A cloud computing-based scoliosis screening method and system

Abstract

A kind of cloud computing-based scoliosis screening method and system, human back image is acquired using depth camera, depth image is obtained by edge control host pre-processing, and continuous depth value area is taken as human back key area;Depth data of human back area obtained is converted into spatial three-dimensional coordinate point, and three-dimensional coordinate point is combined to form point cloud data;Point cloud data is sent to cloud computing center, and cloud computing center calculates the ATR angle of each vertebra of spine according to the point cloud data of human back;Cloud computing center sends the ATR angle data of each vertebra position calculated back to edge control host, for the screening of scoliosis.The method of the application not only reduces the dependence on local device performance, but also improves the efficiency and accuracy of data processing through centralized processing of cloud, while reducing the overall cost.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of data processing, and relates to a method and system for screening scoliosis based on cloud computing. BACKGROUND

[0002] Scoliosis refers to the lateral bending of the spine in addition to the normal anterior and posterior bending. This condition can cause body posture asymmetry and, in severe cases, can affect heart and lung function. Scoliosis screening is a method to help doctors identify potential scoliosis cases from a large population. The purpose of screening is to detect scoliosis as early as possible so that timely treatment can be provided to prevent the condition from worsening. Scoliosis screening is usually targeted at children and adolescents, especially during the pre-pubertal and post-pubertal periods, as this is the period of fastest spinal growth and the fastest development of scoliosis.

[0003] Traditional scoliosis screening devices usually collect the physical data of the subject through image acquisition devices to generate point cloud images. These devices process the data locally and analyze the point cloud through built-in algorithms to identify the shape and position of the spine for further judgment by doctors. The screening process includes comparing the natural curvature of the patient's spine with the standard model and detecting any abnormal lateral bending. If the bending angle exceeds a certain threshold, it is usually considered a sign of scoliosis. This screening method has the following problems:

[0004] Firstly, the point cloud data processing and scoliosis screening algorithm have high requirements for computing resources, which requires the screening device to be equipped with high-performance hardware, thereby increasing the cost of the device. This is a limiting factor for large-scale screening activities, as it limits the popularization and application of screening technology. Secondly, once the screening device is sold, the updating and maintenance of the algorithm program become complex. This makes it difficult for the device to update to the latest screening algorithm in a timely manner, affecting the accuracy and reliability of the screening results. Finally, due to the lack of timely algorithm updates, there may be differences in screening results on different devices, which is not conducive to the uniformity of screening standards and the improvement of screening quality.

[0005] Cloud computing is a new data computing model this year, but data transmission efficiency is still one of the problems to be solved. For scoliosis screening, how to quickly transmit a large amount of basic data collected on site to the cloud computing and obtain accurate results needs to be studied. In addition, existing point cloud analysis and calculation methods still have deficiencies in how to accurately calculate and analyze the spine state from a large amount of point cloud data. Due to individual differences in detection, there are still some errors in the existing ATR angle calculation method. SUMMARY

[0006] The technical problem to be solved by the present application is that the existing scoliosis screening scheme relies on local equipment when detecting spinal data, has high requirements for the equipment, and it is difficult to conduct large-scale screening of scoliosis.

[0007] The technical scheme of the present application is a scoliosis screening method based on cloud computing, comprising the following steps:

[0008] Step 1, use a depth camera to collect a human back image, continuously capture depth image data by an edge control host, and perform filtering preprocessing;

[0009] Step 2, the edge control host obtains a region with continuous depth values in the depth image after preprocessing, which has a depth value range of 100mm-2000mm, and takes this region as a key region of the human back;

[0010] Step 3, the edge control host converts the depth data of the human back region obtained, converts the depth image data into spatial three-dimensional coordinate points according to the focal length and optical center parameters of the depth camera, and the three-dimensional coordinate points are combined to form point cloud data;

[0011] Step 4, send the point cloud data to a cloud computing center, and the cloud computing center calculates the ATR angle of each vertebra of the spine according to the point cloud data of the human back;

[0012] Step 5, the cloud computing center sends the calculated ATR angle data of each vertebra position back to the edge control host for scoliosis screening.

[0013] The present application also provides a scoliosis screening system based on cloud computing, which is used to implement the above-mentioned scoliosis screening method, comprising:

[0014] A data acquisition module is used to acquire human back depth image data;

[0015] A data analysis module is used to obtain a region with continuous depth values in a depth image within a preset depth value range, and take this region as a key region of the human back;

[0016] A data conversion module is used to convert the depth data of the human back region to form human back point cloud data;

[0017] A calculation module is used to calculate the ATR angle of each vertebra of the spine according to the human back point cloud data.

[0018] The cloud computing center in the application can be configured with a CPU computing resource with strong computing power, and can use a multi-core CPU for parallel computing, thereby significantly reducing the computing time. Through cloud computing, resource sharing of point cloud data calculation can be realized, and the cost is reduced. The algorithm is concentrated in the cloud, which is convenient for upgrading and optimizing the algorithm and deployment. At the same time, the centralized management of the cloud is also convenient for unified updating and maintenance of the algorithm, so that all devices can use the latest screening algorithm, and the accuracy and consistency of the screening result are improved. When the cloud computing is used, the amount of data to be transmitted is reduced through the preprocessing of the on-site detection data, and the effective calculation data is reserved for the later calculation, and the accurate angle information is obtained through the designed ATR angle calculation method in the cloud, which is effectively used for scoliosis screening. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 The flowchart of the method of the application.

[0020] Figure 2 The human back key area mask image example in the method of the application.

[0021] Figure 3 The schematic diagram of the system of the application. DETAILED DESCRIPTION

[0022] The application provides a scoliosis screening method based on cloud computing, as shown in the figure, comprising the following steps. Figure 1

[0023] Step 1, using the human back image collected by the depth camera, the edge control host continuously captures the depth image data and pre-processes it.

[0024] (1) The depth camera is a 3D depth camera, which is electrically connected with the edge control host. Adam's forward bend test (Adam's Forward Bend Test, Adam test) is a simple and practical examination method for scoliosis in clinical practice. The depth camera collects the human back image in Adam's position, and transmits the back image to the edge control host to obtain the depth image of H*W, wherein H represents the height of the image, that is, the number of pixels in the vertical direction, and W represents the width of the image, that is, the number of pixels in the horizontal direction. Each pixel point of the depth image has a depth value, and the depth value corresponds to the distance from the pixel point in the camera field of view to the camera sensor plane.

[0025] (2) The edge control host continuously obtains 10-20 frames of depth image data, and performs average value filtering processing on the depth image data, thereby obtaining the first step filtered image data frame D, and then performing the second step filtering, calculating the depth average value d'(x,y) at each pixel point in the image data frame D, and obtaining the value through the following formula: ​

[0026]

[0027] wherein d(x, y) is the depth value of the position (x, y) in the image data frame D, Similarly, n x m is the window area, the depth average value is calculated by defining a window range, m and n are the side length of the window with (x, y) as the center point, in pixels, m and n are 3-7 by experience value; i represents the index of the window in the x-axis direction, the value range is from 1 to n; j represents the index of the window in the y-axis direction, the value range is from 1 to m. The depth average value is taken as the depth value of the pixel position after filtering, and the preprocessed depth image is obtained.

[0028] By two-step filtering processing on the depth image data, the noise in the depth image can be reduced, the depth image is more smooth and stable, and the stability and accuracy of the data are improved.

[0029] Step 2, the edge control host acquires the depth image with the depth value range of 100mm-2000mm as a continuous depth value region, and takes the region as a key region of the back of the human body, the method comprising:

[0030] (1) Shift the depth image by n pixel points, the value range of n is 1-10, and the embodiment takes 2 pixel points. Subtract the shifted depth image from the depth image before shifting to generate a difference image of the depth image, the value of each pixel point in the difference image represents the depth difference of the corresponding position in the two images, and the difference image is binarized:

[0031] Check the value of each pixel point in the difference image, if the value of the pixel point is less than or equal to 10, set the value of the pixel point to 0, if greater than 10, set to 255.

[0032] (2) In OpenCV, use the center point filling algorithm function floodFill to fill the processed difference image, and perform difference operation again on the filled and unfilled images to obtain a mask image, the calculation formula of the difference operation is:

[0033] F(p, q) = |I before (p, q) - I after (p, q)|

[0034] Wherein I before (p, q) is the pixel value of the image before filling at position (p, q), I after (p, q) is the pixel value of the image at (p, q) after filling, F(p, q) is the pixel value of the difference image at position (p, q).

[0035] The mask image is obtained by comparing the pixel value of the difference image with a set threshold value, setting the pixel value less than or equal to the threshold value to 0, and setting the pixel value greater than the threshold value to 255, and the calculation formula is:

[0036]

[0037] wherein M(p,q) is the pixel value of the mask image at position (p,q), T is the set threshold value, and T is usually 10, indicating that only when the pixel value of the difference image is greater than 10, the corresponding pixel value in the mask image is set to 255, otherwise 0.

[0038] (3) The effective area on the mask image, i.e. the area composed of pixel points with pixel value 255, is defined as an area with continuous depth value, and is taken as the human back area, as shown in Figure 2 .

[0039] Step 3: The edge control host converts the depth data of the human back area obtained to three-dimensional coordinate points, and converts the two-dimensional depth image data into three-dimensional coordinate points according to the current camera focal length and optical center parameters, and the three-dimensional coordinate points are combined to form point cloud data, and the formula is:

[0040]

[0041] Z=Z

[0042] wherein (u,v) is the pixel coordinate in the depth image, Z is the corresponding depth value, (cx,cy) is the optical center coordinate, and (fx,fy) is the focal length.

[0043] Step 4: The point cloud data is sent to the cloud computing center, the cloud computing center restores the point cloud data to obtain the human back point cloud data, and calculates the ATR angle of each vertebra of the spine according to the human back point cloud data.

[0044] The edge control host packages and compresses the point cloud data of the human back area obtained after preprocessing into a compressed file; the edge control host sends the compressed file to the cloud computing center through the Internet.

[0045] The cloud computing center performs slice processing on the point cloud data according to the number of vertebrae, obtains the start and end indexes of each vertebra, and each vertebra is thoracic vertebra and lumbar vertebra, which is divided from top to bottom, and there are 12 thoracic vertebrae and 5 lumbar vertebrae.

[0046] In the three-dimensional coordinate points of the point cloud data, the x-axis direction corresponds to the left-right direction of the back, the y-axis direction corresponds to the front-back direction of the back in the bending state, and the z-axis direction corresponds to the depth direction. When the depth camera collects the image of the human back, the human body is in a bending state, and the camera collects the back image directly above the human body. Therefore, the z-axis also represents the depth size of the back in the pixel position distance from the camera plane. The ATR angle is calculated from the point cloud data as follows.

[0047] Mean filtering in the x-axis direction: for each point cloud data and its left and right sides in the x-axis direction, n is 2-5, and 3 points are taken in this embodiment, that is, the z coordinate values of 2n+1 points including itself are used for mean calculation, and the calculation result is used as the new z coordinate value of the point.

[0048] Mean filtering in the y-axis direction: the back data between the indexes is obtained through the y value start and end indexes of each vertebra, that is, the data of the current vertebra slice, and then a one-dimensional slice data is obtained by mean processing of the data of the current vertebra slice in the y-axis direction. 17 vertebrae get 17 slices.

[0049] The mean filtering is used to eliminate the influence of individual screening on the accuracy of the calculation result.

[0050] The one-dimensional slice data is subjected to curve fitting of a function to obtain a function y=f(x), and the bending point of the curve is analyzed through the function.

[0051] The bending point is analyzed to obtain the spinal center point of the vertebra, and the spinal center point coordinate of the vertebra is obtained.

[0052] The points on the vertebra slice within the horizontal distance of 22 mm on the left and right of the spinal center point coordinate of the vertebra are removed, and the remaining points on the slice after removal are used as a set M. The length of the scoliosis measuring ruler is 22 mm, which corresponds to the value taken here. Different A points and B points are selected from the set M, the A point is higher than the B point, and the high here also corresponds to the high in the physical world, that is, the physical height of the A point is higher than that of the B point. Because it is in the bending state, the depth camera is above the human body to shoot the image. If the depth value of the depth camera is used to describe it, the depth value (z value) of the A point is less than that of the B point. The function of the straight line l on which the two points are located is calculated. All points in M are judged using the straight line function to see if there is a point C higher than the straight line l. If a point C higher than the straight line l is found, the C point is used to update and replace the original lower B point, and the function of the straight line l is recalculated. Until all points in the set M are below the straight line l.

[0053] The coordinates of the last updated A point and B point are (x1, y, z1) and (x2, y, z2) respectively, and the ATR angle of the current vertebra is calculated, and the calculation formula is:

[0054]

[0055] Wherein, (x1, y, z1) is the coordinate of point A, and (x2, y, z2) is the coordinate of point B.

[0056] Step 5, the cloud computing center packs and compresses the calculated ATR angle data of each vertebral position into a file, and sends it back to the edge control host. The edge control host is connected with the display terminal, and the results are presented to the display terminal to view the ATR angle of each vertebral position. The ATR angle range of the first scoliosis vertebral position is 5-7°; the ATR angle range of the second scoliosis vertebral position is 7-10°; and the ATR angle range of the third scoliosis vertebral position is more than 10°. This angle can help doctors assess the scoliosis of the spine and decide whether further diagnosis is needed.

[0057] The application also provides a cloud computing-based scoliosis screening system, as shown in Figure 3 The system is used to implement the above-mentioned scoliosis screening method and comprises:

[0058] A data acquisition module is configured to acquire the depth image data of the human back;

[0059] A data analysis module is configured to obtain a continuous depth value region in the depth image within a preset depth value range, and take the region as a key region of the human back;

[0060] A data conversion module is configured to convert the depth data of the human back region to form point cloud data of the human back;

[0061] A calculation module is configured to calculate the ATR angle of each vertebral position of the spine according to the point cloud data of the human back.

[0062] The computer program code for executing the above-mentioned scoliosis screening method can be written in one or more programming languages or a combination thereof. The data acquisition module is a depth camera, and the data analysis module and the data conversion module are implemented in the edge control host. The edge control host comprises a processing device such as a central processing unit, a graphics processing unit, etc. It can perform various appropriate actions and processes according to the computer program stored in the computer readable storage medium, which can be included in the above-mentioned edge control host, or can exist separately without being assembled into the edge control host. The calculation module is a cloud server, which can communicate with the edge control host by using any currently known or future developed network protocol such as HTTP (HyperText Transfer Protocol, HyperText Transfer Protocol).

[0063] The above description is merely exemplary of the disclosure and the application of the principles thereof and it is not intended to limit the scope of the disclosure to the specific forms set forth. The disclosure is susceptible to numerous modifications and alternative forms of implementation, all of which are intended to be within the scope of the disclosure. Indeed, the specific forms set forth are only exemplary embodiments of the disclosure and other embodiments are possible. The scope of the disclosure is therefore not intended to be limited to the specific forms set forth and any modifications and / or improvements are intended to be within the scope of the disclosure.

Claims

1. A cloud-based scoliosis screening method, characterized by: Includes the following steps: Step 1: Use a depth camera to acquire images of the human back. The edge control host continuously captures depth image data and performs filtering preprocessing. Step 2: In the pre-processed depth image, the edge control host obtains a region with a continuous depth value range of 100mm to 2000mm and uses this region as the key area of ​​the human back. Step 3: The edge control host converts the depth data of the acquired human back area. Based on the focal length and optical center parameters of the depth camera, the depth image data is converted into spatial three-dimensional coordinate points. The three-dimensional coordinate points are then collected to form point cloud data. Step 4: Send the point cloud data to the cloud computing center. The cloud computing center calculates the ATR angle of each vertebra of the spine based on the point cloud data of the human back. The cloud computing center slices the point cloud data according to the number of vertebral segments to obtain the start and end indices of each segment. These segments include all segments of the thoracic and lumbar vertebrae. The calculation of the ATR angle includes: In the three-dimensional coordinate points of point cloud data, the x-axis corresponds to the left and right direction of the back, the y-axis corresponds to the front and back direction of the back when bending over, and the z-axis corresponds to the depth direction. Mean filtering is performed in the x-axis direction: For each point cloud data and the n points on its left and right sides in the x-axis direction, where n takes the value of 2 to 5, that is, the z coordinate values ​​of the 2n+1 points including itself are averaged, and the result is used as the new z coordinate value of the point. Mean filtering in the y-axis direction: By using the start and end indices of the y-values ​​of each vertebral segment, the back data between the indices is obtained, which is the data of the current vertebral segment slice. Then, by averaging the data of the current vertebral segment slice in the y-axis direction, a one-dimensional slice data is obtained. Curve fitting of a function to one-dimensional slice data yields the function. The bending points of the curve are analyzed using functions; By analyzing the curvature points, the center point of the spine can be determined, and the coordinates of the center point of the spine at that vertebral segment can be obtained. Using the coordinates of the spinal center point of each vertebral segment, remove points on the vertebral slices within a horizontal distance of 22mm to the left and right. The remaining points on the slices are set M. Randomly select different points A and B from this set, where point A is higher than point B (i.e., the depth value of point A is less than that of point B). Calculate the function of the line l containing the two points. Use the line function to judge all points in M ​​to see if there are any points higher than line l. If a point C higher than line l is found, replace point B with point C and recalculate the function of line l until all points in set M are below line l. The coordinates of point A and the finally updated point B are (x1, y, z1) and (x2, y, z2) respectively. The ATR angle of the current vertebral segment is calculated using the following formula: ； Step 5: The cloud computing center sends the calculated ATR angle data of each vertebral segment back to the edge control host for scoliosis screening.

2. The scoliosis screening method based on cloud computing according to claim 1, Its characteristic is that in step 1, the depth camera acquires an image of the human back in Adam's pose, the edge control host obtains depth image data based on the acquired image, continuously acquires 10 to 20 frames of depth image data, performs average filtering processing to obtain the first-step filtered image data frame D, and then performs the second-step filtering to calculate the average depth at each pixel in image data frame D. It can be obtained through the following formula: in, It is the pixel in image data frame D The depth value of the location; It is the window area, m and n are in terms of The side length of the window centered at the specified point is in pixels, and the values ​​of m and n range from 3 to 7. Indicates the window is in The index along the axis ranges from 1 to n; Indicates the window is in The index along the axis ranges from 1 to m; The average depth value is used as the filtered depth value for that pixel location to obtain the preprocessed depth image.

3. The scoliosis screening method based on cloud computing according to claim 1, characterized in that: Step 2 specifically includes: 1) Offset the depth image by n pixels, where n ranges from 1 to 10. Subtract the offset depth image from the original depth image to generate a difference image. The value of each pixel in the difference image represents the depth difference between the two images at the corresponding position. Binarize the difference image by checking the value of each pixel. If the value of a pixel is less than or equal to 10, set the value of the pixel to 0. If it is greater than 10, set it to 255. 2) The OpenCV center-fill algorithm function `floodFill` is used to fill the processed difference image. The images before and after filling are then differenced again to obtain a mask image. The formula for calculating the difference operation is: in, The image before filling is in position pixel values, The image after filling is in pixel values, The position after the difference operation Pixel values; The mask image is obtained by comparing pixel values ​​obtained through a difference operation with a set threshold. Pixel values ​​less than or equal to the threshold are set to 0, and pixel values ​​greater than the threshold are set to 255. The calculation formula is as follows: in, It is the mask image at the position The pixel value, where T is the set threshold; 3) Define the effective area on the mask image, that is, the area composed of pixels with a pixel value of 255, as the area with continuous depth values, as the human back area.

4. The scoliosis screening method based on cloud computing according to claim 1, characterized in that, in step 3, the conversion of two-dimensional depth image data into three-dimensional coordinate points specifically involves: in, Z represents the pixel coordinates in the depth image, Z is the corresponding depth value, and (X,Y,Z) is the 3D coordinate point. These are the coordinates of the optical center. It is the focal length.

5. A cloud-based scoliosis screening system, characterized in that: To implement the scoliosis screening method according to any one of claims 1-4, it includes: The data acquisition module is used to acquire depth image data of the human back. The data analysis module is used to obtain a continuous depth value region from a depth image within a preset depth value range, and to use this region as the key region of the human back. The data conversion module is used to convert the depth data of the human back region into point cloud data of the human back. The calculation module is used to calculate the ATR angle of each vertebral segment of the spine based on the point cloud data of the human back.

Images