment is: when testing the cervical spine of a patient with bone injury, in order to improve the accuracy of the test results, it is necessary to extract important areas from the entire cervical spine X-ray film of the patient with bone injury to assist the doctor in performing accurate testing.

[0020] This embodiment proposes a cervical spine X-ray film segmentation and identification method for obtaining an accurate treatment plan for bone injuries, such as Figure 1 As shown, the cervical spine X-ray film segmentation and identification method for obtaining a precise treatment plan for bone injuries in this embodiment includes the following steps: Step S1, obtaining a cervical spine X-ray of a patient with bone injury.

[0021] First, collect cervical spine X-rays of patients with bone injuries. Specifically, adjust the position of patients with bone injuries appropriately according to the required angle to ensure that the collected cervical spine X-rays of patients with bone injuries can clearly show the cervical spine structure. During the shooting process, an X-ray machine is used to generate images. Radiology technicians adjust exposure time, angle and other parameters as needed. Digital X-ray equipment is usually used, and the images are directly stored as electronic files. In digital X-ray equipment, image data will be converted into digital signals through detectors to generate electronic images, that is, X-rays of cervical spine of patients with bone injuries are obtained.

[0022] So far, this embodiment has obtained the cervical spine X-ray film of the patient with bone injury.

[0023] Step S2, according to the grayscale difference between each pixel and its neighboring pixels in the cervical spine X-ray film and the grayscale value of each pixel, obtain the image representation value of each pixel; according to the position distribution of each pixel on each quasi-transverse edge line and the image representation value, obtain the edge feature value of each pixel on each quasi-transverse edge line; according to the grayscale difference between each pixel on each quasi-transverse edge line and its surrounding pixels and the edge feature value, obtain the edge regularity of the corresponding pixel.

[0024] In the X-ray of the cervical spine, different types of tissues are contained, including bones, soft tissues, and intervertebral discs. The cervical spine structure is mainly composed of multiple vertebrae, which are arranged together to form the specific anatomical structure of the cervical spine. Due to its physiological characteristics, the density and composition of each tissue are different, resulting in obvious grayscale differences in the X-ray. The vertebral bodies of the cervical spine are mainly composed of bone tissue, which is a high-density tissue. In X-rays, bones usually appear as obvious high-density white shadows in the image due to their high calcium content and density. These shadows show clear vertebral boundaries. Relatively speaking, the display ability of soft tissues is poor, usually appearing as low-density shadows or almost invisible in X-rays. For example, muscles, ligaments, and intervertebral discs are mainly composed of water and organic matter, and their density is much lower than that of vertebrae, so they appear as darker areas on X-ray images. In the captured images, the intervertebral disc and the soft tissues nearby are lower-density areas, usually appearing as gray or darker areas. These density differences have different manifestations in different image areas, so the performance characteristics of each pixel in the image are first analyzed.

[0025] Specifically, the Canny edge detection algorithm is used to detect the cervical spine X-ray film of the bone injury patient to obtain all edge lines and edge pixels in the cervical spine X-ray film of the bone injury patient. The Canny edge detection algorithm is a prior art and will not be described in detail here.

[0026] Next, this embodiment is described by taking any edge pixel point in the cervical spine X-ray film of a bone injury patient as an example. Other edge pixel points in the cervical spine X-ray film of a bone injury patient can be processed using the method provided in this embodiment.

[0027] Specifically, any edge pixel point in the cervical spine X-ray film of a bone injury patient is recorded as a candidate pixel point, and the image performance value of the candidate pixel point is obtained based on the grayscale difference between the candidate pixel point and all the pixels in its neighborhood and the grayscale value of the candidate pixel point. The grayscale difference between the candidate pixel point and all the pixels in its neighborhood and the grayscale value of the candidate pixel point are positively correlated with the image performance value.

[0028] Among them, the positive correlation means that the dependent variable will increase as the independent variable increases, and the dependent variable will decrease as the independent variable decreases. It can be an additive relationship, a multiplicative relationship, etc., which is determined by practical applications.

[0029] In this embodiment, a specific calculation formula for the image performance value is given. The image performance value of the i-th pixel point in the cervical spine X-ray of a bone injury patient can be expressed as: in, represents the image performance value of the i-th pixel in the cervical spine X-ray of a patient with bone injury, represents the gray value of the i-th pixel, Represents the gray value of the jth pixel in the neighborhood of the i-th pixel, It represents the gray value when the number of pixels with the same gray level is the largest and the gray value is the largest in the cervical spine X-ray film of patients with bone injuries. n represents the number of pixels in the neighborhood of the i-th pixel. represents the absolute value sign, norm() represents the normalization function, Indicates the preset first adjustment parameter.

[0030] In this embodiment, the preset first adjustment parameter is introduced into the calculation formula of the image performance value to prevent the denominator from being 0. In this embodiment, the preset first adjustment parameter is 0.01. In specific applications, the implementer can set it according to specific circumstances. In this embodiment, the size of the neighborhood is eight neighborhoods. In specific applications, the implementer can set it according to specific circumstances.

[0031] It represents the grayscale difference between the ith pixel and the surrounding pixels. The larger the difference, the greater the degree of change of the pixel in the image. It is likely to be a pixel on the edge of the bone vertebra. It indicates the difference between the ith pixel and most of the pixels with higher grayscale values ​​in the image. The smaller the value, the closer the grayscale value of the ith pixel is to these pixels, and the more important the pixel is in the image, that is, the greater the image performance value.

[0032] Since the vertebral edges of the cervical vertebrae in the vertical direction are more obvious, after edge detection, a relatively clear and complete vertical edge of each vertebra can be obtained, the overall shape and distribution of each cervical vertebra can be observed, and the relationship between the edges connected to each edge can be analyzed to determine the distribution of the correct edge of the vertebra at each edge point.

[0033] Based on the above characteristics, the absolute value of the average slope of all pixel points on each non-closed edge line in the cervical spine X-ray film of the bone injury patient is calculated respectively, and recorded as the inclination index of each non-closed edge line; the non-closed edge line with an inclination index less than the preset slope threshold is determined as a quasi-lateral edge line, and the non-closed edge line with an inclination index greater than or equal to the preset slope threshold is determined as a quasi-lateral edge line. In this embodiment, the preset slope threshold is 1, and in specific applications, the implementer can set it according to specific circumstances.

[0034] For each quasi-longitudinal edge line, draw a perpendicular line through the midpoint of each quasi-longitudinal edge line.

[0035] Any pixel point on any type of horizontal edge line is recorded as a pixel point to be analyzed, and the difference between the minimum value of the distance between the pixel point to be analyzed and all vertical lines and half the length of the horizontal edge line where the pixel point to be analyzed is located is calculated and recorded as the length difference; based on the image performance value of the pixel point to be analyzed and the length difference, the edge feature value of the pixel point to be analyzed is obtained, and the image performance value of the pixel point to be analyzed is positively correlated with the edge feature value, and the length difference is negatively correlated with the edge feature value.

[0036] Among them, a positive correlation means that the dependent variable will increase as the independent variable increases, and the dependent variable will decrease as the independent variable decreases. It can be an additive relationship, a multiplicative relationship, etc., which is determined by actual application; a negative correlation means that the dependent variable will decrease as the independent variable increases, and the dependent variable will increase as the independent variable decreases. It can be a subtractive relationship, a division relationship, etc., which is determined by actual application.

[0037] In this embodiment, a specific calculation formula for the edge feature value is given, and the edge feature value of the u-th pixel point on the t-th quasi-lateral edge line can be expressed as: in, represents the edge feature value of the u-th pixel on the t-th horizontal edge line, represents the image performance value of the u-th pixel on the t-th horizontal edge line. It represents the minimum value of the distance between the u-th pixel point on the t-th horizontal edge line and all vertical lines. represents the total number of pixels on the t-th horizontal edge line, represents half of the total number of pixels on the t-th horizontal edge line. Indicates the preset second adjustment parameter.

[0038] In this embodiment, the preset second adjustment is introduced into the calculation formula of the edge feature value to prevent the denominator from being 0. In this embodiment, the preset second adjustment parameter is 0.01. In specific applications, the implementer can set it according to specific circumstances.

[0039] The larger the image performance value of the u-th pixel on the t-th horizontal edge line, the more likely the pixel is the pixel on the edge of the vertebral body. Under normal circumstances, the distance between the pixel on the edge of the vertebral body and the vertical line should be close to half the length of the vertical edge, while the pixel inside the vertebral body or other positions will not be close. The smaller the value is, the more likely it is that the feature point is located at the edge of the vertebral body, that is, the larger the edge feature value is.

[0040] Under normal circumstances, the X-ray image of the vertebral body is shown as two upper and lower transverse edges connected to the longitudinal edge to form a clear rectangular or elliptical outline, representing the morphology of the vertebral body. However, the imaging characteristics of the X-ray film may cause multiple transverse edges connected to the longitudinal edge to appear in the image, resulting in blurred and irregular images. This may be due to factors such as shooting angle, equipment limitations, or patient position. Therefore, when analyzing cervical spine X-rays, special attention should be paid to the structural changes and edges of the vertebral body edges.

[0041] In X-ray images, the grayscale value of a normal vertebral edge usually changes from light to dark from the edge to the inside. The grayscale value changes more gently and evenly toward the inside of the vertebra, showing the continuity and stability of the bone structure. In comparison, the grayscale value changes in the intervertebral space and other marginal areas appear more irregular. In the intervertebral space, due to the presence of soft tissue or nucleus pulposus, the absorption of X-rays is lower, and since the shape and position of the intervertebral disc may change, the distribution of grayscale values ​​may also show uneven characteristics, which makes these areas more likely to show blurred or irregular edges in X-rays.

[0042] Next, the pixel to be analyzed is still used as an example for explanation. Specifically, the average value of the difference between the grayscale values ​​of the pixel to be analyzed and its adjacent pixels on the straight line where the pixel to be analyzed is located is calculated, and recorded as the average grayscale difference of the pixel to be analyzed; according to the edge feature value of the pixel to be analyzed and the average grayscale difference, the edge regularity of the pixel to be analyzed is obtained, the edge feature value is positively correlated with the edge regularity, and the average grayscale difference is negatively correlated with the edge regularity.

[0043] Among them, a positive correlation means that the dependent variable will increase as the independent variable increases, and the dependent variable will decrease as the independent variable decreases. It can be an additive relationship, a multiplicative relationship, etc., which is determined by actual application; a negative correlation means that the dependent variable will decrease as the independent variable increases, and the dependent variable will increase as the independent variable decreases. It can be a subtractive relationship, a division relationship, etc., which is determined by actual application.

[0044] In this embodiment, a specific calculation formula for the edge regularity is given. The edge regularity of the u-th pixel point on the t-th quasi-lateral edge line can be expressed as: in, Indicates the edge regularity of the u-th pixel on the t-th horizontal edge line. represents the edge feature value of the u-th pixel on the t-th horizontal edge line, It represents the average value of the difference between the grayscale values ​​of the u-th pixel and its adjacent pixels on the t-th horizontal edge line. Indicates the preset third adjustment parameter.

[0045] It should be noted that: in this embodiment, the method for obtaining the difference between the grayscale values ​​of two pixels is: taking the absolute value of the difference between the grayscale values ​​of the two pixels as the difference between the grayscale values ​​of the two pixels.

[0046] The third adjustment parameter is introduced into the calculation formula of the edge regularity to prevent the denominator from being 0. In this embodiment, the third adjustment parameter is preset to 0.1. In specific applications, the implementer can set it according to specific circumstances. The larger the edge feature value of the u-th pixel on the t-th quasi-lateral edge line and the smaller the average value of the difference between the grayscale values ​​of the u-th pixel and its adjacent pixels, the more uniform the grayscale value change is. The greater the edge regularity of the u-th pixel, the more likely it is to be an important edge of the vertebral body, that is, the greater the edge regularity.

[0047] By adopting the above method, the edge regularity of each pixel point on each quasi-transverse edge line in the cervical spine X-ray film of the bone injury patient can be obtained.

[0048] Step S3, screening feature pixels based on the edge regularity; determining the position importance of each pixel based on the difference in edge regularity between each feature pixel and its adjacent pixels, the edge regularity of each feature pixel, the relative position distribution of each feature pixel and the quasi-longitudinal edge line, and the image performance value of each pixel.

[0049] In X-ray images, the vertebral body usually appears in a rectangular shape, and the edges around it gradually sink inward. This phenomenon reflects the structural characteristics of the vertebral bone. Therefore, when observing the edge near the intervertebral space, if there is a feature of being sunken into the vertebral body, it means that the edge belongs to the edge of the current vertebra, not part of other vertebrae. This depression is consistent with the anatomical structure of the normal vertebra. If the edge near the intervertebral space shows a shape similar to being sunken into the vertebral body, it indicates that the edge characteristics of this area are consistent with the morphological characteristics of the current vertebra.

[0050] The pixel points whose edge regularity is greater than the preset regularity threshold are determined as feature pixel points. In this embodiment, the preset regularity threshold is 0.65, and in specific applications, the implementer can set it according to specific circumstances.

[0051] Next, any characteristic pixel point is taken as an example for description, and other characteristic pixel points can be processed by the method provided in this embodiment.

[0052] Specifically, any feature pixel is recorded as the feature pixel to be evaluated, and the line between the feature pixel to be evaluated and the nearest endpoint of the quasi-longitudinal edge line closest to it is recorded as the reference line segment of the feature pixel to be evaluated; the rotation angle of the quasi-transverse edge line where the feature pixel to be evaluated is rotated counterclockwise from the horizontal right direction as the starting direction, and if the rotation angle is less than 90 degrees, it means that the pixel may not be sunken into the inside of the vertebral body, and may be a pixel in the intervertebral space or other edges on the other side of the vertebral body, and the angle characteristic value of the feature pixel to be evaluated is -1; if the rotation angle is greater than or equal to 90 degrees, it means that the pixel is sunken toward the inside of the cone, and the greater the possibility of being the edge of the vertebral body, the higher the importance, and the angle characteristic value of the feature pixel to be evaluated is 1. According to the difference in the degree of edge regularity between the feature pixel to be evaluated and its adjacent pixels, the degree of edge regularity of the feature pixel to be evaluated and the angle characteristic value, the position importance factor of the feature pixel to be evaluated is obtained.

[0053] In this embodiment, a specific calculation formula of the position importance factor is given, and the position importance factor of the vth feature pixel point can be expressed as: in, Represents the position importance factor of the vth feature pixel, represents the angle eigenvalue of the vth feature pixel, Indicates the edge regularity of the vth feature pixel, Indicates the degree of edge regularity of the adjacent pixels of the vth feature pixel. represents the absolute value sign, and exp() represents the exponential function with a natural constant as the base.

[0054] Indicates the difference between the edge regularity of the vth feature pixel and its adjacent pixels. The smaller the value, the smaller the difference between the edge regularity of the two pixels, the more likely the two pixels are to belong to the same continuous edge, and the greater the accuracy of the position. The larger the value, the higher the accuracy of the position of the vth feature pixel when it is located inside the vertebral body. At the same time, the greater the edge regularity, the greater the possibility that the vth feature pixel is the edge of the vertebral body, that is, the greater the position importance factor of the vth feature pixel. It should be noted that: if there is more than one neighboring pixel of the vth feature pixel, the average value of the edge regularity of all neighboring pixels is used as the edge regularity of the neighboring pixels of the vth feature pixel.

[0055] By using the above method, the position importance factor of each feature pixel can be obtained.

[0056] For the normal edges of the vertebrae, these areas usually carry key information to help doctors accurately identify the morphology, boundaries and structural status of the vertebrae. The edge of the vertebrae is one of the most critical parts of the X-ray image. Any slight morphological changes, edge irregularities or bone abnormalities may be a hint of potential lesions. Therefore, the pixels of the normal edges of the vertebrae should be more strongly enhanced to make these edges clearer and more prominent in the image. By enhancing the contrast of the vertebral edges, doctors can more accurately observe the morphological changes of the vertebrae, especially when the vertebrae have degenerative changes, fractures or other abnormalities. This enhancement can help improve the accuracy of diagnosis.

[0057] For the pixels of the intervertebral space and other interfering edges, the degree of enhancement should be moderately reduced. Due to its different tissue structure from the vertebral body, the intervertebral space usually appears as a relatively flat or blurred area, and its edge changes are relatively irregular. Over-enhancing the contrast of these areas may cause more noise or artifacts in the image, which will interfere with the observation of the vertebral structure. Therefore, the degree of enhancement should be relatively low in the intervertebral space and other areas that may carry interfering information to avoid adding unnecessary details in the image and ensure that these areas do not affect the normal structural analysis of the vertebral body. In addition, other interfering edges, such as false edges formed due to shooting angles, patient positions, equipment inaccuracies, etc., also need to be reduced in enhancement. Over-enhancing these false or non-real edge information may lead to misdiagnosis and affect the correct understanding of the vertebral body and surrounding structures.

[0058] Based on the above characteristics, this embodiment sets the position importance factor of each pixel point except the characteristic pixel point in the cervical spine X-ray film of the bone injury patient to 0, and then determines the position importance of each pixel point as the sum of the position importance factor of each pixel point and its image performance value. The greater the position importance, the more important the position is, and the more it needs to be enhanced.

[0059] So far, the position importance of each pixel in the cervical spine X-ray film of the bone injury patient is obtained.

[0060] Step S4, dividing all pixel points using the importance of all positions to obtain the target area.

[0061] In step S3 of this embodiment, the position importance of each pixel in the cervical spine X-ray of the bone injury patient is obtained, and then all the pixels are clustered based on the position importance.

[0062] Specifically, based on the position importance of all pixels in the cervical spine X-rays of patients with bone injuries, the K-means clustering algorithm is used to cluster all pixels in the cervical spine X-rays of patients with bone injuries to obtain multiple clusters. The pixels in each cluster have similar and highly correlated features, representing important areas in the image. The K-means clustering algorithm is a prior art and will not be described in detail here.

[0063] After clustering, histogram equalization is performed on each area according to the different divided areas to enhance the contrast and ensure that the details of each area are clearer. Finally, the U-Net neural network is used to segment the image and extract the target area of ​​the cervical spine X-ray. The target area is the area that needs to be focused on, which assists doctors in analyzing the morphology of each vertebra and the condition of the intervertebral space, determining whether there are signs of disc herniation or degeneration, observing whether there are bone spurs or other abnormalities on the edge of the vertebra, and analyzing the morphological changes of the vertebra, including whether it is deformed or deformed. If the edge of the cone is deformed, it may indicate osteoporosis or other lesions.

[0064] This embodiment first analyzes the image performance characteristics of each pixel point based on the grayscale difference between each pixel point and the neighboring pixel points and the grayscale value of each pixel point in the cervical spine X-ray film, and obtains the image performance value. Then, based on the distribution of the cervical spine, the edge regularity of the pixel point is evaluated according to the position distribution of each pixel point on each quasi-transverse edge line and the grayscale difference between each pixel point on each quasi-transverse edge line and the surrounding pixel points, and then a plurality of characteristic pixel points are screened out. Then, based on the different change characteristics on the edge of each vertebral body of the cervical spine, the distribution change of the pixel points around the intervertebral space is obtained, so as to evaluate the importance of the positions of the pixels at different positions, and then divide all the pixels points into a plurality of different regions based on the position importance, so as to facilitate the extraction of the target region. The method provided in this embodiment realizes the effective segmentation of the vertebral body area in the image, improves the situation where the edges between the intervertebral spaces are difficult to distinguish or unclear, and improves the accuracy of the division results of different regions, providing support for subsequent medical diagnosis and helping doctors to identify cervical spine lesions.

[0065] Example of a cervical spine X-ray segmentation and identification system for obtaining an accurate treatment plan for bone injuries: See also Figure 2 , which shows a structural block diagram of a cervical spine X-ray segmentation and identification system for obtaining a precise treatment plan for bone injuries provided by an embodiment of the present invention. The system may include an image acquisition module, a first calculation module, a second calculation module and a target area determination module.

[0066] Among them, the image acquisition module is used to obtain cervical spine X-rays of patients with bone injuries; The first calculation module is used to obtain the image performance value of each pixel point according to the grayscale difference between each pixel point and the neighboring pixel points in the cervical spine X-ray film and the grayscale value of each pixel point; obtain the edge feature value of each pixel point on each quasi-transverse edge line according to the position distribution of each pixel point and the image performance value; obtain the edge regularity degree of the corresponding pixel point according to the grayscale difference between each pixel point on each quasi-transverse edge line and the surrounding pixel points and the edge feature value; The second calculation module is used to select characteristic pixel points based on the edge regularity; determine the position importance of each pixel point based on the difference in edge regularity between each characteristic pixel point and its adjacent pixel points, the edge regularity of each characteristic pixel point, the relative position distribution of each characteristic pixel point and the quasi-longitudinal edge line, and the image performance value of each pixel point; wherein the quasi-horizontal edge line and the quasi-longitudinal edge line are determined according to the inclination of the non-closed edge line; The target area determination module is used to divide all pixel points according to the importance of all positions to obtain the target area.

[0067] It should be understood that Figure 2 The structural block diagram of the cervical spine X-ray segmentation identification system for obtaining a precise treatment plan for bone injuries and its modules shown can be implemented in various ways. For example, in some embodiments, the system and its modules can be implemented by hardware, software, or a combination of software and hardware. Among them, the hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. Those skilled in the art can understand that the above-mentioned methods and systems can be implemented using computer executable instructions and / or included in processor control codes, such as carrier media such as disks, CDs or DVD-ROMs, programmable memories such as read-only memories (firmware), or data carriers such as optical or electronic signal carriers. Such codes are provided on. The system and its modules of this specification can not only be implemented by hardware circuits such as ultra-large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., but can also be implemented by software executed by various types of processors, and can also be implemented by a combination of the above-mentioned hardware circuits and software (for example, firmware).

[0068] For more details about the above modules, please refer to other places in this manual and will not be repeated here.

[0069] In other embodiments, a cervical vertebra X-ray film segmentation and identification device for obtaining a precise treatment plan for bone injuries is also provided, including a memory and a processor. The memory is used to store executable program code, and the processor is used to call and run the executable program code from the memory, so that the device executes the above-mentioned cervical vertebra X-ray film segmentation and identification method for obtaining a precise treatment plan for bone injuries. The device can specifically be a chip, a component or a module, and the chip may include a connected processor and a memory; wherein the memory is used to store instructions, and when the processor calls and executes the instructions, the chip can execute the cervical vertebra X-ray film segmentation and identification method for obtaining a precise treatment plan for bone injuries provided in the above-mentioned embodiment.

[0070] In other embodiments, a computer program product is also provided. When the computer program product is run on a computer, the computer executes the above-mentioned related steps to implement the cervical spine X-ray film segmentation and identification method for obtaining a precise treatment plan for bone injuries provided in the above-mentioned embodiment.

[0071] In other embodiments, a computer-readable storage medium is also provided, in which a computer program code is stored. When the computer program code is run on a computer, the computer executes the above-mentioned related method steps to implement the cervical spine X-ray film segmentation and identification method for obtaining a precise treatment plan for bone injuries provided in the above-mentioned embodiment.

[0072] Among them, the provided system, electronic device, computer program product, and computer-readable storage medium are all used to execute the corresponding methods provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding methods provided above, and will not be repeated here.

[0073] It should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A cervical spine X-ray film segmentation and marking method for obtaining an accurate treatment plan for bone injuries, characterized in that: The method comprises the following steps: Obtain cervical spine radiographs for patients with bone injuries; According to the grayscale difference between each pixel point and the neighboring pixel points in the cervical spine X-ray film and the grayscale value of each pixel point, the image performance value of each pixel point is obtained; according to the position distribution of each pixel point on each quasi-transverse edge line and the image performance value, the edge feature value of each pixel point on each quasi-transverse edge line is obtained; according to the grayscale difference between each pixel point on each quasi-transverse edge line and the surrounding pixel points and the edge feature value, the edge regularity of the corresponding pixel point is obtained; The characteristic pixel points are selected based on the edge regularity; the position importance of each pixel point is determined by combining the difference in edge regularity between each characteristic pixel point and its adjacent pixel points, the edge regularity of each characteristic pixel point, the relative position distribution of each characteristic pixel point and the quasi-longitudinal edge line, and the image performance value of each pixel point; wherein the quasi-horizontal edge line and the quasi-longitudinal edge line are determined according to the inclination of the non-closed edge line; All pixels are divided using the importance of all positions to obtain the target area.

2. The cervical spine X-ray film segmentation and marking method for obtaining a precise treatment plan for bone injuries according to claim 1, characterized in that: The image performance value of each pixel is obtained according to the grayscale difference between each pixel and the neighboring pixel in the cervical spine X-ray film and the grayscale value of each pixel, including: According to the grayscale difference between the candidate pixel and all the pixels in its neighborhood and the grayscale value of the candidate pixel, an image representation value of the candidate pixel is obtained, and the grayscale difference between the candidate pixel and all the pixels in its neighborhood and the grayscale value of the candidate pixel are both positively correlated with the image representation value; The candidate pixel point is any pixel point in the cervical spine X-ray film.

3. The cervical vertebra X-ray film segmentation and marking method for obtaining a precise treatment plan for bone injuries according to claim 1, characterized in that: The acquisition of quasi-transverse edge lines and quasi-longitudinal edge lines includes: The absolute value of the average slope of all pixel points on each non-closed edge line in the cervical spine X-ray film of patients with bone injuries was calculated and recorded as the inclination index of each non-closed edge line; A non-closed edge line with a tilt index less than a preset slope threshold is determined as a quasi-transverse edge line, and a non-closed edge line with a tilt index greater than or equal to the preset slope threshold is determined as a quasi-longitudinal edge line.

4. The cervical vertebra X-ray film segmentation and marking method for obtaining a precise treatment plan for bone injuries according to claim 1, characterized in that: The step of obtaining the edge feature value of each pixel point on each quasi-lateral edge line according to the position distribution of each pixel point on each quasi-lateral edge line and the image performance value includes: For each quasi-longitudinal edge line, draw a perpendicular line through the midpoint of each quasi-longitudinal edge line; Calculate the difference between the minimum value of the distance between the pixel to be analyzed and all vertical lines and half the length of the horizontal edge line where the pixel to be analyzed is located, and record it as the length difference; According to the image representation value of the pixel to be analyzed and the length difference, an edge feature value of the pixel to be analyzed is obtained, wherein the image representation value of the pixel to be analyzed is positively correlated with the edge feature value, and the length difference is negatively correlated with the edge feature value; The pixel point to be analyzed is any pixel point on any type of horizontal edge line.

5. The cervical vertebra X-ray film segmentation and marking method for obtaining a precise treatment plan for bone injuries according to claim 4, characterized in that: The step of obtaining the edge regularity of the corresponding pixel point according to the grayscale difference between each pixel point and its surrounding pixels on each quasi-lateral edge line and the edge feature value includes: Calculate the average value of the grayscale value difference between the pixel to be analyzed and its adjacent pixels on the straight line where the pixel to be analyzed is located, and record it as the average grayscale difference of the pixel to be analyzed; The edge regularity of the pixel to be analyzed is obtained according to the edge feature value of the pixel to be analyzed and the grayscale average difference, wherein the edge feature value is positively correlated with the edge regularity, and the grayscale average difference is negatively correlated with the edge regularity.

6. The cervical vertebra X-ray film segmentation and marking method for obtaining a precise treatment plan for bone injuries according to claim 1, characterized in that: The screening of characteristic pixel points based on the edge regularity degree includes: determining pixel points whose edge regularity degree is greater than a preset regularity threshold as characteristic pixel points.

7. The cervical spine X-ray film segmentation and marking method for obtaining a precise treatment plan for bone injuries according to claim 1, characterized in that: The determining of the position importance of each pixel point by combining the difference in edge regularity between each feature pixel point and its adjacent pixels, the edge regularity of each feature pixel point, the relative position distribution of each feature pixel point and the quasi-vertical edge line, and the image performance value of each pixel point includes: The line connecting the feature pixel to be evaluated and the nearest endpoint of the quasi-vertical edge line closest to it is recorded as the reference line segment of the feature pixel to be evaluated; obtaining the rotation angle of the quasi-horizontal edge line where the feature pixel to be evaluated is located, with the horizontal right direction as the starting direction and rotated counterclockwise to the quasi-horizontal edge line where the feature pixel to be evaluated is located; if the rotation angle is less than 90 degrees, the angle characteristic value of the feature pixel to be evaluated is set to -1; if the rotation angle is greater than or equal to 90 degrees, the angle characteristic value of the feature pixel to be evaluated is set to 1; Obtaining a position importance factor of the feature pixel to be evaluated according to the difference in edge regularity between the feature pixel to be evaluated and its adjacent pixel points, the edge regularity of the feature pixel to be evaluated and the angle characteristic value; the feature pixel to be evaluated is any feature pixel point; Set the position importance factor of pixels other than feature pixels to 0; The position importance factor of each pixel and the image performance value are combined to determine the position importance of each pixel.

8. The cervical vertebra X-ray film segmentation and marking method for obtaining a precise treatment plan for bone injuries according to claim 7, characterized in that: The determining the position importance of each pixel point by combining the position importance factor of each pixel point and the image performance value includes: The position importance factor of each pixel and its image performance value are summed to determine the position importance of each pixel.

9. The cervical vertebra X-ray film segmentation and marking method for obtaining a precise treatment plan for bone injuries according to claim 1, characterized in that: The method of dividing all pixels using the importance of all positions to obtain the target area includes: All pixels in the cervical spine X-ray films of patients with bone injuries are clustered based on their positional importance, and the target area is extracted based on the clustering results.

10. The cervical spine X-ray film segmentation and marking method for obtaining a precise treatment plan for bone injuries according to claim 9, characterized in that: The K-means clustering algorithm was used to cluster all the pixels in the cervical spine X-rays of patients with bone injuries.

Images