Auxiliary analysis system and method for bone monitoring data

Abstract

The invention discloses a bone monitoring data auxiliary analysis system and method, and the system comprises an obtaining module which is used for obtaining bone monitoring data; the bone monitoring data comprises a DXA image and a CT image; the first extraction module is used for performing feature extraction on the DXA image to obtain a first feature extraction result; the second extraction module is used for performing feature extraction on the CT image to obtain a second feature extraction result; and the determination module is used for determining an auxiliary analysis suggestion according to the first feature extraction result and the second feature extraction result. Comprehensive and accurate bone monitoring data are obtained, data analysis is carried out on the bone monitoring data, and accurate auxiliary analysis suggestions are obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a bone quality monitoring data auxiliary analysis system and method. Background Art

[0002] Osteoporosis is a bone disease characterized by decreased bone mass and deterioration of bone microstructure. Osteoporosis patients are prone to brittle (low-traumatic) fractures, which lead to pain, decreased function, increased hospitalization rate, impaired quality of life, significantly increased disability and mortality rates, and a huge burden on families and society. Early screening and early prevention and treatment are of great significance in preventing fractures.

[0003] At present, the BMD value measured by DXA is the gold standard for diagnosing osteoporosis. In fact, BMD can only explain 70% to 75% of the changes in bone strength. The remaining factors affecting bone strength include the accumulation of microfractures, changes in bone microstructure, bone remodeling disorders, bone mineralization, etc. BMD can only evaluate bone mass and cannot directly reflect changes in bone microstructure. In addition, there is a partial overlap in BMD values ​​between fracture and non-fracture patients, indicating that relying solely on BMD to diagnose osteoporosis, especially relying solely on BMD to exclude osteoporosis, is not comprehensive. The bone monitoring data obtained in the existing technology is not accurate or comprehensive, and there are also deviations in determining auxiliary analysis recommendations. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems in the above-mentioned technology to a certain extent. To this end, the purpose of the present invention is to propose a bone monitoring data auxiliary analysis system and method to obtain comprehensive and accurate bone monitoring data, and to perform data analysis on the bone monitoring data to obtain accurate auxiliary analysis suggestions.

[0005] To achieve the above objectives, the embodiment of the present invention proposes a bone quality monitoring data auxiliary analysis system, comprising:

[0006] An acquisition module is used to acquire bone monitoring data; the bone monitoring data includes DXA images and CT images;

[0007] A first extraction module, used for performing feature extraction on the DXA image to obtain a first feature extraction result;

[0008] The second extraction module is used to extract features from the CT image to obtain a second feature extraction result;

[0009] The determination module is used to determine auxiliary analysis suggestions based on the first feature extraction result and the second feature extraction result.

[0010] According to some embodiments of the present invention, the acquisition module includes:

[0011] DXA scanners are used to irradiate patients in a static state with X-rays. After the scan is completed, DXA images are generated;

[0012] The CT scanner is used to perform cross-sectional scanning on a patient using X-rays to obtain cross-sectional images of bones as CT images.

[0013] According to some embodiments of the present invention, the first extraction module includes:

[0014] A preprocessing module, used for performing image preprocessing on the DXA image to obtain a preprocessed image;

[0015] The generation module is used to extract features from the preprocessed image based on the SBCFormer network to obtain a first feature extraction result.

[0016] According to some embodiments of the present invention, the preprocessing module includes:

[0017] A smoothing processing module is used for smoothing the DXA image to obtain a smoothed image;

[0018] A segmentation module is used to perform image segmentation based on the Otsu method, taking the gray value with the minimum variance within the smoothed image and the maximum variance between classes as the target threshold, and obtain a segmented image with a bone tissue area according to the target threshold;

[0019] The contour marking module is used to perform edge detection on the segmented image based on the edge detection operator, and mark the detected contour points to obtain a preprocessed image.

[0020] According to some embodiments of the present invention, the smoothing processing module performs smoothing processing on the DXA image to obtain a smoothed image, including:

[0021]

[0022] Among them, G(x,y) is the data at (x,y) in the smoothed image; x is the distance from the origin on the horizontal axis; y is the distance from the origin on the vertical axis; e is a natural constant; σ is the standard deviation of the Gaussian distribution, and σ is a variable value between 0 and 3.

[0023] According to some embodiments of the present invention, the generating module comprises:

[0024] The propagation module is used to forward propagate the preprocessed image based on the convolutional layer, pooling layer, and fully connected layer structure in the SBCFormer network;

[0025] The first calculation module is used to calculate the first correlation between different channel features in the SBCFormer network and the second correlation between features at different positions in the image during the forward propagation process; optimize the features in the forward propagation process according to the first correlation and the second correlation to obtain the first feature extraction result.

[0026] According to some embodiments of the present invention, the second extraction module includes:

[0027] The first analysis module is used to:

[0028] Based on the preset window size and step length, a rectangular window is slid on the CT image; based on the structural similarity algorithm, the similarity between the central area of ​​the rectangular window and the adjacent area is calculated; the center distance between the central area and the adjacent area is calculated, and the weight coefficient of the adjacent area and the central area is set according to the ratio of the preset standard distance to the center distance;

[0029] According to the similarity between the adjacent areas and the central area and the corresponding weight coefficient, the position feature value of the central area of ​​the rectangular window is determined;

[0030] The second analysis module is used to:

[0031] Apply the principal component analysis algorithm to the CT image, reduce the dimension of the CT image to the feature dimension space, project the central area of ​​all rectangular windows onto the feature dimension space, and determine the dimensional feature value of the central area of ​​the rectangular window;

[0032] A central area whose position feature value is greater than a preset position feature value and whose dimension feature value is greater than a preset dimension feature value is screened out as a second feature extraction result.

[0033] According to some embodiments of the present invention, the determining module includes:

[0034] A comparison module, used to determine TBS data according to the first feature extraction result, and compare the TBS data with the TBS threshold to obtain a comparison result;

[0035] Matching modules for:

[0036] Determine a color change matrix of the central area according to the second feature extraction result, and match the color change matrix with a preset color change matrix in a preset database to obtain a matching result;

[0037] Auxiliary analysis suggestions are determined based on the comparison results and the matching results.

[0038] According to some embodiments of the present invention, the matching module includes:

[0039] A transformation module, used for transforming the central area in the second feature extraction result into an N*M image matrix of pixels; N is the length of the image matrix; M is the width of the image matrix;

[0040] The first calculation module is used to obtain the value of each pixel in the R channel, the G channel and the B channel, and calculate the color feature value according to the value of each pixel in the R channel, the G channel and the B channel to obtain the color change matrix;

[0041]

[0042] Among them, W λ is the color feature value of the λth pixel in the color change matrix; S i is the optimization parameter of the i-th channel in the image matrix Q; Q i,λ is the value of the i-th channel of the λ-th pixel of the image matrix Q; Q j,λ is the value of the j-th channel of the λ-th pixel of the image matrix Q.

[0043] According to some embodiments of the present invention, an analysis method of a bone monitoring data auxiliary analysis system includes:

[0044] Obtaining bone monitoring data; the bone monitoring data includes DXA images and CT images;

[0045] Performing feature extraction on the DXA image to obtain a first feature extraction result;

[0046] Performing feature extraction on the CT image to obtain a second feature extraction result;

[0047] An auxiliary analysis suggestion is determined according to the first feature extraction result and the second feature extraction result.

[0048] The present invention proposes a bone monitoring data auxiliary analysis system and method, which can more comprehensively evaluate the health status of bones by combining DXA and CT image data. It realizes automated bone monitoring data analysis and improves diagnostic efficiency and accuracy. It provides personalized auxiliary analysis suggestions based on the specific conditions of the patient, which helps to formulate a more accurate treatment plan.

[0049] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0050] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0052] Figure 1 is a block diagram of a bone quality monitoring data auxiliary analysis system according to an embodiment of the present invention;

[0053] Figure 2 is a block diagram of an acquisition module according to an embodiment of the present invention;

[0054] Figure 3 The present invention is a flowchart of a bone monitoring data auxiliary analysis method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0055] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0056] like Figure 1 As shown, the embodiment of the present invention proposes a bone monitoring data auxiliary analysis system, including:

[0057] An acquisition module is used to acquire bone monitoring data; the bone monitoring data includes DXA images and CT images;

[0058] A first extraction module, used for performing feature extraction on the DXA image to obtain a first feature extraction result;

[0059] The second extraction module is used to extract features from the CT image to obtain a second feature extraction result;

[0060] The determination module is used to determine auxiliary analysis suggestions based on the first feature extraction result and the second feature extraction result.

[0061] The working principle of the above technical solution is: feature extraction is performed on DXA images based on the SBCFormer network to determine features that can effectively evaluate the microstructure of bones and describe the quality of bones. Feature extraction is performed on CT images to determine areas containing detailed features, for example, including bone structure, joints, cortical bone, cancellous bone, and possible lesion areas. Based on the first feature extraction results and the second feature extraction results, the health status of bones is comprehensively evaluated, and auxiliary analysis suggestions are provided, including the severity of osteoporosis, fracture risk, treatment suggestions, etc.

[0062] The beneficial effects of the above technical solution: Combining the data of DXA and CT images can more comprehensively evaluate the health status of bones. Automated bone monitoring data analysis is achieved to improve diagnostic efficiency and accuracy. According to the specific situation of the patient, personalized auxiliary analysis suggestions are provided to help formulate more accurate treatment plans.

[0063] like Figure 2 As shown, according to some embodiments of the present invention, the acquisition module includes:

[0064] DXA scanners are used to irradiate patients in a static state with X-rays. After the scan is completed, DXA images are generated;

[0065] The CT scanner is used to perform cross-sectional scanning on a patient using X-rays to obtain cross-sectional images of bones as CT images.

[0066] The working principle of the above technical solution is: DXA scanner uses X-rays to irradiate patients in a static state, uses dual-energy X-ray technology to measure bone density, and generates DXA images. CT scanner uses X-rays to perform cross-sectional scans on patients to obtain bone cross-sectional images, namely CT images. CT images can clearly show the subtle structure and pathological changes of bones, providing doctors with more comprehensive bone information.

[0067] Beneficial effects of the above technical solution: The acquisition module is an important component of the bone monitoring data auxiliary analysis system. It acquires the patient's bone monitoring data through a DXA scanner and a CT scanner, making it easy to obtain comprehensive and accurate data.

[0068] According to some embodiments of the present invention, the first extraction module includes:

[0069] A preprocessing module, used for performing image preprocessing on the DXA image to obtain a preprocessed image;

[0070] The generation module is used to extract features from the preprocessed image based on the SBCFormer network to obtain a first feature extraction result.

[0071] The working principle of the above technical solution is: DXA images are preprocessed to improve image quality, reduce noise and interference, and achieve contours for easy marking. The SBCFormer network is used to process DXA images and extract features related to bone health. Through the SBCFormer network, feature information related to bone density, bone structure, osteoporosis degree, etc. can be extracted from the preprocessed DXA images.

[0072] Beneficial effects of the above technical solution: The first extraction module extracts feature information related to bone health from the DXA image through image preprocessing and feature extraction steps.

[0073] According to some embodiments of the present invention, the preprocessing module includes:

[0074] A smoothing processing module is used for smoothing the DXA image to obtain a smoothed image;

[0075] A segmentation module is used to perform image segmentation based on the Otsu method, taking the gray value with the minimum variance within the smoothed image and the maximum variance between classes as the target threshold, and obtain a segmented image with a bone tissue area according to the target threshold;

[0076] The contour marking module is used to perform edge detection on the segmented image based on the edge detection operator, and mark the detected contour points to obtain a preprocessed image.

[0077] The working principle of the above technical solution is as follows: the smoothing module performs smoothing on the DXA image to reduce noise and interference in the image and improve the signal-to-noise ratio and clarity of the image. The segmentation module performs image segmentation on the smoothed image based on the Otsu method to extract the bone tissue area. The Otsu method is an adaptive threshold segmentation method that automatically selects a threshold value based on the grayscale histogram of the image so that the segmented image has the smallest intra-class variance and the largest inter-class variance. This threshold is used as the target threshold to segment the image into the foreground (bone tissue area) and the background. After segmentation by the Otsu method, a segmented image with a bone tissue area is obtained. This image only contains the bone tissue area and the background, and other irrelevant parts are removed. The detected edge points (i.e., contour points) are marked to obtain a preprocessed image. This image only contains the contour information of the bone tissue area, providing a more accurate basis for subsequent feature extraction.

[0078] Beneficial effects of the above technical solution: The preprocessing module preprocesses the DXA image through steps such as smoothing, image segmentation and contour marking, providing a clearer and more accurate basis for subsequent feature extraction.

[0079] According to some embodiments of the present invention, the smoothing processing module performs smoothing processing on the DXA image to obtain a smoothed image, including:

[0080]

[0081] Among them, G(x,y) is the data at (x,y) in the smoothed image; x is the distance from the origin on the horizontal axis; y is the distance from the origin on the vertical axis; e is a natural constant; σ is the standard deviation of the Gaussian distribution, and σ is a variable value between 0 and 3.

[0082] The working principle of the above technical solution is: σ is the standard deviation of the Gaussian distribution, which determines the degree of smoothing. The larger the value of σ, the more obvious the smoothing effect, and the more details and noise in the image are removed; conversely, the smaller the value of σ, the weaker the smoothing effect, and the more details and noise in the image are retained. σ is a variable value between 0 and 3. In practical applications, the value of σ can be adjusted according to the specific situation of the DXA image and the needs of subsequent analysis to achieve the best smoothing effect.

[0083] The beneficial effect of the above technical solution is that the image quality can be improved by smoothing the DXA image using a Gaussian function.

[0084] According to some embodiments of the present invention, the generating module comprises:

[0085] The propagation module is used to forward propagate the preprocessed image based on the convolutional layer, pooling layer, and fully connected layer structure in the SBCFormer network;

[0086] The first calculation module is used to calculate the first correlation between different channel features in the SBCFormer network and the second correlation between features at different positions in the image during the forward propagation process; optimize the features in the forward propagation process according to the first correlation and the second correlation to obtain the first feature extraction result.

[0087] The working principle of the above technical solution is as follows: the propagation module is responsible for forward propagating the preprocessed image through the convolution layer, pooling layer, fully connected layer and other structures in the SBCFormer network. The preprocessed image is used as input, and the spatial features of the image are extracted through the convolution layer in the network. The resolution of the feature map is reduced and the amount of calculation is reduced through the pooling layer. Finally, the features are mapped to the feature space through the fully connected layer to obtain a preliminary feature representation. In the forward propagation process, the first correlation between the features of different channels in the SBCFormer network and the second correlation between the features at different positions in the image are calculated. The first correlation refers to the correlation or dependency between the features of different channels in the network, which can be achieved by calculating the covariance, correlation coefficient or attention mechanism between the feature maps. Based on the first correlation, it helps to capture the global information of the image. The second correlation refers to the spatial relationship or contextual information between the features at different positions in the image, which can be achieved by calculating the position correlation on the feature map, using the spatial attention mechanism or the self-attention mechanism. Based on the second correlation, it helps to capture the local structure and texture information of the image. According to the calculated first correlation and second correlation, the features in the forward propagation process are optimized by re-weighting the feature map, fusing features of different channels or positions, suppressing irrelevant features, etc., so as to improve the accuracy and robustness of the feature representation.

[0088] The beneficial effect of the above technical solution is that the generation module optimizes the features by using the SBCFormer network for forward propagation and calculating the correlation between different channel and position features, thereby obtaining an accurate and robust first feature extraction result.

[0089] According to some embodiments of the present invention, the second extraction module includes:

[0090] The first analysis module is used to:

[0091] Based on the preset window size and step length, a rectangular window is slid on the CT image; based on the structural similarity algorithm, the similarity between the central area of ​​the rectangular window and the adjacent area is calculated; the center distance between the central area and the adjacent area is calculated, and the weight coefficient of the adjacent area and the central area is set according to the ratio of the preset standard distance to the center distance;

[0092] According to the similarity between the adjacent areas and the central area and the corresponding weight coefficient, the position feature value of the central area of ​​the rectangular window is determined;

[0093] The second analysis module is used to:

[0094] Apply the principal component analysis algorithm to the CT image, reduce the dimension of the CT image to the feature dimension space, project the central area of ​​all rectangular windows onto the feature dimension space, and determine the dimensional feature value of the central area of ​​the rectangular window;

[0095] A central area whose position feature value is greater than a preset position feature value and whose dimension feature value is greater than a preset dimension feature value is screened out as a second feature extraction result.

[0096] The working principle of the above technical solution is as follows: a rectangular window of a preset size is set on the CT image, and the window is slid on the image according to a preset step size. This window is used to gradually scan the entire image to capture features at different positions. For each sliding position, the similarity between the central area of ​​the rectangular window and the adjacent area is calculated. The structural similarity algorithm (Structural Similarity Index, SSIM) is used here, which is an indicator to measure the similarity between two images. SSIM takes into account the brightness, contrast and structural information of the image and can more accurately reflect the similarity between images. The center distance between the central area and the adjacent area is calculated. This distance is the Euclidean distance. According to the ratio of the preset standard distance to the calculated center distance, the weight coefficient of the adjacent area and the central area is set. This weight coefficient reflects the importance of the adjacent area for the calculation of the characteristic value of the central area. The area closer to the central area will be given a higher weight. According to the similarity between the adjacent area and the central area and the corresponding weight coefficient, the position characteristic value of the central area of ​​the rectangular window is calculated. This characteristic value is a value that combines the similarity of the adjacent area and the weight coefficient, which reflects the relative position and characteristics of the central area in the image.

[0097] Flatten the central region of each window into a one-dimensional vector for PCA processing. Combine all the flattened central region vectors into a two-dimensional data matrix, where each row represents an eigenvector of the central region. Standardize the data so that each feature has zero mean and unit variance. Calculate the covariance matrix of the data matrix. Use eigendecomposition or singular value decomposition (SVD) to calculate the eigenvectors and eigenvalues ​​of the covariance matrix. Select the first k principal components (i.e., the eigenvectors corresponding to the largest k eigenvalues) based on the size of the eigenvalue. Project the original data onto the feature space composed of the first k principal components to obtain the reduced-dimensional data. For the central region of each window, extract its representation in the feature space (i.e., the reduced-dimensional vector). Project the central region of the rectangular window in the CT image onto the feature dimension space and determine its dimensional eigenvalue.

[0098] The beneficial effects of the above technical solution are: determining the position feature value of the central area of ​​the rectangular window and the dimension feature value of the central area of ​​the rectangular window, and screening out the central area whose position feature value is greater than the preset position feature value and whose dimension feature value is greater than the preset dimension feature value as the valid area and as the second feature extraction result.

[0099] According to some embodiments of the present invention, the determining module includes:

[0100] A comparison module, used to determine TBS data according to the first feature extraction result, and compare the TBS data with the TBS threshold to obtain a comparison result;

[0101] Matching modules for:

[0102] Determine a color change matrix of the central area according to the second feature extraction result, and match the color change matrix with a preset color change matrix in a preset database to obtain a matching result;

[0103] Auxiliary analysis suggestions are determined based on the comparison results and the matching results.

[0104] The working principle of the above technical solution: trabecular bone score (TBS) is a structural parameter that can be obtained from DXA images to reflect the grayscale changes on the image, which can effectively evaluate the microstructure of the bone and describe the quality of the bone. The comparison results are: TBS data greater than or equal to the TBS threshold indicates that the bone microstructure is dense, the trabeculae are well connected, and the gaps between the trabeculae are small; TBS data less than the TBS threshold indicates that the bone microstructure is incomplete, the trabeculae are poorly connected, and the gaps between the trabeculae are large.

[0105] The preset color change matrix database may contain multiple color change patterns and their corresponding labels or classifications. The similarity between the extracted matrix and the matrix in the database is measured based on the Euclidean distance to obtain a matching result, and the corresponding bone information is determined based on the matching preset color change matrix.

[0106] The beneficial effect of the above technical solution is that it is convenient to accurately determine the auxiliary analysis suggestions based on the comparison results and the matching results.

[0107] According to some embodiments of the present invention, the matching module includes:

[0108] A transformation module, used for transforming the central area in the second feature extraction result into an N*M image matrix of pixels; N is the length of the image matrix; M is the width of the image matrix;

[0109] The first calculation module is used to obtain the value of each pixel in the R channel, the G channel and the B channel, and calculate the color feature value according to the value of each pixel in the R channel, the G channel and the B channel to obtain the color change matrix;

[0110]

[0111] Among them, W λ is the color feature value of the λth pixel in the color change matrix; S i is the optimization parameter of the i-th channel in the image matrix Q; Q i,λ is the value of the i-th channel of the λ-th pixel of the image matrix Q; Q j,λ is the value of the j-th channel of the λ-th pixel of the image matrix Q.

[0112] The working principle of the above technical solution: N×M image matrix, each element of which represents the color value of a pixel. The values ​​of each pixel in the three color channels of R (red), G (green), and B (blue) are calculated, and the color feature values ​​are calculated based on these values ​​to generate a color change matrix. The summation operation in the formula involves traversing the three channels of R, G, and B.

[0113] The beneficial effects of the above technical solution are: it is easy to accurately determine the color change matrix of the central area, thereby improving the accuracy of matching the color change matrix with the preset color change matrix in the preset database.

[0114] like Figure 3 As shown, according to some embodiments of the present invention, an analysis method of a bone monitoring data auxiliary analysis system includes steps S1-S4:

[0115] S1. Obtain bone monitoring data; the bone monitoring data includes DXA images and CT images;

[0116] S2, extracting features from the DXA image to obtain a first feature extraction result;

[0117] S3, extracting features from the CT image to obtain a second feature extraction result;

[0118] S4. Determine auxiliary analysis suggestions according to the first feature extraction result and the second feature extraction result.

[0119] The beneficial effects of the above technical solution: Combining the data of DXA and CT images can more comprehensively evaluate the health status of bones. Automated bone monitoring data analysis is achieved to improve diagnostic efficiency and accuracy. According to the specific situation of the patient, personalized auxiliary analysis suggestions are provided to help formulate more accurate treatment plans.

[0120] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A bone monitoring data auxiliary analysis system, characterized in that: include: An acquisition module is used to acquire bone monitoring data; The bone monitoring data includes DXA images and CT images; A first extraction module, used for performing feature extraction on the DXA image to obtain a first feature extraction result; The second extraction module is used to extract features from the CT image to obtain a second feature extraction result; The determination module is used to determine auxiliary analysis suggestions based on the first feature extraction result and the second feature extraction result.

2. The bone monitoring data auxiliary analysis system according to claim 1, characterized in that: The acquisition module comprises: DXA scanners are used to irradiate patients in a static state with X-rays. After the scan is completed, DXA images are generated; The CT scanner is used to perform cross-sectional scanning on a patient using X-rays to obtain cross-sectional images of bones as CT images.

3. The bone monitoring data auxiliary analysis system according to claim 1, characterized in that: The first extraction module comprises: A preprocessing module, used for performing image preprocessing on the DXA image to obtain a preprocessed image; The generation module is used to extract features from the preprocessed image based on the SBCFormer network to obtain a first feature extraction result.

4. The bone monitoring data auxiliary analysis system according to claim 3, characterized in that: The preprocessing module comprises: A smoothing processing module is used for smoothing the DXA image to obtain a smoothed image; A segmentation module is used to perform image segmentation based on the Otsu method, taking the gray value with the minimum variance within the smoothed image and the maximum variance between classes as the target threshold, and obtain a segmented image with a bone tissue area according to the target threshold; The contour marking module is used to perform edge detection on the segmented image based on the edge detection operator, and mark the detected contour points to obtain a preprocessed image.

5. The bone monitoring data auxiliary analysis system according to claim 4, characterized in that: The smoothing processing module performs smoothing processing on the DXA image to obtain a smoothed image, including: Among them, G(x,y) is the data at (x,y) in the smoothed image; x is the distance from the origin on the horizontal axis; y is the distance from the origin on the vertical axis; e is a natural constant; σ is the standard deviation of the Gaussian distribution, and σ is a variable value between 0 and 3.

6. The bone monitoring data auxiliary analysis system according to claim 3, characterized in that: The generating module comprises: The propagation module is used to forward propagate the preprocessed image based on the convolutional layer, pooling layer, and fully connected layer structure in the SBCFormer network; The first calculation module is used to calculate the first correlation between different channel features in the SBCFormer network and the second correlation between features at different positions in the image during the forward propagation process; optimize the features in the forward propagation process according to the first correlation and the second correlation to obtain the first feature extraction result.

7. The bone monitoring data auxiliary analysis system according to claim 1, characterized in that: The second extraction module comprises: The first analysis module is used to: Based on the preset window size and step length, a rectangular window is slid on the CT image; based on the structural similarity algorithm, the similarity between the central area of ​​the rectangular window and the adjacent area is calculated; the center distance between the central area and the adjacent area is calculated, and the weight coefficient of the adjacent area and the central area is set according to the ratio of the preset standard distance to the center distance; According to the similarity between the adjacent areas and the central area and the corresponding weight coefficient, the position feature value of the central area of ​​the rectangular window is determined; The second analysis module is used to: Apply the principal component analysis algorithm to the CT image, reduce the dimension of the CT image to the feature dimension space, project the central area of ​​all rectangular windows onto the feature dimension space, and determine the dimensional feature value of the central area of ​​the rectangular window; A central area whose position feature value is greater than a preset position feature value and whose dimension feature value is greater than a preset dimension feature value is screened out as a second feature extraction result.

8. The bone monitoring data auxiliary analysis system according to claim 7, characterized in that: The determining module comprises: A comparison module, used to determine TBS data according to the first feature extraction result, and compare the TBS data with the TBS threshold to obtain a comparison result; Matching modules for: Determine a color change matrix of the central area according to the second feature extraction result, and match the color change matrix with a preset color change matrix in a preset database to obtain a matching result; Auxiliary analysis suggestions are determined based on the comparison results and the matching results.

9. The bone monitoring data auxiliary analysis system according to claim 8, characterized in that: The matching module comprises: A transformation module, used for transforming the central area in the second feature extraction result into an N*M image matrix of pixels; N is the length of the image matrix; M is the width of the image matrix; The first calculation module is used to obtain the value of each pixel in the R channel, the G channel and the B channel, and calculate the color feature value according to the value of each pixel in the R channel, the G channel and the B channel to obtain the color change matrix; Among them, W λ is the color feature value of the λth pixel in the color change matrix; S i is the optimization parameter of the i-th channel in the image matrix Q; Q i,λ is the value of the i-th channel of the λ-th pixel of the image matrix Q; Q j,λ is the value of the j-th channel of the λ-th pixel of the image matrix Q.

10. The analysis method of the bone monitoring data auxiliary analysis system according to any one of claims 1 to 9, characterized in that: include: Obtain bone mass monitoring data; The bone monitoring data includes DXA images and CT images; Performing feature extraction on the DXA image to obtain a first feature extraction result; Performing feature extraction on the CT image to obtain a second feature extraction result; An auxiliary analysis suggestion is determined according to the first feature extraction result and the second feature extraction result.

Images