Method and system for detecting bone damage in medical images

Abstract

The present invention provides a bone medical image damage detection method and system, and relates to the field of medical image processing technology. The method comprises: obtaining a bone binary image; obtaining the bone contour line in the bone binary image; determining the bone fracture judgment score; obtaining the midline of two adjacent bone contour lines; determining the bone crack judgment score; in the case where at least one of the bone crack judgment score and the bone fracture judgment score is greater than or equal to a preset first damage score threshold, setting a detection line parallel to the contour line in the background area outside the bone area; obtaining the bone fragment judgment score; determining the bone damage judgment score; if the bone damage judgment score is greater than or equal to a preset second damage score threshold, marking the bone medical image to be processed. According to the present invention, image processing technology is used to assist doctors in diagnosis, thereby improving the detection accuracy and efficiency of bone damage.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and in particular to a method and system for detecting bone medical image damage. Background Art

[0002] In the related art, doctors often identify various types of injuries such as subtle injuries and complex injuries in the diagnosis process for injuries in bone medical images, which may lead to low work efficiency when processing a large number of images at the same time. In addition, the detection of injuries in bone medical images is overly dependent on manual work, which may lead to misidentification and other phenomena.

[0003] The information disclosed in the background technology section of this application is only intended to deepen the understanding of the general background technology of this application, and should not be regarded as an admission or any form of suggestion that the information constitutes the prior art already known to those skilled in the art. Summary of the invention

[0004] The embodiments of the present invention provide a method and system for detecting bone medical image damage, which can solve the technical problems of low detection efficiency and easy misidentification of bone damage.

[0005] According to a first aspect of an embodiment of the present invention, a bone medical image damage detection method is provided, comprising: performing binarization processing on a bone medical image to be processed to obtain a bone binary image; detecting the bone binary image through a contour detection model to obtain a bone contour line in the bone binary image; determining a bone fracture judgment score according to the position coordinates of a plurality of pixel points on the contour line; obtaining a midline of two adjacent bone contour lines according to the bone contour line; determining a bone crack judgment score according to the midline and the pixel values ​​of the pixel points in the bone area within the contour line in the bone binary image; in the case where at least one of the bone crack judgment score and the bone fracture judgment score is greater than or equal to a preset first damage score threshold, setting a detection line parallel to the contour line in a background area outside the bone area according to the bone contour line; obtaining a bone fragment judgment score according to the pixel values ​​of the pixel points in the background area and the detection line; determining a bone fracture judgment score according to the bone fracture judgment score, the bone crack judgment score and the bone fragment judgment score; if the bone injury judgment score is greater than or equal to a preset second damage score threshold, marking the bone medical image to be processed.

[0006] According to a second aspect of an embodiment of the present invention, a bone medical image damage detection system is provided, the system comprising: a bone binary image module, used to perform binary processing on the bone medical image to be processed to obtain a bone binary image; a contour line module, used to detect the bone binary image through a contour detection model to obtain the contour line of the bone in the bone binary image; a bone fracture judgment score module, used to determine the bone fracture judgment score according to the position coordinates of a plurality of pixel points on the contour line; a midline module, used to obtain the midline of two adjacent bone contour lines according to the bone contour line; a bone crack judgment score module, used to determine the bone crack judgment score according to the midline and the pixel values ​​of the pixel points of the bone area within the contour line in the bone binary image. a detection line module, which is used to set a detection line parallel to the contour line in the background area outside the bone area according to the contour line of the bone when at least one of the bone crack judgment score and the bone fracture judgment score is greater than or equal to a preset first injury score threshold; a bone fragment judgment score module, which is used to obtain the bone fragment judgment score according to the pixel values ​​of the pixel points in the background area and the detection line; a bone injury judgment score module, which is used to determine the bone injury judgment score according to the bone fracture judgment score, the bone crack judgment score and the bone fragment judgment score; a marking module, which is used to mark the bone medical image to be processed if the bone injury judgment score is greater than or equal to a preset second injury score threshold.

[0007] Technical effect: According to the present invention, by binarizing and detecting the bone medical image, combined with the analysis of pixel position coordinates and pixel values, the detection of injuries such as bone fractures, bone cracks and bone fragments is realized, and the comprehensiveness of the detection of multiple types of injuries is improved. By analyzing the bone fracture judgment score, the bone crack judgment score and the bone fragment judgment score, the degree of bone injury is comprehensively evaluated. The final bone injury judgment score can provide an important reference for doctors, assist doctors in detection, improve detection efficiency, and reduce the probability of misidentification. By marking the image, it is convenient for doctors to understand the injury more intuitively, and provide an accurate data basis for doctors to formulate accurate treatment plans. When determining the bone fracture judgment score, the angle between two adjacent discriminant vectors on the bone contour line can be used to judge by using a preset angle threshold to realize the detection of the bone fracture part that may exist in the bone medical image. It can assist doctors in quickly and accurately identifying potential bone fractures, improve diagnostic efficiency, and improve the accuracy and objectivity of bone fracture recognition. When determining the bone crack judgment score, the pixel values ​​on multiple discriminant curves can be analyzed, combined with the gradient size of the pixel values ​​along the direction vector and the preset first change rate threshold, to evaluate the possible crack conditions in the bone area, and finally combine the crack conditions judged by multiple discriminant curves to obtain the bone crack judgment score, which can assist doctors in bone crack detection and diagnosis, and improve the accuracy and objectivity of crack identification. When determining the fragment discrimination score corresponding to the p-th translation, the pixel values ​​of the pixel points on the detection line after the p-th translation can be analyzed, combined with the gradient size of the pixel values ​​along the direction vector and the preset second change rate threshold, the fragment discrimination score corresponding to the p-th translation can be calculated, and the possible bone fragments in the bone medical image can be quantitatively evaluated, providing doctors with more objective information to assist diagnosis and treatment, and improving the accuracy of bone fragment identification. When determining the bone fragment judgment score, the meaningful translation positions can be screened out, the fragment judgment scores of each translation position can be weighted and summed, and the distance between the bone fragment and the bone area can be represented by the number of translations, so that the weight can be set based on the distance, so that the weight can objectively express the severity of bone damage. The feature changes of the bone image under different translations can be comprehensively considered, which improves the accuracy and objectivity of the recognition of bone fragments.

[0008] It should be understood that the above general description and the following detailed description are exemplary and explanatory only and do not limit the present invention. Other features and aspects of the present invention will become more apparent from the following detailed description of exemplary embodiments with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other embodiments can be obtained based on these drawings without creative work.

[0010] Figure 1 A schematic diagram of a flow chart of a method for detecting bone medical image damage according to an embodiment of the present invention is exemplarily shown;

[0011] Figure 2 A block diagram of a skeletal medical image damage detection system according to an embodiment of the present invention is exemplarily shown. DETAILED DESCRIPTION

[0012] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0013] The technical solution of the present invention is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0014] Figure 1A flow chart of a bone medical image damage detection method according to an embodiment of the present invention is exemplarily shown, the method comprising: step S101, binarizing the bone medical image to be processed to obtain a bone binary image; step S102, detecting the bone binary image through a contour detection model to obtain the bone contour line in the bone binary image; step S103, determining a bone fracture judgment score according to the position coordinates of a plurality of pixel points on the contour line; step S104, obtaining the midline of two adjacent bone contour lines according to the bone contour line; step S105, determining a bone crack judgment score according to the midline and the pixel values ​​of the pixel points of the bone area within the contour line in the bone binary image. Score; step S106, when at least one of the bone crack judgment score and the bone fracture judgment score is greater than or equal to a preset first injury score threshold, a detection line parallel to the contour line is set in the background area outside the bone area according to the contour line of the bone; step S107, obtaining a bone fragment judgment score according to the pixel values ​​of the pixel points in the background area and the detection line; step S108, determining a bone injury judgment score according to the bone fracture judgment score, the bone crack judgment score and the bone fragment judgment score; step S109, if the bone injury judgment score is greater than or equal to a preset second injury score threshold, marking the bone medical image to be processed.

[0015] According to the bone medical image injury detection method of the embodiment of the present invention, by binarizing and contour detecting the bone medical image, combined with the analysis of the pixel position coordinates and pixel values, the detection of injuries such as bone fractures, bone cracks and bone fragments is achieved, and the comprehensiveness of the detection of various types of injuries is improved. By analyzing the bone fracture judgment score, the bone crack judgment score and the bone fragment judgment score, the degree of bone injury is comprehensively evaluated. The final bone injury judgment score can provide an important reference for doctors, assist doctors in detection, improve detection efficiency, and reduce the probability of misidentification. By marking the image, it is convenient for doctors to understand the injury more intuitively, and provide an accurate data basis for doctors to formulate accurate treatment plans.

[0016] According to one embodiment of the present invention, in step S101, the binarization process is to convert the pixel values ​​in the image into black and white colors (for example, the gray value is set to 0 or 1), that is, to convert the image into an image containing only black and white pixels. For example, when the gray value of a pixel is greater than or equal to a preset threshold, the pixel is set to a black pixel, otherwise, the pixel is set to a white pixel. In skeletal medical images, this process can simplify the image, reduce noise interference in skeletal medical images, highlight the contours and features of the bone structure, and provide a clear image basis for subsequent contour detection and damage analysis.

[0017] According to one embodiment of the present invention, in step S102, the contour detection model is a model used in the field of image processing and computer vision, such as an active contour model, a geometric active contour model, etc., which is intended to identify the edge contours of objects in an image. In bone medical image processing, the contour detection model is used to identify and extract the contours of bone structures, providing a basis for subsequent analysis and diagnosis.

[0018] According to one embodiment of the present invention, in step S103, by analyzing the positions of the pixels on the contour line, it is possible to detect whether the bone is broken. The degree or possibility of bone fracture can be evaluated based on the position coordinates of multiple pixels to provide auxiliary diagnosis information for doctors.

[0019] According to an embodiment of the present invention, step S103 includes: evenly setting a plurality of sampling points on each contour line; determining a discriminant vector between adjacent sampling points; and determining a bone fracture judgment score according to the plurality of discriminant vectors.

[0020] According to one embodiment of the present invention, multiple sampling points are evenly set on the contour line of each bone, and the vector between adjacent sampling points on the same contour line is a discriminant vector. The angle change between two adjacent discriminant vectors is observed. At the fracture site, the angle may change sharply or exceed the normal range. Based on the characteristics of multiple discriminant vectors, the possibility of bone fracture can be calculated and a corresponding judgment score can be given.

[0021] According to one embodiment of the present invention, a bone fracture judgment score is determined based on a plurality of discriminant vectors, including: determining a bone fracture judgment score F according to formula (1),

[0022]

[0023] Among them, V i，(j)-(j+1) is the discriminant vector between the jth sampling point and the j+1th sampling point on the i-th contour line, V i，(j+1)-(j+2) is the discriminant vector between the j+1th sampling point and the j+2th sampling point on the i-th contour line, n i is the number of sampling points on the i-th contour line, n is the number of contour lines, θ T is the preset angle threshold, arccos is the arccosine function, max is the maximum value function, if is the conditional function, 1≤j≤n i -2, i≤n, and i, j, n and n i All are positive integers.

[0024] According to one embodiment of the present invention, in formula (1), is the cosine similarity between the discriminant vector between the jth sampling point and the j+1th sampling point on the i-th contour line and the discriminant vector between the j+1th sampling point and the j+2th sampling point on the i-th contour line. is the arccosine value of the cosine similarity, indicating the angle between the two adjacent discriminant vectors. When the maximum value of the angle between two adjacent discriminant vectors on the same contour line is greater than or equal to the preset angle threshold, the conditional function value is 1, otherwise, the conditional function value is 0. The conditional function value is 1, indicating that there is a break on the contour line, which can indicate a bone break. The preset angle threshold can be set to 30°, 35°, etc., and the present invention is not limited to this. Indicates the average value of the angles between adjacent discriminant vectors on all contour lines that exceed the preset angle threshold, that is, the bone fracture judgment score. In the example, if the score is greater than or equal to 0.5, at least half of the contour lines are broken, and the bone fracture can be judged.

[0025] In this way, the angle between two adjacent discriminant vectors on the bone contour line can be used to determine the fractured bones in the bone medical image. This can help doctors quickly and accurately identify potential bone fractures, improve diagnostic efficiency, and improve the accuracy and objectivity of bone fracture identification.

[0026] According to an embodiment of the present invention, in step S104, by obtaining the midline of two adjacent contour lines of the bone, the morphology, structure and possible changes of the bone can be further analyzed. The extraction of the midline can be used for more detailed morphological analysis, bone feature extraction or further damage assessment.

[0027] According to one embodiment of the present invention, step S104 includes: when the area between two adjacent contour lines is a skeleton area, determining the minimum circumscribed rectangle of the skeleton area; determining the direction vector in the width direction of the minimum circumscribed rectangle; at any pixel point on the first contour line of the two adjacent contour lines, setting a straight line along the direction of the direction vector to obtain the intersection of the straight line and the second contour line; obtaining the midpoint between the arbitrary pixel point and the intersection; and obtaining the midline of the two adjacent contour lines based on the midpoints corresponding to multiple pixel points on the first contour line.

[0028] According to one embodiment of the present invention, the portion between two adjacent contour lines representing the bone region is obtained, and the minimum circumscribed rectangle of this region is calculated, that is, the rectangle that can completely surround the bone region and has the smallest area. The width direction of the minimum circumscribed rectangle is the direction of the short side of the rectangle, and the bone region is extended in the length direction of the minimum circumscribed rectangle. The midpoint can represent the middle position of the two contour lines along the width direction. The midpoints corresponding to multiple pixel points on the two contour lines are used to connect and obtain the midline between the two adjacent contour lines. The midline can represent the center line of the bone region between the two adjacent contour lines, which is parallel to the bone extension direction, and is helpful to further analyze the morphological characteristics of the bone structure.

[0029] According to one embodiment of the present invention, in step S105, by analyzing the pixel values ​​inside the bone region, especially the pixel value changes of the pixels in the direction parallel to the midline, it can be determined whether there is a crack in the bone. For example, if the midline pixel value changes twice, it means that the midline passes through a bone crack once, that is, there is a bone crack. Abnormal conditions of the bone structure can be discovered in time, providing an important reference for diagnosis and treatment.

[0030] According to one embodiment of the present invention, step S105 includes: setting a plurality of discriminant curves parallel to the midline in the bone area between the two adjacent contour lines, wherein the midline serves as the first discriminant curve; and determining the bone crack judgment score according to the pixel values ​​of the pixel points on the plurality of discriminant curves.

[0031] According to one embodiment of the present invention, the midline itself is used as the first discriminant curve, and other discriminant curves may be located on both sides of the midline to more comprehensively cover the bone area. For each pixel point on the discriminant curve, the change of its pixel value is observed, and the change of these pixel values ​​may reflect the condition of the bone crack.

[0032] According to one embodiment of the present invention, determining a bone crack judgment score according to the pixel values ​​of the pixel points on a plurality of discriminant curves includes: determining a bone crack judgment score C according to formula (2),

[0033]

[0034] Among them, (x k，t ,y k，t ) is the coordinate of the t-th pixel point on the k-th discriminant curve, p(x k，t ,y k，t ) is the pixel value of the t-th pixel point on the k-th discriminant curve, is the direction vector at the position of the t-th pixel point on the k-th discriminant curve, is the pixel value of the t-th pixel point on the k-th discriminant curve along the direction vector The gradient of RT，1 is the preset first change rate threshold, n k is the number of pixels on the kth discriminant curve, m is the number of discriminant curves, t≤n k , k≤m, and t, k, n k and m are both positive integers.

[0035] According to one embodiment of the present invention, in formula (2), Represents the pixel value of the t-th pixel point on the k-th discriminant curve along the direction vector The gradient size. It can be used to measure the gradient along the direction vector The rate of change of pixel values ​​in the direction of the discriminant curve is the size of the gradient of the pixel values ​​on the discriminant curve. If the discriminant curve passes through the bone crack, that is, the discriminant curve passes from the bone area to the crack area, and then from the crack area to the bone area, the number of times the gradient of the pixel values ​​on the discriminant curve changes significantly is two. If the discriminant curve does not pass through the bone crack, the gradient of the pixel values ​​on the discriminant curve remains unchanged or changes slightly. Represents the pixel value of the t-th pixel point on the k-th discriminant curve along the direction vector When the gradient size is greater than or equal to the preset first change rate threshold, the conditional function value is 1, otherwise, the conditional function value is 0. If the gradient value is large enough, it indicates that there is a significant change in the pixel value at the position of the pixel point on the kth discriminant curve, indicating that there may be a bone crack. It represents the sum of the conditional judgments of all the pixels on the k-th discriminant curve, and the result is divided by 2, that is, the number of bone cracks that the k-th discriminant curve passes through.

[0036] It represents the average number of bone cracks passed by all the discriminant curves, that is, the bone crack judgment score.

[0037] In this way, the possible crack conditions in the bone area can be evaluated by analyzing the pixel values ​​on multiple discriminant curves, combining the gradient size of the pixel values ​​along the direction vector and the preset first change rate threshold, and finally combining the existence of cracks determined by multiple discriminant curves to obtain a bone crack judgment score, which can assist doctors in bone crack detection and diagnosis and improve the accuracy and objectivity of crack identification.

[0038] According to one embodiment of the present invention, in step S106, if at least one of the crack judgment score and the fracture judgment score is greater than or equal to a preset first injury score threshold, it indicates that the bone may be seriously damaged. At this time, the situation around the bone can be further analyzed. For example, a detection line can be set to detect bone fragments, the impact range, and the degree of damage. If neither the crack judgment score nor the fracture judgment score is greater than or equal to the preset first injury score threshold, the bone fragment judgment score does not need to be obtained.

[0039] According to one embodiment of the present invention, in step S107, the pixel values ​​of the pixels on the detection line can be used to determine whether there are bone fragments in the background area. By analyzing these pixel values ​​and the position of the detection line, it is possible to evaluate whether the bone is broken, as well as the distance and number of the bone fragments from the bone contour line. For example, the farther the detection line is from the bone contour line, the more bone fragments are detected, and the more serious the bone damage is.

[0040] According to one embodiment of the present invention, step S107 includes: setting the detection line at an initial position in the background area, wherein the distance between the detection line at the initial position and the contour line is a preset distance, and the preset distance is the distance along the width direction of the minimum circumscribed rectangle; determining the fragment discrimination score corresponding to the initial position according to the pixel values ​​of the pixel points on the detection line at the initial position; during the p-th translation, translating the detection line along the width direction so that the detection line is away from the contour line, wherein p is a positive integer greater than or equal to 1, and the distance of each translation is the preset distance; after the p-th translation, determining the fragment discrimination score corresponding to the p-th translation according to the pixel values ​​of the pixel points on the detection line after the p-th translation; determining the bone fragment discrimination score according to the fragment discrimination score corresponding to the initial position and the fragment discrimination scores corresponding to each translation.

[0041] According to one embodiment of the present invention, a detection line at an initial position is determined in the background area, and the distance between the line and the contour line is a preset distance. The preset distance is the distance measured in the width direction of the minimum circumscribed rectangle of the bone area. By analyzing the pixel values ​​of the pixel points at the initial position on the detection line, the fragment discrimination score corresponding to the initial position is determined. The score helps to evaluate the possible bone fragments at the initial position. Each time the contour line is moved away from the contour line, the possible bone fragments in the background area can be gradually detected.

[0042] According to an embodiment of the present invention, after the p-th translation, according to the pixel values ​​of the pixels on the detection line after the p-th translation, determining the fragment discrimination score corresponding to the p-th translation includes: obtaining the fragment discrimination score D corresponding to the p-th translation according to formula (3) p ,

[0043]

[0044] Among them, (x p，q ,y p，q ) is the coordinate of the qth pixel on the detection line after the pth translation, p(x p，q ,y p，q ) is the pixel value of the qth pixel point on the detection line after the pth translation, is the direction vector of the qth pixel on the detection line after the pth translation, is the pixel value of the qth pixel point on the detection line after the pth translation along the direction vector The gradient of R T，2 is the preset second change rate threshold, n p is the number of pixels on the detection line, and p, q and n p All are positive integers.

[0045] According to one embodiment of the present invention, in formula (3), Represents the pixel value of the qth pixel point on the detection line after the pth translation along the direction vector The gradient size. It can be used to measure the gradient along the direction vector The value of the rate of change of pixel values ​​in the direction of the detection line, that is, the size of the gradient of the pixel values ​​on the detection line. If the detection line passes through the bone fragment, the size of the gradient of the pixel values ​​on the detection line changes significantly twice, that is, it changes significantly once when it passes from the background area to the area where the bone fragment is located, and changes significantly once when it passes from the area where the bone fragment is located to the background area. If the detection line does not pass through the bone fragment, the size of the gradient of the pixel values ​​on the detection line remains unchanged, or the change is not significant.

[0046] Represents the pixel value of the qth pixel point on the detection line after the pth translation along the direction vector When the gradient size is greater than or equal to the preset second change rate threshold, the conditional function value is 1, otherwise, the conditional function value is 0. If the gradient value is large enough, it indicates that the gradient of the detection line at the position of the pixel point has a significant change, indicating that there may be bone fragments. It means that the conditional judgment results of all pixels on the detection line after the p-th translation are summed and divided by 2, that is, the number of bone fragments passed by the detection line after the p-th translation, which represents the fragment discrimination score corresponding to the p-th translation.

[0047] In this way, the pixel values ​​of the pixel points on the detection line after the p-th translation can be analyzed, and the fragment discrimination score corresponding to the p-th translation can be calculated by combining the gradient size of the pixel value along the direction vector and the preset second change rate threshold, so as to quantitatively evaluate the possible bone fragments in the skeletal medical image, provide doctors with more objective information to assist diagnosis and treatment, and improve the accuracy of bone fragment recognition.

[0048] According to one embodiment of the present invention, the bone fragment judgment score is determined according to the fragment judgment score corresponding to the initial position and the fragment judgment score corresponding to each translation, including: determining the bone fragment judgment score D according to formula (4),

[0049]

[0050] Among them, D p is the fragment discrimination score corresponding to the pth translation, N is the total number of translations, 0≤p≤N, and when p=0, D0 is the fragment discrimination score corresponding to the initial position, p and N are both positive integers, and if is a conditional function.

[0051] According to one embodiment of the present invention, in formula (4), Represents the weight of the fragment discrimination score corresponding to each translation, and the weight is adjusted according to the number of translations. Among them, N+1 is the total number of positions, and p+1 is the position to which the translation is made. The distance of each translation is equal. The distance between the p+1th position and the contour line is p+1 times the preset distance, and the distance between the total position and the contour line is N+1 times the preset distance. The farther the translation position is from the contour line, the more serious the bone damage, and the greater the weight. represents the weighted summation of the fragment discrimination scores corresponding to each translation. p ≠0, 1, 0) means that when the fragment discrimination score corresponding to the p-th translation is not equal to 0, the conditional function value is 1, otherwise, the conditional function value is 0. It means summing the conditional function results corresponding to each translation, that is, counting the number of detection lines passing through the bone fragments. It represents the weighted average value of the fragment discrimination scores corresponding to each detection line passing through the bone fragment. Indicates that when the sum of the fragment discrimination scores corresponding to all translations is not equal to 0, the conditional function value is Otherwise, the conditional function value is 0. That is, when the sum of the fragment discrimination scores corresponding to all translations is not 0, the final bone fragment discrimination score is calculated, otherwise, the bone fragment discrimination score is directly set to 0.

[0052] In this way, by screening out meaningful translation positions, the fragment discrimination scores of each translation position are weighted and summed, and the number of translations is used to represent the distance between the bone fragment and the bone area, so that the weight is set based on the distance, so that the weight can objectively express the severity of bone damage. The feature changes of the bone image under different translations can be comprehensively considered, which improves the accuracy and objectivity of the recognition of bone fragments.

[0053] According to one embodiment of the present invention, in step S108, according to the severity, the bone fracture judgment score has a medium weight, the bone crack judgment score has the lowest weight, and the bone fragment judgment score has the highest weight. These three weighted scores are then summed to obtain a bone injury judgment score to assist doctors in assessing the degree of bone injury and thus formulating a corresponding treatment plan.

[0054] According to an embodiment of the present invention, in step S109, the marking process may include marking the location, type, severity and other information of the damaged part in the image so that the doctor can more clearly understand the bone injury. The marking process can assist the medical team to diagnose and track bone injuries more effectively and provide more reference and support for formulating corresponding treatment plans.

[0055] According to the bone medical image injury detection method of the embodiment of the present invention, by binarizing and contour detection of the bone medical image, combined with the analysis of the pixel position coordinates and pixel values, the detection of injuries such as bone fracture, bone crack and bone fragment is realized, and the comprehensiveness of the detection of multiple types of injuries is improved. By analyzing the bone fracture judgment score, the bone crack judgment score and the bone fragment judgment score, the degree of bone injury is comprehensively evaluated. The final bone injury judgment score can provide an important reference for doctors, assist doctors in detection, improve detection efficiency, and reduce the probability of misidentification. By marking the image, it is convenient for doctors to understand the injury more intuitively, and provide an accurate data basis for doctors to formulate accurate treatment plans. When determining the bone fracture judgment score, the angle between two adjacent discriminant vectors on the bone contour line can be judged by using a preset angle threshold to realize the detection of the bone fracture part that may exist in the bone medical image. It can assist doctors in quickly and accurately identifying potential bone fractures, improve diagnostic efficiency, and improve the accuracy and objectivity of bone fracture recognition. When determining the bone crack judgment score, the pixel values ​​on multiple discriminant curves can be analyzed, combined with the gradient size of the pixel values ​​along the direction vector and the preset first change rate threshold, to evaluate the possible crack conditions in the bone area, and finally combine the crack conditions judged by multiple discriminant curves to obtain the bone crack judgment score, which can assist doctors in bone crack detection and diagnosis, and improve the accuracy and objectivity of crack identification. When determining the fragment discrimination score corresponding to the p-th translation, the pixel values ​​of the pixel points on the detection line after the p-th translation can be analyzed, combined with the gradient size of the pixel values ​​along the direction vector and the preset second change rate threshold, the fragment discrimination score corresponding to the p-th translation can be calculated, and the possible bone fragments in the bone medical image can be quantitatively evaluated, providing doctors with more objective information to assist diagnosis and treatment, and improving the accuracy of bone fragment identification. When determining the bone fragment judgment score, the meaningful translation positions can be screened out, the fragment judgment scores of each translation position can be weighted and summed, and the distance between the bone fragment and the bone area can be represented by the number of translations, so that the weight can be set based on the distance, so that the weight can objectively express the severity of bone damage. The feature changes of the bone image under different translations can be comprehensively considered, which improves the accuracy and objectivity of the recognition of bone fragments.

[0056] Figure 2A block diagram of a bone medical image damage detection system according to an embodiment of the present invention is exemplarily shown, wherein the system comprises: a bone binary image module, which is used to perform binary processing on the bone medical image to be processed to obtain a bone binary image; a contour line module, which is used to detect the bone binary image through a contour detection model to obtain the contour line of the bone in the bone binary image; a bone fracture judgment score module, which is used to determine the bone fracture judgment score according to the position coordinates of multiple pixel points on the contour line; a midline module, which is used to obtain the midline of two adjacent contour lines of the bone according to the contour line of the bone; and a bone crack judgment score module, which is used to determine the bone crack judgment score according to the midline and the pixel values ​​of the pixel points of the bone area within the contour line in the bone binary image. score; a detection line module, used to set a detection line parallel to the contour line in the background area outside the bone area according to the contour line of the bone when at least one of the bone crack judgment score and the bone fracture judgment score is greater than or equal to a preset first injury score threshold; a bone fragment judgment score module, used to obtain the bone fragment judgment score according to the pixel value of the pixel point in the background area and the detection line; a bone injury judgment score module, used to determine the bone injury judgment score according to the bone fracture judgment score, the bone crack judgment score and the bone fragment judgment score; a marking module, used to mark the bone medical image to be processed if the bone injury judgment score is greater than or equal to a preset second injury score threshold.

[0057] It should be understood by those skilled in the art that the embodiments of the present invention described above and shown in the accompanying drawings are only examples and do not limit the present invention. The purpose of the present invention has been fully and effectively achieved. The functional and structural principles of the present invention have been demonstrated and explained in the embodiments, and the embodiments of the present invention may be deformed or modified in any way without departing from the principles.

[0058] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A bone medical image damage detection method, characterized in that: include: Binarization is performed on the bone medical image to be processed to obtain a bone binary image; The bone binary image is detected by a contour detection model to obtain the contour line of the bone in the bone binary image; the bone fracture judgment score is determined according to the position coordinates of multiple pixel points on the contour line; the midline of two adjacent bone contour lines is obtained according to the bone contour line; the bone crack judgment score is determined according to the midline and the pixel values ​​of the pixel points of the bone area within the contour line in the bone binary image; When at least one of the bone crack judgment score and the bone fracture judgment score is greater than or equal to a preset first injury score threshold, a detection line parallel to the contour line is set in the background area outside the bone area according to the contour line of the bone; a bone fragment judgment score is obtained according to the pixel values ​​of the pixel points in the background area and the detection line; a bone injury judgment score is determined according to the bone fracture judgment score, the bone crack judgment score and the bone fragment judgment score; if the bone injury judgment score is greater than or equal to a preset second injury score threshold, the bone medical image to be processed is marked.

2. The bone medical image damage detection method according to claim 1, characterized in that: Determining the bone fracture judgment score according to the position coordinates of multiple pixel points on the contour line includes: evenly setting multiple sampling points on each contour line; determining the discriminant vector between adjacent sampling points; and determining the bone fracture judgment score according to the multiple discriminant vectors.

3. The bone medical image damage detection method according to claim 2, characterized in that: According to multiple discriminant vectors, the bone fracture judgment score is determined, including: according to the formula Determine the bone fracture judgment score F, where V i，(j)-(j+1) is the discriminant vector between the jth sampling point and the j+1th sampling point on the i-th contour line, V i，(j+1)-(j+2) is the discriminant vector between the j+1th sampling point and the j+2th sampling point on the i-th contour line, n i is the number of sampling points on the i-th contour line, n is the number of contour lines, θ T is the preset angle threshold, arccos is the arccosine function, max is the maximum value function, if is the conditional function, 1≤j≤n i -2, i≤n, and i, j, n and n i All are positive integers.

4. The bone medical image damage detection method according to claim 1, characterized in that: According to the contour line of the skeleton, the midline of two adjacent contour lines of the skeleton is obtained, including: when the area between the two adjacent contour lines is a skeleton area, determining the minimum circumscribed rectangle of the skeleton area; determining the direction vector in the width direction of the minimum circumscribed rectangle; at any pixel point on the first contour line of the two adjacent contour lines, setting a straight line along the direction of the direction vector to obtain the intersection of the straight line and the second contour line; obtaining the midpoint between the arbitrary pixel point and the intersection; and obtaining the midline of the two adjacent contour lines according to the midpoints corresponding to multiple pixel points on the first contour line.

5. The bone medical image damage detection method according to claim 1, characterized in that: The bone crack judgment score is determined according to the center line and the pixel values ​​of the pixel points in the bone area within the contour line in the bone binary image, including: in the bone area between the two adjacent contour lines, a plurality of discriminant curves parallel to the center line are set, wherein the center line is used as the first discriminant curve; and the bone crack judgment score is determined according to the pixel values ​​of the pixel points on the plurality of discriminant curves.

6. The bone medical image damage detection method according to claim 5, characterized in that: According to the pixel values ​​of the pixel points on the multiple discriminant curves, the bone crack judgment score is determined, including: according to the formula Determine the bone crack judgment score C, where (x k，t ,y k，t ) is the coordinate of the t-th pixel point on the k-th discriminant curve, p(x k，t ,y k，t ) is the pixel value of the t-th pixel point on the k-th discriminant curve, is the direction vector at the position of the t-th pixel point on the k-th discriminant curve, is the pixel value of the t-th pixel point on the k-th discriminant curve along the direction vector The gradient of R T，1 is the preset first change rate threshold, n k is the number of pixels on the kth discriminant curve, m is the number of discriminant curves, t≤n k , k≤m, and t, k, n k and m are both positive integers.

7. The bone medical image damage detection method according to claim 4, characterized in that: According to the pixel values ​​of the pixel points in the background area and the detection line, a bone fragment judgment score is obtained, including: setting the detection line at an initial position in the background area, wherein the distance between the detection line at the initial position and the contour line is a preset distance, and the preset distance is the distance along the width direction of the minimum circumscribed rectangle; determining the fragment judgment score corresponding to the initial position according to the pixel values ​​of the pixel points on the detection line at the initial position; during the p-th translation, translating the detection line along the width direction to make the detection line away from the contour line, wherein p is a positive integer greater than or equal to 1, and the distance of each translation is the preset distance; after the p-th translation, determining the fragment judgment score corresponding to the p-th translation according to the pixel values ​​of the pixel points on the detection line after the p-th translation; determining the bone fragment judgment score according to the fragment judgment score corresponding to the initial position and the fragment judgment scores corresponding to each translation.

8. The bone medical image damage detection method according to claim 7, characterized in that: After the p-th translation, the fragment discrimination score corresponding to the p-th translation is determined according to the pixel values ​​of the pixels on the detection line after the p-th translation, including: according to the formula Get the fragment discrimination score D corresponding to the pth translation p , where (x p，q ,y p，q ) is the coordinate of the qth pixel on the detection line after the pth translation, p(x p，q ,y p，q ) is the pixel value of the qth pixel on the detection line after the pth translation, is the direction vector of the qth pixel on the detection line after the pth translation, is the pixel value of the qth pixel point on the detection line after the pth translation along the direction vector The gradient of R T，2 is the preset second change rate threshold, n p is the number of pixels on the detection line, and p, q and n p All are positive integers.

9. The bone medical image damage detection method according to claim 7, characterized in that: Determining the bone fragment judgment score according to the fragment judgment score corresponding to the initial position and the fragment judgment score corresponding to each translation includes: according to the formula Determine the bone fragment judgment score D, where D p is the fragment discrimination score corresponding to the pth translation, N is the total number of translations, 0≤p≤N, and when p=0, D0 is the fragment discrimination score corresponding to the initial position, p and N are both positive integers, and if is a conditional function.

10. A bone medical image damage detection system, characterized in that: include: A bone binary image module is used to perform binary processing on the bone medical image to be processed to obtain a bone binary image; The contour line module is used to detect the bone binary image through the contour detection model to obtain the bone contour line in the bone binary image; the bone fracture judgment score module is used to determine the bone fracture judgment score according to the position coordinates of multiple pixel points on the contour line; the midline module is used to obtain the midline of two adjacent bone contour lines according to the bone contour line; the bone crack judgment score module is used to determine the bone crack judgment score according to the midline and the pixel values ​​of the pixel points of the bone area within the contour line in the bone binary image; the detection line module is used to determine at least one of the bone crack judgment score and the bone fracture judgment score. When the bone fracture judgment score is greater than or equal to a preset first injury score threshold, a detection line parallel to the contour line is set in the background area outside the bone area according to the contour line of the bone; a bone fragment judgment score module is used to obtain the bone fragment judgment score according to the pixel value of the pixel point in the background area and the detection line; a bone injury judgment score module is used to determine the bone injury judgment score according to the bone fracture judgment score, the bone crack judgment score and the bone fragment judgment score; a marking module is used to mark the bone medical image to be processed if the bone injury judgment score is greater than or equal to a preset second injury score threshold.

Images