Scoliosis detection method and device, computer device and storage medium

Abstract

The present application relates to the field of image recognition, and discloses a scoliosis detection method and device, computer equipment and a storage medium, comprising: collecting video information of the back of a subject performing a predetermined action through a collection device, and selecting a preset number of key frames from the video information as recognition images; obtaining the edge contour of the subject according to each recognition image, standardizing each recognition image according to the edge contour, and obtaining a standardized image; calculating the scoliosis deviation angle in each standardized image, and outputting a corresponding waveform graph according to the scoliosis deviation angle and the change in time; and identifying the waveform graph to determine whether the subject has scoliosis. The detection of scoliosis can be carried out without professional measuring tools, and is convenient and universal.

Description

Technical Field

[0001] This invention relates to the field of image recognition, and more particularly to a method, apparatus, computer device, and storage medium for detecting scoliosis. Background Technology

[0002] Scoliosis is the third leading cause of serious harm to the physical and mental health of adolescents after obesity and myopia, and it is the most common spinal deformity among children and adolescents in my country. The simplest and most effective method for early detection of scoliosis is the Adams flexion test. The specific method involves the patient facing away from the examiner, bending forward with elbows hanging naturally and hands clasped, and then bending forward as if diving or trying to touch their toes with their fingers. If one side of the back bulges significantly or the trunk rotates during the flexion, the result is positive. Because this deformity is much more noticeable during flexion than when the patient is standing, it is widely used for screening adolescent scoliosis. However, assessing scoliosis based on this flexion test requires specialized medical knowledge and tools (such as a scoliosis measuring ruler), making it unsuitable for widespread home, school, or individual self-screening. Summary of the Invention

[0003] In a first aspect, this application provides a method for detecting scoliosis, including:

[0004] The device captures video information of the subject's back as they perform a predetermined action, and selects a preset number of keyframes from the video information as recognition images.

[0005] The edge contour of the subject is obtained from each recognition image, and the edge contour is used to standardize the output of each recognition image to obtain a standardized image.

[0006] Calculate the scoliosis deviation angle in each of the standardized images, and output the corresponding waveform based on the change of the scoliosis deviation angle over time;

[0007] The waveform is identified and judged to determine whether the subject has scoliosis.

[0008] Furthermore, the calculation of the scoliosis deviation angle in each standardized output recognition image includes:

[0009] The subject's shoulder contour is determined based on the edge contour, and the midline is determined based on the shoulder contour. Using the midline as a reference, two perpendicular lines are marked on each side of the midline according to the subject's shoulder width. The first and second intersection points formed by the left and right perpendicular lines intersecting with the shoulder or the edge contour are determined respectively.

[0010] Calculate the angle between the straight line connecting the first intersection point and the second intersection point and the horizontal line to obtain the scoliosis deviation angle.

[0011] Furthermore, the step of identifying and judging the waveform to determine whether the subject has scoliosis includes:

[0012] The peak value of the scoliosis deviation angle in the waveform is obtained, and it is determined whether the peak value exceeds a preset angle. If it exceeds the preset angle, the subject is determined to have scoliosis; otherwise, there is no scoliosis. The preset angle ranges from 3 to 4 degrees.

[0013] Furthermore, the step of identifying and judging the waveform to determine whether the subject has scoliosis includes:

[0014] The waveform is input into a pre-trained recognition model, which then determines whether the subject has scoliosis.

[0015] Furthermore, the training method for the recognition model includes:

[0016] Acquire back video information of multiple groups of scoliosis patients when performing the preset movements, and back video information of normal people when performing the preset movements;

[0017] The scoliosis deviation angle in each action video is obtained and a corresponding waveform is generated. Each waveform is labeled and used as training data to be input into a selected learning model for training to obtain the recognition model.

[0018] Furthermore, there are at least two acquisition devices;

[0019] The step of capturing back video information of the subject performing a predetermined action using a data acquisition device, and selecting a preset number of keyframes from the video information as recognition images, includes:

[0020] Video information from each of the acquisition devices is acquired. For each video information, a preset number of image frames are extracted as recognition images according to the playback order of the video. The recognition images extracted from the same video information are grouped together to obtain multiple groups of recognition images.

[0021] Furthermore, the step of obtaining the edge contours of the image subject based on each recognition image, and standardizing the output of each recognition image based on the contours, includes:

[0022] The various recognition images are compared to obtain the edge contours of the subject in each recognition image. Based on the edge contours, each recognition image is centered and cropped, and then output as a standardized image with the same preset ratio.

[0023] Secondly, this application also provides a scoliosis detection device, comprising:

[0024] The shooting module is used to capture video information of the back of the subject when performing a predetermined action through the acquisition device, and select a preset number of key frames from the video information as recognition images;

[0025] The preprocessing module is used to obtain the edge contour of the subject based on each recognition image, and to standardize the output of each recognition image based on the edge contour to obtain a standardized image.

[0026] The calculation module is used to calculate the scoliosis deviation angle in each of the standardized images and output the corresponding waveform based on the change of the scoliosis deviation angle over time.

[0027] The identification module is used to identify and judge the waveform to determine whether the subject has scoliosis.

[0028] Thirdly, this application also provides a computer device, including a processor and a memory, wherein the memory stores a computer program, and the computer program executes the scoliosis detection method when it is run on the processor.

[0029] Fourthly, this application also provides a readable storage medium storing a computer program that executes the scoliosis detection method when run on a processor.

[0030] This invention relates to the field of image recognition and discloses a method, apparatus, computer device, and storage medium for detecting scoliosis. The method includes: capturing video information of a subject's back as they perform a predetermined action using an acquisition device, and selecting a preset number of keyframes from the video information as recognition images; obtaining the subject's edge contours based on each recognition image, and standardizing each recognition image according to the edge contours to obtain standardized images; calculating the scoliosis deviation angle in each standardized image, and outputting a corresponding waveform based on the change of the scoliosis deviation angle over time; and identifying and judging the waveform to determine whether the subject has scoliosis. This method allows for scoliosis detection without specialized measurement tools, enabling convenient and universal detection. Attached Figure Description

[0031] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope of protection of the present invention. In the various drawings, similar components are numbered similarly.

[0032] Figure 1 A schematic flowchart of a scoliosis detection method according to an embodiment of this application is shown;

[0033] Figure 2 This illustration shows a schematic diagram of a scoliosis detection imaging scenario according to an embodiment of this application;

[0034] Figure 3 A standardized image schematic diagram of an embodiment of this application is shown;

[0035] Figure 4 A schematic diagram of standardized image processing according to an embodiment of this application is shown;

[0036] Figure 5 A waveform diagram of an embodiment of this application is shown;

[0037] Figure 6 A schematic diagram of a scoliosis detection device according to an embodiment of this application is shown. Detailed Implementation

[0038] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0039] The components of the embodiments of the invention described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0040] In the following, the terms “comprising,” “having,” and their cognates, which may be used in various embodiments of the invention, are intended only to indicate a particular feature, number, step, operation, element, component, or combination thereof, and should not be construed as excluding, firstly, the presence of one or more other features, numbers, steps, operations, elements, components, or combinations thereof, or adding the possibility of one or more features, numbers, steps, operations, elements, components, or combinations thereof.

[0041] Furthermore, the terms "first," "second," and "third" are used only to distinguish descriptions and should not be interpreted as indicating or implying relative importance.

[0042] Unless otherwise specified, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of the invention pertain. Terms (such as those defined in commonly used dictionaries) shall be interpreted as having the same meaning as in their contextual meaning in the relevant technical field and shall not be interpreted as having an idealized or overly formal meaning, unless clearly defined in the various embodiments of the invention.

[0043] The technical solution of this application is applied to the process of identifying scoliosis in the human body. It mainly uses a camera device to collect video of the subject performing the required actions, and then extracts a number of images from the video as recognition images. The scoliosis deviation angle is then obtained from these images, thus obtaining a waveform diagram of the subject's scoliosis deviation angle changing over time. Based on this waveform diagram, it can be determined whether the subject has scoliosis. Moreover, the detection process is fully automated, does not require manual operation by a doctor, and is suitable for non-professionals to conduct the detection in any location.

[0044] The technical solution of this application will now be described with reference to specific embodiments.

[0045] Example 1

[0046] like Figure 1 As shown, the scoliosis detection method of this application includes:

[0047] Step S100: Capture video information of the subject's back as they perform a predetermined action using a data acquisition device, and select a preset number of keyframes from the video information as recognition images.

[0048] Since this embodiment is for checking whether a subject has scoliosis, it mainly requires recording video of the back. The predetermined action performed by the subject can be the Adams flexion test used to check for scoliosis. The acquisition device records video information of the subject's movements while the subject performs the predetermined action. The video consists of multiple image frames, so a certain number of keyframes can be extracted in chronological order as recognition images. The specific number of keyframes extracted can be determined based on the length of the recording time, such as 30 or 60 frames. The extracted keyframes should cover the subject's entire set of movements. For example, if one movement is from standing upright to bending forward 90 degrees, the acquired keyframes should include images of the subject standing upright and images of the subject bending forward at the end.

[0049] like Figure 2As shown, it demonstrates an actual shooting scenario. The data acquisition equipment can be a common camera or camcorder. To accommodate subjects of different heights, multiple data acquisition devices can be set up for shooting, such as... Figure 2 Three acquisition devices are set up from top to bottom to take pictures simultaneously. In addition, acquisition device 4 is set up on the side of the subject to take pictures to confirm whether the subject's movements are accurate, so as to ensure the effectiveness of the test.

[0050] Since there is more than one acquisition device, multiple videos will be obtained. In this embodiment, a recognition image will be acquired for each video, and a corresponding number of recognition image groups will be obtained according to the number of videos. These recognition images will be used for subsequent recognition operations.

[0051] Step S200: Obtain the edge contour of the subject based on each recognition image, and standardize the output of each recognition image based on the edge contour to obtain a standardized image.

[0052] Before performing the recognition operation, these recognition images need to be preprocessed to obtain standardized images.

[0053] It is understandable that, due to limitations such as shooting location, the final image may have issues such as inappropriate distances and inconsistent sizes of people. Therefore, it is necessary to preprocess each recognition image to standardize the output.

[0054] Specifically, for recognition images from the same video, the edge contours of the subject, including the head, can be obtained by comparing the different recognition images. These edge contours represent the position of the human body in the image. Obtaining these contours is equivalent to identifying the position of the person in the image. Then, through operations such as cropping, the human body is centered in the image and enlarged. At the same time, the final output ratio of the image is standardized, and the preprocessed recognition image is standardized to produce a standardized image.

[0055] For example, the aspect ratio mentioned above can be output in a ratio such as 3:2 or 2:1. The specific ratio can be determined according to the height of the examinee and the shooting angle.

[0056] Therefore, after the above operations, a series of standardized images will be obtained. These images, obtained from the captured video, can reflect the changes in the subject's back when performing preset actions.

[0057] The specific output image is as follows: Figure 3 As shown in this embodiment, a set of image data is illustrated using 30 images as an example.

[0058] Step S300: Calculate the scoliosis deviation angle in each of the standardized images, and output the corresponding waveform based on the change of the scoliosis deviation angle over time.

[0059] The scoliosis deviation angle is used to represent the angle of curvature of the spine compared to a straight, normal spine, and can indicate to some extent whether the examinee has scoliosis. Specifically, this embodiment needs to address... Figure 3 For each frame of the image, a corresponding scoliosis deviation angle is calculated to continue subsequent recognition operations.

[0060] Specifically, such as Figure 4 As shown, firstly, the shoulder contour of the subject is determined based on the edge contour. It can be understood that because the shoulder contour is connected to the head contour and has abrupt change features, the abrupt change features here can be used for identification and positioning.

[0061] After determining the shoulder contour, a midline can be determined based on the shoulder contour. Using the midline as a reference, two perpendicular lines (auxiliary lines, not drawn) are marked on both sides of the midline according to the subject's shoulder width. The first intersection point 100 and the second intersection point 200 formed by the intersection of the left and right perpendicular lines with the shoulder or the edge contour are determined. The left and right perpendicular lines can be drawn at a distance of one-sixth of the shoulder width from the midline, or at a proportional distance of one-fifth, to obtain the first intersection point 100 and the second intersection point 200.

[0062] The midline can be understood as the location of the spine. The midline is perpendicular to the horizontal line. It is determined by identifying two shoulder points that define the shoulder contour. Figure 4 The first shoulder point 300 and the second shoulder point 400 represent the position of the shoulder. The straight line connecting the two points is a horizontal line. After determining these two points, the midpoint between them can be found. A straight line perpendicular to the line segment connecting the first shoulder point 300 and the second shoulder point 400 is drawn through this midpoint. This straight line is the centerline.

[0063] Calculate the angle between the straight line connecting the first intersection point and the second intersection point and the horizontal line to obtain the scoliosis deviation angle.

[0064] from Figure 4 As can be seen from the diagram, there exists a first line segment connecting the first intersection point 100 and the second intersection point 200, and a horizontal line connecting the first shoulder point 300 and the second shoulder point 400. These two lines, after being extended, will intersect at a certain point, and the angle formed by the intersection is the scoliosis deviation angle required in this step. Specifically, this angle can be obtained by direct measurement through image recognition or by calculation using trigonometric functions.

[0065] Because there are multiple recognition images, the scoliosis deviation angle of the subject in different images can be obtained. Since the recognition images are actually related to time, a waveform diagram can be generated with the frame number as the horizontal axis and the basic scoliosis deviation angle as the vertical axis.

[0066] The specific waveform is as follows: Figure 5 The image shown is a waveform obtained by processing the scoliosis deviation angle of a set of recognition images. The number of frames on the horizontal axis corresponds to time. Therefore, this waveform reflects the change in the scoliosis deviation angle of the subject's back when performing a predetermined action. It can be understood that the scoliosis deviation angle of a person without scoliosis will only fluctuate within a low range, while the scoliosis deviation angle of a subject with scoliosis will fluctuate within a large range. Therefore, this waveform can be used to identify and confirm whether the subject has scoliosis.

[0067] It should be noted that the waveform is actually composed of discrete data. For example, if only 30 images are extracted for recognition, the maximum horizontal axis of the waveform is only 30. Each frame of the recognition image is associated with a scoliosis deviation angle to form 30 coordinates. Based on these 30 coordinates, a waveform is formed as follows: Figure 5 The waveform diagram shown.

[0068] It is understandable that different groups of recognition images will have different waveforms due to different shooting angles. Overlaying these waveforms can form a composite waveform, and the interval formed by the space between multiple waveform curves can also represent the fluctuation range of the scoliosis deviation angle to a certain extent.

[0069] Step S400: Identify and determine the waveform to ascertain whether the subject has scoliosis.

[0070] It is understandable that people with scoliosis and those without scoliosis will have different waveforms and fluctuation ranges. Based on these characteristics, it is possible to make a judgment to determine whether the subject in the image has scoliosis.

[0071] For example, the presence of scoliosis in a subject can be determined by analyzing the peak value of the scoliosis deviation angle in the waveform. Specifically, the peak value of the scoliosis deviation angle in the waveform can be obtained, and it can be determined whether the peak value exceeds a preset angle. If it exceeds the preset angle, the subject is determined to have scoliosis; otherwise, no scoliosis is found.

[0072] The preset angle ranges from 3 to 4 degrees.

[0073] In addition, machine learning can be used for identification. First, videos of normal people and people with scoliosis performing the above-mentioned predetermined actions are obtained. After obtaining waveforms by following the same steps, these waveforms are labeled according to whether the examinee has scoliosis. Then, these data are divided into training set and test set. For the information in the training set, machine learning algorithms such as Support Vector Machine can be used to mathematically model the information, and the ability to distinguish scoliosis patients can be evaluated using Receiver Operating Characteristic curves. Then, the data in the test set is used to input into the model to further optimize the algorithm. The identification model trained in this way can use the waveform obtained in step S300 as input to identify whether the examinee corresponding to the waveform has scoliosis.

[0074] The scoliosis detection method in this embodiment involves capturing video of the subject's back using a camera, extracting a certain number of keyframes, processing each keyframe to calculate the scoliosis deviation angle in each frame, and generating a waveform based on the calculated deviation angle and the number of image frames. This waveform is then used for identification to determine if the subject has scoliosis. As can be seen, this embodiment's technical solution requires no medical knowledge from the operator; the entire process only requires capturing an image, which is then processed by a machine to obtain a conclusion. It also eliminates the need for specialized medical equipment, making the overall detection process simple, efficient, and suitable for various applications.

[0075] Example 2

[0076] like Figure 6 As shown, this application also provides a scoliosis detection device, comprising:

[0077] The shooting module 10 is used to capture back video information of the subject performing a predetermined action through the acquisition device, and select a preset number of key frames from the video information as recognition images;

[0078] Preprocessing module 20 is used to obtain the edge contour of the subject based on each recognition image, and to standardize each recognition image based on the edge contour to obtain a standardized image;

[0079] The calculation module 30 is used to calculate the scoliosis deviation angle in each of the standardized images and output the corresponding waveform based on the change of the scoliosis deviation angle over time.

[0080] The identification module 40 is used to identify and determine the waveform to determine whether the subject has scoliosis.

[0081] This application also provides a computer device, including a processor and a memory, wherein the memory stores a computer program that executes the scoliosis detection method when the computer program is run on the processor.

[0082] This application also provides a readable storage medium storing a computer program that, when run on a processor, executes the scoliosis detection method. The method includes: capturing back video information of a subject performing a predetermined action using a data acquisition device, and selecting a preset number of keyframes from the video information as recognition images; obtaining the edge contour of the subject based on each recognition image, and standardizing each recognition image based on the edge contour to obtain a standardized image; calculating the scoliosis deviation angle in each standardized image, and outputting a corresponding waveform based on the change of the scoliosis deviation angle over time; and identifying and judging the waveform to determine whether the subject exhibits scoliosis.

[0083] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative; for example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that, as an alternative implementation, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0084] In addition, the functional modules or units in the various embodiments of the present invention can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0085] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a smartphone, personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0086] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for detecting scoliosis, characterized in that, include: The device captures video information of the subject's back as they perform a predetermined action, and selects a preset number of keyframes from the video information as recognition images. The edge contour of the subject is obtained from each recognition image, and the edge contour is used to standardize the output of each recognition image to obtain a standardized image. The subject's shoulder contour is determined based on the edge contour, and the midline is determined based on the shoulder contour. Using the midline as a reference, two perpendicular lines are marked on each side of the midline according to the subject's shoulder width. The first and second intersection points formed by the left and right perpendicular lines intersecting with the shoulder or the edge contour are determined respectively. Calculate the angle between the straight line connecting the first intersection point and the second intersection point and the horizontal line to obtain the scoliosis deviation angle, and output the corresponding waveform based on the change of the scoliosis deviation angle over time. The waveform is identified and judged to determine whether the subject has scoliosis. There are at least two acquisition devices; The step of capturing back video information of the subject performing a predetermined action using a data acquisition device, and selecting a preset number of keyframes from the video information as recognition images, includes: Video information from each of the acquisition devices is acquired. For each video information, a preset number of image frames are extracted as recognition images according to the playback order of the video. The recognition images extracted from the same video information are grouped together to obtain multiple groups of recognition images.

2. The scoliosis detection method according to claim 1, characterized in that, The step of identifying and judging the waveform to determine whether the subject has scoliosis includes: The peak value of the scoliosis deviation angle in the waveform is obtained, and it is determined whether the peak value exceeds a preset angle. If it exceeds the preset angle, the subject is determined to have scoliosis; otherwise, there is no scoliosis. The preset angle ranges from 3 to 4 degrees.

3. The scoliosis detection method according to claim 1, characterized in that, The step of identifying and judging the waveform to determine whether the subject has scoliosis includes: The waveform is input into a pre-trained recognition model, which then determines whether the subject has scoliosis.

4. The scoliosis detection method according to claim 3, characterized in that, The training method for the recognition model includes: Acquire back video information of multiple groups of scoliosis patients when performing the preset movements, and back video information of normal people when performing the preset movements; The scoliosis deviation angle in each action video is obtained and a corresponding waveform is generated. Each waveform is labeled and used as training data to be input into a selected learning model for training to obtain the recognition model.

5. The scoliosis detection method according to claim 1, characterized in that, The step of obtaining the edge contour of the subject in each recognition image and standardizing the output of each recognition image based on the contour includes: The various recognition images are compared to obtain the edge contours of the subject in each recognition image. Based on the edge contours, each recognition image is centered and cropped, and then output as a standardized image with the same preset ratio.

6. A scoliosis detection device, characterized in that, include: The shooting module is used to capture video information of the back of the subject when performing a predetermined action through the acquisition device, and select a preset number of key frames from the video information as recognition images; The preprocessing module is used to obtain the edge contour of the subject based on each recognition image, and to standardize the output of each recognition image based on the edge contour to obtain a standardized image. The calculation module is used to determine the shoulder contour of the examinee based on the edge contour, determine the midline based on the shoulder contour, mark two perpendicular lines on the left and right sides of the midline based on the midline and the width of the examinee's shoulders, and determine the first intersection point and the second intersection point formed by the left and right perpendicular lines intersecting with the shoulder or the edge contour, respectively. Calculate the angle between the straight line connecting the first intersection point and the second intersection point and the horizontal line to obtain the scoliosis deviation angle, and output the corresponding waveform based on the change of the scoliosis deviation angle over time. The identification module is used to identify and judge the waveform to determine whether the subject has scoliosis. There are at least two acquisition devices; The step of capturing back video information of the subject performing a predetermined action using a data acquisition device, and selecting a preset number of keyframes from the video information as recognition images, includes: Video information from each of the acquisition devices is acquired. For each video information, a preset number of image frames are extracted as recognition images according to the playback order of the video. The recognition images extracted from the same video information are grouped together to obtain multiple groups of recognition images.

7. A computer device, characterized in that, It includes a processor and a memory, the memory storing a computer program that, when run on the processor, executes the scoliosis detection method according to any one of claims 1 to 5.

8. A readable storage medium, characterized in that, It stores a computer program that, when run on a processor, executes the scoliosis detection method according to any one of claims 1 to 5.

Images