A method and system for intelligently evaluating sports injury status

By analyzing the feature changes of ultrasound image maps, an enhanced reconstruction ultrasound map is constructed and inputted to the evaluation network model, the problem of blurred image interference in ultrasound examination is solved and the accuracy of motion injury assessment is improved.

CN118781125BActive Publication Date: 2024-11-15CIVIL AVIATION FLIGHT UNIV OF CHINA
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202411276386.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-12
Publication Date
2024-11-15
Estimated Expiration
2044-09-12

AI Technical Summary

Technical Problem

In ultrasound examination, due to blurred image interference caused by patient movement, it is difficult to accurately evaluate motion injuries, and the prior art is difficult to effectively eliminate blurred interference, affecting the accuracy of the degree of damage.

Method used

By obtaining the patient's medical history and multiple continuous ultrasound images, analyzing the image feature changes, determining the fuzzy factor combination, building an enhanced reconstructed ultrasound image, eliminating fuzzy interference, and inputting the reconstructed ultrasound image and medical history to the evaluation network model , obtain the evaluation results of the sports injury status.

Benefits of technology

It improves the accuracy of determining the degree of tissue damage and the accuracy of the evaluation results of the motor injury status, eliminates image blur interference, and enhances the accuracy of the damage assessment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118781125B_ABST
    Figure CN118781125B_ABST
Patent Text Reader

Abstract

The present invention relates to the technical field of medical care informatics, and in particular to an intelligent evaluation method and system for a motion injury state. The method comprises obtaining a plurality of continuous ultrasonic images of a patient's medical history and a lesion location, analyzing image feature changes between the plurality of continuous ultrasonic images, determining a fuzzy factor combination, and constructing an enhanced reconstructed ultrasonic image based on the fuzzy factor combination. The enhanced reconstructed ultrasonic image is constructed by the fuzzy factor combination of the image feature change analysis, which is conducive to eliminating interference such as blur in the image and obtaining a clearer reconstructed ultrasonic image. The degree of tissue damage is determined based on the reconstructed ultrasonic image, thereby improving the accuracy of determining the degree of tissue damage. The reconstructed ultrasonic image, the medical history and the degree of tissue damage are input into an evaluation network model to obtain an evaluation result of the motion injury state, thereby improving the accuracy of the evaluation result of the motion injury state.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of medical care informatics, and in particular to a method and system for intelligently evaluating a sports injury state. Background Art

[0002] Sports injuries refer to injuries to body tissues caused by sports training, competition, physical exercise, or other physical activities. Sports injuries are common in muscles, tendons, joints, ligaments, and other tissue structures. Common types of injuries include joint injuries, muscle strains, ligament cartilage injuries, and tendinitis. In summary, sports injuries are mostly soft tissue injuries, and accurate assessment of sports injuries is of great significance for doctors, nutritionists, and others to develop reasonable treatment plans and conditioning plans.

[0003] At present, when detecting physical injuries, commonly used detection methods include X-ray examination, CT scan, MRI examination, ultrasound examination and other detection methods. However, since sports injuries are mostly soft tissue injuries, ultrasound examination and MRI examination are commonly used. Since ultrasound detection is an instant detection feedback image, it is necessary to move the ultrasound probe during the process to facilitate the detection of human injuries from different angles. However, during the detection process, the person being examined may twist his body due to discomfort or other reasons, which will cause blurring and other related interference in the image during the ultrasound examination, resulting in an inability to accurately evaluate the final result. Summary of the invention

[0004] In order to solve the above technical problems, the purpose of the present invention is to provide a method and system for intelligent assessment of sports injury status, and the technical solutions adopted are as follows:

[0005] In a first aspect, the present application provides a method for intelligently assessing sports injury status, comprising:

[0006] Obtain multiple consecutive ultrasound images of the patient's medical history and lesion location;

[0007] Analyzing changes in image features between a plurality of consecutive ultrasound images, determining a fuzzy factor combination, and constructing an enhanced reconstructed ultrasound image based on the fuzzy factor combination;

[0008] determining the extent of tissue damage according to the reconstructed ultrasound image;

[0009] The reconstructed ultrasound image, the medical history and the degree of tissue damage are input into an evaluation network model to obtain an evaluation result of the motion injury state.

[0010] In one embodiment, analyzing the image feature changes between a plurality of consecutive ultrasound images to determine the fuzzy factor combination includes:

[0011] According to a sliding window of a preset size and a preset step size, synchronous sliding is performed in each of the ultrasonic image images to obtain image windows corresponding to a plurality of image areas in each of the ultrasonic image images;

[0012] Determining similarity parameters between pixels in each of the image windows according to the grayscale values ​​of each of the image windows;

[0013] Determining a fuzzy factor combination according to the similarity parameter;

[0014] The fuzzy factor combination includes the fuzzy factors of each image window in each of the ultrasonic images; the similarity parameter represents the motion blur of the pixel points, and the size of the similarity parameter is positively correlated with the degree of motion blur of the pixel points.

[0015] In one implementation, determining the similarity parameter between pixels in each of the image windows according to the grayscale value of each of the image windows includes:

[0016] Determine a first ultrasonic image and a second ultrasonic image from each of the ultrasonic image images, determine a first image window corresponding to a target image area from the first ultrasonic image, and determine a second image window corresponding to the target image area from the second ultrasonic image;

[0017] Determine a similarity parameter between each first pixel and each second pixel in the first image window according to a first gray value of each first pixel and a second gray value of each second pixel in the second image window at the same position as the first pixel, re-determine the first image window, and return to the step of determining the second image window corresponding to the target image area from the second ultrasonic image until a similarity parameter between each first pixel and each second pixel in each image window in the first ultrasonic image is determined;

[0018] Determine a new first ultrasonic image and a new second ultrasonic image, and return to the step of determining the first image window corresponding to the target image area from the first ultrasonic image until the similarity parameters between the pixel points in each image window in each ultrasonic image are determined.

[0019] In one implementation, determining the fuzzy factor combination according to the similarity parameter includes:

[0020] Performing a summation process according to the similarity parameters between each first pixel point and the second pixel point in each of the image windows to obtain a sum of similarity parameters of each of the image windows;

[0021] Determining a similarity parameter mean value of each image window according to the similarity parameter and the number of pixels in each image window;

[0022] According to the average function and the mean value of the similarity parameter of each image window, the fuzzy factor of each image window in each ultrasound image is calculated respectively.

[0023] In one embodiment, constructing an enhanced reconstructed ultrasound image according to the combination of fuzzy factors includes:

[0024] Constructing enlarged reference areas with each of the image windows as the center;

[0025] Determine each target pixel point included in the reference area, and determine the target grayscale value of each target pixel point according to the fuzzy factor corresponding to the image window where each target pixel point is located;

[0026] According to each of the target grayscale values, construct a fuzzy view of each corresponding image window, and determine a reconstruction index of each of the image windows in each of the ultrasound images according to the fuzzy view;

[0027] Using the reconstruction index of the image window as a weight respectively, calculating the weighted pixel mean of the pixel values ​​in the corresponding image window;

[0028] An enhanced reconstructed ultrasound image is constructed according to the weighted pixel means.

[0029] In one embodiment, determining the reconstruction index of each image window in each ultrasound image according to the fuzzy view includes:

[0030] Determine a specified angle, determine a grayscale gradient value of a pixel at the specified angle in each of the blurred views, update the specified angle, and return to the step of determining the grayscale gradient value of the pixel at the specified angle in each of the blurred views, until the grayscale gradient value of each pixel at different angles in each of the blurred views is obtained, wherein the grayscale gradient value refers to the grayscale difference between adjacent pixels on both sides of the pixel at the specified angle;

[0031] Calculate the sum of the grayscale gradient values ​​of each pixel at each corresponding angle respectively, and determine the total number of pixels whose grayscale gradient values ​​are in a decreasing trend at each corresponding angle;

[0032] Taking the total amount of reduction as a weight, and calculating the mean grayscale gradient value corresponding to each blurred view according to the sum of the grayscale gradient values ​​at each angle and the weight;

[0033] According to the grayscale gradient value mean and the hyperbolic tangent function, a reconstruction index of each image window in each of the ultrasonic image graphs is obtained.

[0034] In one embodiment, determining the degree of tissue damage according to the reconstructed ultrasound image comprises:

[0035] According to the detection windows of the specified size and the specified step length, a plurality of detection windows are determined in the reconstructed ultrasound image, and the detection windows are processed by an edge detection algorithm to obtain corresponding edge detection images;

[0036] Binarizing the edge detection image, scanning the binarization result by a connected domain labeling algorithm, and determining at least one tissue in each detection window;

[0037] Determining the texture factor of each of the tissues in each of the detection windows respectively;

[0038] Determining the consistency of tissue texture between the tissues in the same detection window according to the texture factor;

[0039] The product of the texture factor of each tissue and the consistency of the tissue texture is determined respectively, and the degree of damage of each tissue is determined according to the product and the exponential function.

[0040] In one embodiment, the determining the texture factor of each tissue in each detection window comprises:

[0041] Determining target tissues in each of the tissues in the detection window, and determining each edge of the target tissues;

[0042] Determine the angle between the gradient directions of each first pixel point on each of the edges and a second pixel point adjacent to the first pixel point, and determine the function value corresponding to each of the edges according to each of the angles and the cosine function;

[0043] Determine a mean value of the function value corresponding to each edge according to each of the function values ​​and the number of the function values ​​in each of the detection windows;

[0044] Determining the texture factor of the target tissue according to the ratio of the sum of the mean values ​​of the function values ​​to the number of the edges;

[0045] New target tissues are re-determined in each of the tissues in the detection window, and the steps of determining and each edge of the target tissue are returned to until the texture factors of each of the tissues are obtained.

[0046] In one embodiment, determining the consistency of tissue texture between the tissues in the same detection window according to the texture factor includes:

[0047] Determining target tissues in each of the tissues in the detection window respectively;

[0048] Determine the absolute values ​​of texture factor differences of the texture factor of the target tissue and the texture factors of other tissues adjacent to the target tissue respectively, and determine a mean value of the texture factor differences according to the sum of the absolute values ​​of the texture factor differences and the number of the texture factor differences;

[0049] Determining the consistency of tissue textures of the target tissue and other tissues adjacent to the target tissue in the same detection window according to an exponential function and the mean value of the texture factor difference;

[0050] Re-determine a new target tissue in each of the tissues in the detection window, and return to the step of respectively determining the absolute value of the texture factor difference between the texture factor of the target tissue and the texture factor of other tissues adjacent to the target tissue, until the tissue texture consistency between the various tissues in the same detection window in each of the detection windows is determined.

[0051] In a second aspect, the present application provides a sports injury status intelligent assessment system, comprising:

[0052] An acquisition module, used to acquire multiple continuous ultrasound images of the patient's medical history and the location of the lesion;

[0053] An analysis module, used for analyzing changes in image features between a plurality of consecutive ultrasonic images, determining a fuzzy factor combination, and constructing an enhanced reconstructed ultrasonic image according to the fuzzy factor combination;

[0054] A determination module, used to determine the degree of tissue damage based on the reconstructed ultrasound image;

[0055] The evaluation module is used to input the reconstructed ultrasound image, the medical history and the degree of tissue damage into the evaluation network model to obtain an evaluation result of the sports injury state.

[0056] The present invention has the following beneficial effects:

[0057] By obtaining the patient's medical history and multiple continuous ultrasound images of the lesion location, the image feature changes between the multiple continuous ultrasound images are analyzed, the fuzzy factor combination is determined, and based on the fuzzy factor combination, an enhanced reconstructed ultrasound image is constructed. The enhanced reconstructed ultrasound image is constructed by the fuzzy factor combination of the image feature change analysis, which is beneficial to eliminate interference such as blur in the image and obtain a clearer reconstructed ultrasound image. The degree of tissue damage is determined based on the reconstructed ultrasound image, thereby improving the accuracy of determining the degree of tissue damage; the reconstructed ultrasound image, medical history and tissue damage degree are input into the evaluation network model to obtain the evaluation result of the motion injury state, thereby improving the accuracy of the evaluation result of the motion injury state. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] 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.

[0059] Figure 1 A schematic diagram of the steps of a method for intelligently assessing sports injury status provided by one embodiment of the present invention;

[0060] Figure 2 A schematic diagram of a plurality of continuous ultrasonic images according to an embodiment of the present invention;

[0061] Figure 3 A schematic diagram of the movement of an image window according to an embodiment of the present invention;

[0062] Figure 4 A structural block diagram of a sports injury status intelligent assessment system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0063] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of a method and system for intelligently evaluating the state of sports injuries proposed by the present invention, its specific implementation, structure, features and effects, in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0064] 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.

[0065] The specific scheme of the intelligent assessment method and system for sports injury status provided by the present invention is described in detail below with reference to the accompanying drawings.

[0066] See also Figure 1 , which shows a flow chart of a method for intelligently evaluating a sports injury state provided by an embodiment of the present invention. The method for intelligently evaluating a sports injury state may at least include steps S100-S400:

[0067] S100, obtaining a plurality of continuous ultrasound images of the patient's medical history and the location of the lesion.

[0068] S200, analyzing image feature changes between a plurality of consecutive ultrasound images, determining a fuzzy factor combination, and constructing an enhanced reconstructed ultrasound image based on the fuzzy factor combination.

[0069] S300: Determine the degree of tissue damage according to the reconstructed ultrasound image.

[0070] S400, inputting the reconstructed ultrasound image, the medical history and the degree of tissue damage into the evaluation network model to obtain the evaluation result of the sports injury status.

[0071] The technical solution of the embodiment of the present application obtains a plurality of continuous ultrasonic images of the patient's medical history and the location of the lesion, analyzes the image feature changes between the plurality of continuous ultrasonic images, determines a fuzzy factor combination, and constructs an enhanced reconstructed ultrasonic image based on the fuzzy factor combination. The enhanced reconstructed ultrasonic image is constructed by the fuzzy factor combination of the image feature change analysis, which is helpful to eliminate interference such as blur in the image and obtain a clearer reconstructed ultrasonic image. The degree of tissue damage is determined based on the reconstructed ultrasonic image, thereby improving the accuracy of determining the degree of tissue damage. The reconstructed ultrasonic image, the medical history, and the degree of tissue damage are input into the evaluation network model to obtain an evaluation result of the motion injury state, thereby improving the accuracy of the evaluation result of the motion injury state.

[0072] In one embodiment, in order to comprehensively evaluate the patient's sports injury status, the present invention needs to obtain various patient information, such as obtaining and recording the patient's medical history by asking questions, such as whether the patient has had any sports injuries in the past, what kind of sports injuries the patient has had, and what kind of treatment the patient has received; the current location of the lesion, pain, swelling and other related conditions; and checking the condition and locating the lesion through various means such as touch, press and activity testing. Then, with respect to the lesion location, one or more lesion locations are checked through ultrasound detection (B-mode ultrasound, GE Healthcare's "Desktop Voluson E10" desktop ultrasound device), and multiple ultrasound images (referred to as ultrasound images) corresponding to the lesion location that are continuous in time (continuous frames) are obtained, such as Figure 2As shown, the collected medical history and ultrasound images can be stored in an information database.

[0073] It should be noted that, during ultrasound examination, the probe is close to the skin for examination. When the probe is moved for scanning or the patient's body moves, the ultrasound image will also change. However, since the probe is close to the skin surface, the closer the position is to the skin, the smaller the change is. The changes in the body tissues in the middle are the largest, and then the changes close to the bones gradually become smaller. According to these rules, the movable body tissues are determined and suppressed. Then, during ultrasound detection, different body tissues react differently to ultrasound, that is, different body tissues have different echo intensities to ultrasound, and related soft tissues such as muscles have certain textures because they form fibers and then form tissues. When soft tissues are damaged, these textures will cause damage, so the corresponding damage can be identified from the ultrasound image.

[0074] In addition, during ultrasound testing, the doctor takes the ultrasound probe and sticks it to the surface of the body, constantly turning the probe handle or shaking the ultrasound probe in order to obtain a more comprehensive and detailed soft tissue structure. In this process, by changing the position and angle of the probe, the damaged area can be observed from different directions and levels. Moreover, due to the movement of the probe, the patient's damaged part is stimulated and feels uncomfortable, so the body twists, causing the ultrasound image to shake. And it is precisely because of the above two reasons that the image of the upper part of the ultrasound image (close to the skin) is more stable and unchanged, while the lower part of the image (soft tissue) is farther away from the probe and bones, and the image of the corresponding part changes more with the same shaking amplitude, but because it is farther away than the skin. Moreover, due to the different refresh rates of the detector, the part with greater changes will produce partial motion blur, resulting in failure to detect body injuries.

[0075] In one embodiment, the step S200 analyzes the image feature changes between a plurality of consecutive ultrasound images to determine the fuzzy factor combination, including steps S201-S203:

[0076] S201. According to a sliding window of a preset size and a preset step size, synchronous sliding is performed in each ultrasonic image to obtain image windows corresponding to a plurality of image regions in each ultrasonic image.

[0077] Optionally, in a preset size for For example, the preset step size is 2 as an example to illustrate. The size and the preset step size of 2 are used. The lower left corner of the ultrasound image is used as the starting point. The movement method is to move rightward first and then upward in a serpentine manner. The above method is used to synchronously slide in each ultrasound image to obtain image windows corresponding to several image areas in each ultrasound image. For example, the sliding window setting and movement are as follows: Figure 3 As shown, the ultrasound image shows sliding windows corresponding to 5 image areas. For example, at the first moment, there is a first sliding window in the lower left corner, and at the second moment, there is a second sliding window. Correspondingly, at the same time, for example, at the first moment, each ultrasound image will have a corresponding first sliding window at the same lower left corner position, so the corresponding sliding window in each ultrasound image can be obtained at the same moment.

[0078] S202: Determine similarity parameters between pixels in each image window according to the grayscale value of each image window.

[0079] First, a first ultrasound image is determined from each ultrasound image, and is recorded as ultrasound image , and the second ultrasound image, denoted as ultrasound image , determining a first image window corresponding to the target image area from the first ultrasound image , denoted as , and determining a second image window corresponding to the target image region (ie, the same region) from the second ultrasound image , denoted as ; The target image area is any image area among the several image areas in step S201.

[0080] Secondly, according to the first gray value of each first pixel in the first image window (Ultrasound image Internal Image Window Middle Line The grayscale value of the column, different values ​​are taken to represent each first pixel point) and the second grayscale value of each second pixel point at the same position as the first pixel point in the second image window (Indicates ultrasound image Internal Image Window Middle Line The grayscale value of the column, Different values ​​are taken to represent each second pixel point), and the similarity parameter between each first pixel point and each second pixel point is determined: ,in; represents the exponential function, which is used for inverse normalization. It should be noted that for the same two image windows, The ratio obtained may have different calculation results, so in order to improve the accuracy, the similarity parameter can also be expressed as: , for and The minimum value in for and The maximum value among ; or, for and The maximum value in for and For the convenience of explanation, this application uses the similarity parameter as This is for illustrative purposes only and is not intended to be limiting.

[0081] It should be noted that the similarity parameter represents the motion blur of the pixel, and the size of the similarity parameter is positively correlated with the degree of motion blur of the pixel. tends to 1, that is tends to 1, and The similarity parameter tends to 0. The smaller the similarity parameter is, the more likely it is that the ultrasound image The pixel point (the first pixel point) is normal, otherwise it is considered that the pixel point (the first pixel point) is caused by motion blur. This is because the delayed drift of the image caused by motion blur will not appear in multiple image frames, that is, it will gradually return to normal when the image stabilizes.

[0082] Furthermore, the first image window is re-determined, and the step of determining the second image window corresponding to the target image area from the second ultrasonic image is returned until the similarity parameter between each first pixel point and the second pixel point in each image window in the first ultrasonic image is determined, which is equivalent to When it is 1, it represents the first first image window and the first second image window. When different values ​​are taken, the similarity parameter between each first pixel point and the second pixel point in each image window of the first ultrasonic image can be determined. , Different values ​​refer to different image windows.

[0083] Then, a new first ultrasonic image and a new second ultrasonic image are determined, and the step of determining the first image window corresponding to the target image area from the first ultrasonic image is returned until the similarity parameter between the pixels in each image window in each ultrasonic image is determined, that is, middle, , Each image window represents a different ultrasound image when taking different values The similarity parameter between pixels in .

[0084] S203: Determine a fuzzy factor combination according to the similarity parameter.

[0085] First, a summation process is performed based on the similarity parameters between each first pixel and the second pixel in each image window to obtain the similarity parameters and ,in, For image window The total number of banks, For image window The total number of columns.

[0086] Secondly, according to the similarity parameter and the number of pixels in each image window , determine the mean similarity parameter of each image window: .

[0087] Then, according to the average function and the mean value of the similarity parameter of each image window, the fuzzy factor of each image window in each ultrasound image is calculated. , so the final formula is:

[0088] ;

[0089] in, Ultrasound image Middle Image Window The fuzzy factor, Different values ​​are taken to represent each ultrasound image. Take different values ​​to represent each image window, represents the average function for finding the mean value; it should be noted that since there are many pixels in the image window of the ultrasound image, the Is the image window Is there any blur caused by motion in the image? Therefore, multiple The average value of pixels , and because the comparison is done through the same position pixels between multiple frames of ultrasound images, it is necessary to traverse all ultrasound images to investigate, so the average .

[0090] It should be noted that the fuzzy factor combination includes the fuzzy factors of each image window in each ultrasound image.

[0091] In one embodiment, in step S200, an enhanced reconstructed ultrasound image is constructed according to the combination of fuzzy factors, including steps S204-S208:

[0092] S204, constructing an enlarged reference area with each image window as the center.

[0093] Optionally, each image window (referred to as image window , Different values ​​represent different image windows) as the center and the neighborhood side length as Construct an amplified reference field.

[0094] S205 , determining each target pixel point included in the reference area, and determining a target grayscale value of each target pixel point according to a fuzzy factor corresponding to the image window where each target pixel point is located.

[0095] Optionally, each target pixel point included in the reference area is determined, and the fuzzy factor of the image window corresponding to each target pixel point is determined. , using the fuzzy factor Determine the target grayscale value of each target pixel, the target grayscale value , Represents the ceiling function.

[0096] S206 , constructing fuzzy views of corresponding image windows according to the target grayscale values, and determining reconstruction indexes of the image windows in each ultrasound image according to the fuzzy views.

[0097] Then, based on the target grayscale values ​​corresponding to the image windows, fuzzy views of the corresponding image windows are constructed, and the grayscale values ​​of the pixels in the fuzzy views are all target grayscale values.

[0098] Optionally, in S206, determining the reconstruction index of each image window in each ultrasound image according to the fuzzy view includes the steps of:

[0099] First, determine the specified angle , respectively, in each blurred view, determine the specified angle The grayscale gradient value of the pixel on the right. The grayscale gradient value refers to the grayscale difference between the adjacent pixels on both sides of the pixel at a specified angle. For example, there is a pixel A, and the specified angle of pixel A is On the left, pixel A is on one side of pixel B, and pixel A is on the other side of pixel C, that is, at a specified angle In the figure, the previous pixel of pixel A is pixel B, and the next pixel of pixel A is pixel C. At this time, the grayscale gradient value of pixel A is the grayscale difference between pixel B and pixel C. Defined as: 0° to the right, clockwise is the positive direction, update the specified angle, for example, the specified angle at the beginning The specified angle is updated based on a preset rotation angle (for example, including but not limited to a range of 1-5°), and then the gray gradient value of the pixel at the specified angle is determined in each blurred view, until different angles (i.e., different specified angles) are obtained in each blurred view. ) , Refers to the pixels in the blurred view At a specified angle The gray gradient value on =1 represents the first pixel.

[0100] Secondly, calculate the sum of the grayscale gradient values ​​of each pixel at each corresponding angle: ,in, Blur the pixels in the view and determine the total number of pixels whose grayscale gradient values ​​are decreasing at each corresponding angle That is, if there is a trend of grayscale decrease between two pixels, the total number of decreases is recorded as +2, and the same pixel is recorded once. If there is a trend of grayscale increase, the total number of decreases is recorded as 0.

[0101] Furthermore, it will reduce the total amount As a weight, according to the sum of the gray gradient values ​​at each angle And weights, calculate the mean gray gradient value corresponding to each blurred view .in, Indicates the utilization weight , calculate different Corresponding function value The weighted average of , that is, the mean of the grayscale gradient values.

[0102] Then, according to the mean gray gradient value and the hyperbolic tangent function , and obtain the reconstruction index of each image window in each ultrasound image , the specific formula is:

[0103] ;

[0104] Among them, the reconstruction index Ultrasound image Middle Image Window The reconstruction index of Different values ​​are taken to represent each ultrasound image. Take different values ​​to represent each image window; hyperbolic tangent function Used for normalization of positive proportion; The larger the value, the more In this angle The greater the grayscale reduction value, the greater the grayscale here reflects the evaluation of motion blur, that is, the larger the sum value is, the more blurred the corresponding angle is, that is, the blur in this direction is gradually increasing or decreasing. In addition, due to Describes the image window Therefore, it is necessary to traverse all angles in this image window and find the average. In addition, due to the above-mentioned diffusion, the blurring conditions at different angles are different, so weighting is performed when calculating the mean according to the length of the decreasing trend, that is, the greater the length of the trend, the greater the weight of the corresponding sum value.

[0105] It should be noted that since the various tissues of the body are relatively connected, when a certain part of the body moves, other nearby tissues will move along with it. In other words, the motion blur of the body tissue is diffuse and gradually decreases in certain directions. Therefore, the reconstruction index of each image window is considered. .

[0106] S207 , using the reconstruction index of each image window as a weight, and calculating a weighted pixel mean of the pixel values ​​in the corresponding image window.

[0107] The specific formula is:

[0108] ;

[0109] In the formula, is the pixel point in the image window The weighted pixel mean of the pixel values ​​(i.e. the reconstructed grayscale value), Indicates The ultrasonic image corresponds to the pixel point in the window The gray value of Indicates Pixels in the ultrasound image Image window Reconstruction indicators; Represents the reconstruction index of the image window As a weight, After weighted summation, the average value is calculated to obtain the weighted pixel mean of the pixel values ​​in the corresponding image window ; For example, pixel is the first pixel, , at this time the reconstruction index of the image window is , determine the first weighted term as × , pixel is the second pixel, , at this time the reconstruction index of the image window is , determine the second weighted term as × , summing up the multiple weighted items finally determined, and then calculating the average value according to the number of pixels, that is, the number of weighted items.

[0110] S208: construct an enhanced reconstructed ultrasound image according to the weighted pixel means.

[0111] Optionally, based on each weighted pixel mean, a corresponding new enhancement window is constructed, and then the enhancement window in the original ultrasound image is updated to the new enhancement window, and the enhanced reconstructed ultrasound image is constructed respectively. .

[0112] The embodiment of the present invention removes image blur caused by the body and the ultrasound probe by analyzing the texture features of each part of the ultrasound image and the correlation changes between multiple acquired ultrasound images, thereby achieving effective denoising and constructing a clear reconstructed ultrasound image. , avoiding misjudgment of body injuries caused by image blur, based on the reconstructed ultrasound image after enhancing quality , improving the ability and accuracy of identifying tissue damage.

[0113] It should be noted that in ultrasound testing, when ultrasound passes through body tissues, the density and structure of normal and healthy body tissues are relatively uniform, which leads to relatively uniform reflection of ultrasound waves. Therefore, the ultrasound images formed by the same body tissues are relatively consistent. Different body tissues have relatively different densities and different resistances to ultrasound absorption. Ultrasound waves propagate in tissues of different densities, making the edges between different tissues relatively clear. Therefore, normal body tissues are relatively neat and orderly in ultrasound images. When soft tissue contusions occur, the texture of the tissues in the ultrasound images will be destroyed. Subsequent analysis is based on this principle.

[0114] In one implementation, step S300 includes steps S301-S305:

[0115] S301 . Determine a number of detection windows in the reconstructed ultrasound image according to a detection window of a specified size and a specified step size, and process the detection windows by an edge detection algorithm to obtain a corresponding edge detection image.

[0116] Optionally, taking the specified size as 5×5 and the specified step size as 3 as an example, a number of detection windows are determined in the reconstructed ultrasound image, and any detection window is recorded as the detection window , can be moved in a similar way to the image window above, and by The edge detection algorithm processes the detection window to obtain the corresponding edge detection image.

[0117] S302 , binarizing the edge detection image, scanning the binarization result by a connected component labeling algorithm, and determining at least one tissue in each detection window.

[0118] Then, the edge detection image is binarized and the connected component labeling algorithm ( ) Scan the binarization results to determine each detection window At least one tissue (also referred to as local tissue) in the body.

[0119] S303, respectively determine the texture factor of each tissue in each detection window.

[0120] First, in the detection window Determine the target organization (any organization) among the organizations in , get the detection window The edge of the inner image and the gradient size and direction corresponding to the edge pixel on the edge, select one edge at a time, and determine the edges of the target tissue ( Different values ​​represent different edges).

[0121] Secondly, determine each edge separately Each first pixel point (denoted as pixel point) and the second pixel point adjacent to the first pixel point (denoted as the +1 pixel) , that is, organization The Edge Previous The gradient direction of a pixel and its adjacent pixels The angle between the gradient directions; and according to each angle and the cosine function, determine the function value corresponding to each edge .

[0122] Furthermore, according to each function value and the number of function values ​​in each detection window, the mean function value corresponding to each edge is determined: , Indicates the calculation of each edge The mean value of the function among the pixels in .

[0123] Then, according to the ratio of the sum of the mean values ​​of each function value to the number of edges, that is, according to the ratio of the sum of the mean values ​​of each function value and the number of edges, the mean is calculated again to obtain the target tissue (that is, each tissue )'s texture factor , the specific formula is:

[0124] ;

[0125] In the formula, represents mean calculation, that is, calculating the sum of the means of each function value and the ratio of the number of edges, For detection window Medium Organization Texture factor used to describe the detection window Internal Organization Texture.

[0126] Finally, the new target tissue is re-determined in each tissue of the detection window, and the steps of determining each edge of the target tissue are returned until the texture factor of each tissue is obtained, that is, , that is, different detection windows are obtained Different organizations in The texture factor.

[0127] S304: Determine the consistency of tissue texture between tissues in the same detection window according to the texture factor.

[0128] First, the target tissue is determined in each tissue of the detection window. Similarly, the target tissue is recorded as tissue .

[0129] Secondly, determine the texture factor of the target tissue and other tissues adjacent to the target tissue (referred to as tissue )'s texture factor The absolute value of the texture factor difference ,organize It refers to the organization One of the other tissues in the same detection window or in the adjacent step window. And according to the sum of the absolute values ​​of the texture factor differences and the number of texture factor differences (i.e. the number of other tissues), determine the mean value of the texture factor difference: .

[0130] Furthermore, according to the exponential function and the mean value of the texture factor difference, the consistency of the tissue texture of the target tissue and other tissues adjacent to the target tissue in the same detection window is determined respectively, and the formula is:

[0131] ;

[0132] In the formula, For detection window Medium Organization consistency of tissue texture with other adjacent tissues, Represents the exponential function, used for inverse normalization.

[0133] Then, a new target tissue is re-determined in each tissue in the detection window, and the step of determining the absolute value of the texture factor difference between the texture factor of the target tissue and the texture factor of other tissues adjacent to the target tissue is returned to, until the consistency of tissue texture between each tissue in the same detection window is determined, that is, , Different values ​​represent different detection windows. Different values ​​represent different organizations.

[0134] S305 , respectively determining the product of the texture factor of each tissue and the consistency of the tissue texture, and determining the degree of damage of each tissue according to the product and the exponential function.

[0135] Specifically, the product of the texture factor of each tissue and the tissue texture consistency is determined , according to the product and exponential function, determine the degree of damage to each tissue , the specific formula is:

[0136] ;

[0137] In the formula, Reconstructed Ultrasound Internal detection window Medium Organization The degree of damage, Different values ​​represent different detection windows. Different values ​​represent different organizations. Represents the degree of damage to different tissues in different detection windows.

[0138] In one embodiment, a neural network is constructed in advance, and a training set is constructed based on the historical medical history of a large number of patients, a large number of continuous historical ultrasound images corresponding to the lesion location, and the historical damage degree of each corresponding tissue obtained by analyzing the historical ultrasound images. The training set is used to train the neural network to obtain a trained neural network, which is recorded as an evaluation network model. The evaluation network model can output a state evaluation of the sports injury. Therefore, the medical history, reconstructed ultrasound image, and degree of tissue damage in step S100 are input into the evaluation network model, so that the evaluation network model outputs an evaluation result of the sports injury state. The evaluation result of the sports injury state obtained is more accurate and reliable, helping doctors to determine the appropriate treatment method and improve the relevant conditioning treatment effect. It should be understood that the process of constructing a training set and using the training set to train a neural network belongs to conventional techniques in the field and will not be repeated here.

[0139] Reference Figure 4 , shows a structural block diagram of a sports injury status intelligent assessment system according to an embodiment of the present application, the system may include:

[0140] An acquisition module, used to acquire multiple continuous ultrasound images of the patient's medical history and the location of the lesion;

[0141] An analysis module, used for analyzing image feature changes between a plurality of consecutive ultrasound images, determining a fuzzy factor combination, and constructing an enhanced reconstructed ultrasound image according to the fuzzy factor combination;

[0142] A determination module, used for determining the degree of tissue damage based on the reconstructed ultrasound image;

[0143] The evaluation module is used to input the reconstructed ultrasound image, medical history and tissue injury degree into the evaluation network model to obtain the evaluation result of the sports injury status.

[0144] In the embodiment of the present application, the functions of each module in the system can be found in the corresponding description in the above method and will not be repeated here.

[0145] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0146] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for intelligently assessing sports injury status, characterized in that: The method comprises: Obtain multiple consecutive ultrasound images of the patient's medical history and lesion location; Analyzing changes in image features between a plurality of consecutive ultrasound images, determining a fuzzy factor combination, and constructing an enhanced reconstructed ultrasound image based on the fuzzy factor combination; determining the extent of tissue damage according to the reconstructed ultrasound image; Inputting the reconstructed ultrasound image, the medical history and the degree of tissue damage into an evaluation network model to obtain an evaluation result of the sports injury state; The analyzing the image feature changes between the plurality of continuous ultrasonic images to determine the fuzzy factor combination comprises: According to a sliding window of a preset size and a preset step size, synchronous sliding is performed in each of the ultrasonic image images to obtain image windows corresponding to a plurality of image areas in each of the ultrasonic image images; Determining similarity parameters between pixels in each of the image windows according to the grayscale values ​​of each of the image windows; Determining a fuzzy factor combination according to the similarity parameter; The fuzzy factor combination includes the fuzzy factor of each image window in each of the ultrasound images; the similarity parameter represents the motion blur of the pixel point, and the size of the similarity parameter is positively correlated with the degree of motion blur of the pixel point; Determining the similarity parameters between the pixels in each of the image windows according to the grayscale values ​​of each of the image windows comprises: Determine a first ultrasonic image and a second ultrasonic image from each of the ultrasonic image images, determine a first image window corresponding to a target image area from the first ultrasonic image, and determine a second image window corresponding to the target image area from the second ultrasonic image; Determine a similarity parameter between each first pixel and each second pixel in the first image window according to a first gray value of each first pixel and a second gray value of each second pixel in the second image window at the same position as the first pixel, re-determine the first image window, and return to the step of determining the second image window corresponding to the target image area from the second ultrasonic image until a similarity parameter between each first pixel and each second pixel in each image window in the first ultrasonic image is determined; Determine a new first ultrasonic image and a new second ultrasonic image, and return to the step of determining the first image window corresponding to the target image area from the first ultrasonic image until a similarity parameter between pixel points in each image window in each ultrasonic image is determined; Determining the fuzzy factor combination according to the similarity parameter includes: Performing a summation process according to the similarity parameters between each first pixel point and the second pixel point in each of the image windows to obtain a sum of similarity parameters of each of the image windows; Determining a similarity parameter mean value of each image window according to the similarity parameter and the number of pixels in each image window; Calculating the fuzzy factor of each image window in each ultrasound image according to the average function and the mean value of the similarity parameter of each image window; The step of constructing an enhanced reconstructed ultrasound image according to the combination of fuzzy factors comprises: Constructing enlarged reference areas with each of the image windows as the center; Determine each target pixel point included in the reference area, and determine the target grayscale value of each target pixel point according to the fuzzy factor corresponding to the image window where each target pixel point is located; According to each of the target grayscale values, construct a fuzzy view of each corresponding image window, and determine a reconstruction index of each of the image windows in each of the ultrasound images according to the fuzzy view; Using the reconstruction index of the image window as a weight respectively, calculating the weighted pixel mean of the pixel values ​​in the corresponding image window; constructing an enhanced reconstructed ultrasound image according to each of the weighted pixel means; Determining the reconstruction index of each image window in each of the ultrasound image maps according to the fuzzy view includes: Determine a specified angle, determine a grayscale gradient value of a pixel at the specified angle in each of the blurred views, update the specified angle, and return to the step of determining the grayscale gradient value of the pixel at the specified angle in each of the blurred views, until the grayscale gradient value of each pixel at different angles in each of the blurred views is obtained, wherein the grayscale gradient value refers to the grayscale difference between adjacent pixels on both sides of the pixel at the specified angle; Calculate the sum of the grayscale gradient values ​​of each pixel at each corresponding angle respectively, and determine the total number of pixels whose grayscale gradient values ​​are in a decreasing trend at each corresponding angle; Taking the total amount of reduction as a weight, and calculating the mean grayscale gradient value corresponding to each blurred view according to the sum of the grayscale gradient values ​​at each angle and the weight; Obtaining a reconstruction index of each image window in each of the ultrasound images according to the grayscale gradient value mean and the hyperbolic tangent function; Determine the target grayscale value of each target pixel point, and the corresponding calculation formula is: h ζ,α =[255×H ζ,α ] Among them, h ζ,α Indicates the fuzzy factor of the image window corresponding to each target pixel, h ζ,α Represents the target grayscale value of each target pixel, Represents the ceiling function.

2. The method for intelligently assessing sports injury status according to claim 1, characterized in that: Determining the degree of tissue damage according to the reconstructed ultrasound image comprises: According to the detection windows of the specified size and the specified step length, a plurality of detection windows are determined in the reconstructed ultrasound image, and the detection windows are processed by an edge detection algorithm to obtain corresponding edge detection images; Binarizing the edge detection image, scanning the binarization result by a connected domain labeling algorithm, and determining at least one tissue in each detection window; Determining the texture factor of each of the tissues in each of the detection windows respectively; Determining the consistency of tissue texture between the tissues in the same detection window according to the texture factor; The product of the texture factor of each tissue and the consistency of the tissue texture is determined respectively, and the degree of damage of each tissue is determined according to the product and the exponential function.

3. The method for intelligently assessing sports injury status according to claim 2, characterized in that: The respectively determining the texture factor of each tissue in each detection window comprises: Determining target tissues in each of the tissues in the detection window, and determining each edge of the target tissues; Determine the angle between the gradient directions of each first pixel point on each of the edges and a second pixel point adjacent to the first pixel point, and determine the function value corresponding to each of the edges according to each of the angles and the cosine function; Determine a mean value of the function value corresponding to each edge according to each of the function values ​​and the number of the function values ​​in each of the detection windows; Determining the texture factor of the target tissue according to the ratio of the sum of the mean values ​​of the function values ​​to the number of the edges; New target tissues are re-determined in each of the tissues in the detection window, and the steps of determining and each edge of the target tissue are returned to until the texture factors of each of the tissues are obtained.

4. The method for intelligently assessing sports injury status according to claim 2, characterized in that: Determining the consistency of tissue texture between the tissues in the same detection window according to the texture factor includes: Determining target tissues in each of the tissues in the detection window respectively; Determine the absolute values ​​of texture factor differences of the texture factor of the target tissue and the texture factors of other tissues adjacent to the target tissue respectively, and determine a mean value of the texture factor differences according to the sum of the absolute values ​​of the texture factor differences and the number of the texture factor differences; Determining the consistency of tissue textures of the target tissue and other tissues adjacent to the target tissue in the same detection window according to an exponential function and the mean value of the texture factor difference; Re-determine a new target tissue in each of the tissues in the detection window, and return to the step of respectively determining the absolute value of the texture factor difference between the texture factor of the target tissue and the texture factor of other tissues adjacent to the target tissue, until the tissue texture consistency between the various tissues in the same detection window in each of the detection windows is determined.

5. An intelligent sports injury status assessment system, characterized in that: include: An acquisition module is used to obtain multiple continuous ultrasound images of the patient's medical history and the location of the lesion; An analysis module, used for analyzing changes in image features between a plurality of consecutive ultrasonic images, determining a fuzzy factor combination, and constructing an enhanced reconstructed ultrasonic image according to the fuzzy factor combination; A determination module, used to determine the degree of tissue damage based on the reconstructed ultrasound image; An evaluation module, used for inputting the reconstructed ultrasound image, the medical history and the degree of tissue damage into an evaluation network model to obtain an evaluation result of the sports injury state; The analyzing the image feature changes between the plurality of continuous ultrasonic images to determine the fuzzy factor combination comprises: According to a sliding window of a preset size and a preset step size, synchronous sliding is performed in each of the ultrasonic image images to obtain image windows corresponding to a plurality of image areas in each of the ultrasonic image images; Determining similarity parameters between pixels in each of the image windows according to the grayscale values ​​of each of the image windows; Determining a fuzzy factor combination according to the similarity parameter; The fuzzy factor combination includes the fuzzy factor of each image window in each of the ultrasound images; the similarity parameter represents the motion blur of the pixel point, and the size of the similarity parameter is positively correlated with the degree of motion blur of the pixel point; Determining the similarity parameters between the pixels in each of the image windows according to the grayscale values ​​of each of the image windows comprises: Determine a first ultrasonic image and a second ultrasonic image from each of the ultrasonic image images, determine a first image window corresponding to a target image area from the first ultrasonic image, and determine a second image window corresponding to the target image area from the second ultrasonic image; Determine a similarity parameter between each first pixel and each second pixel in the first image window according to a first gray value of each first pixel and a second gray value of each second pixel in the second image window at the same position as the first pixel, re-determine the first image window, and return to the step of determining the second image window corresponding to the target image area from the second ultrasonic image until a similarity parameter between each first pixel and each second pixel in each image window in the first ultrasonic image is determined; Determine a new first ultrasonic image and a new second ultrasonic image, and return to the step of determining the first image window corresponding to the target image area from the first ultrasonic image until a similarity parameter between pixel points in each image window in each ultrasonic image is determined; Determining the fuzzy factor combination according to the similarity parameter includes: Performing a summation process according to the similarity parameters between each first pixel point and the second pixel point in each of the image windows to obtain a sum of similarity parameters of each of the image windows; Determining a similarity parameter mean value of each image window according to the similarity parameter and the number of pixels in each image window; Calculating the fuzzy factor of each image window in each ultrasound image according to the average function and the mean value of the similarity parameter of each image window; The step of constructing an enhanced reconstructed ultrasound image according to the combination of fuzzy factors comprises: Constructing enlarged reference areas with each of the image windows as the center; Determine each target pixel point included in the reference area, and determine the target grayscale value of each target pixel point according to the fuzzy factor corresponding to the image window where each target pixel point is located; According to each of the target grayscale values, construct a fuzzy view of each corresponding image window, and determine a reconstruction index of each of the image windows in each of the ultrasound images according to the fuzzy view; Using the reconstruction index of the image window as a weight respectively, calculating the weighted pixel mean of the pixel values ​​in the corresponding image window; constructing an enhanced reconstructed ultrasound image according to each of the weighted pixel means; Determining the reconstruction index of each image window in each of the ultrasound image maps according to the fuzzy view includes: Determine a specified angle, determine a grayscale gradient value of a pixel at the specified angle in each of the blurred views, update the specified angle, and return to the step of determining the grayscale gradient value of the pixel at the specified angle in each of the blurred views, until the grayscale gradient value of each pixel at different angles in each of the blurred views is obtained, wherein the grayscale gradient value refers to the grayscale difference between adjacent pixels on both sides of the pixel at the specified angle; Calculate the sum of the grayscale gradient values ​​of each pixel at each corresponding angle respectively, and determine the total number of pixels whose grayscale gradient values ​​are in a decreasing trend at each corresponding angle; Taking the total amount of reduction as a weight, and calculating the mean grayscale gradient value corresponding to each blurred view according to the sum of the grayscale gradient values ​​at each angle and the weight; Obtaining a reconstruction index of each image window in each of the ultrasound images according to the grayscale gradient value mean and the hyperbolic tangent function; Determine the target grayscale value of each target pixel point, and the corresponding calculation formula is: h ζ,α =[255×H ζ,α ] Among them, H ζ,α Indicates the fuzzy factor of the image window corresponding to each target pixel, H ζ,α Represents the target grayscale value of each target pixel, Represents the ceiling function.

Citation Information

Patent Citations

  • Athlete athletic injury risk early warning method

    CN104200076A

  • Sports injury prompting method and device and electronic equipment

    CN116784828A

  • Sports injury severity assessment method and assessment device

    CN116798639A

  • Method and system for establishing neurological function prognosis prediction model of cerebral hemorrhage patient

    CN117174313A

  • System and method for monitoring brain trauma exposure

    WO2023035072A1