Bone mineral density detection method

Abstract

The invention relates to the technical field of bone mineral density detection, and discloses a bone mineral density detection method which comprises the following steps: step 1) image acquisition; step 2) image segmentation; 3) reconstructing a three-dimensional surface; 4) reconstructing three-dimensional body distribution; 5) performing calibration calculation; according to the method, the automatic segmentation algorithm, the three-dimensional reconstruction algorithm and the gray density mapping technology based on the Lambert-Beer law are combined, so that the lumbar vertebra bone mineral density is rapidly and accurately detected, and a reliable basis is provided for early diagnosis of osteoporosis; in addition, the method is based on common X-ray images, is low in radiation dose, is lower in equipment cost, is suitable for basic medical institutions and resource-limited regions, and has wide popularization value and application prospect.

Description

Technical Field

[0001] The present invention relates to the technical field of bone density detection, and in particular to a bone density detection method. Background Art

[0002] Bone mineral density (BMD) is an important indicator for assessing bone health and is widely used in the diagnosis of osteoporosis and fracture risk assessment. Osteoporosis is a common chronic disease, especially prevalent among the elderly, characterized by decreased bone mass and degeneration of bone microstructure, leading to increased bone fragility. By measuring bone density, osteoporosis and its progression can be detected earlier, thereby helping doctors develop prevention and treatment plans.

[0003] As the global aging trend intensifies, the incidence of osteoporosis has also increased significantly. According to statistics from the World Health Organization (WHO), a fracture caused by osteoporosis occurs every three seconds in the world. Bone density measurement, as a key means of assessing bone health, has become an important tool for preventing and managing osteoporosis.

[0004] Dual-energy X-ray Absorptiometry (DXA): This is the most commonly used bone density measurement technology in clinical practice. It measures the difference in the intensity of absorbed X-rays by passing X-rays of different energies through the bones to calculate bone density. DXA has high accuracy and low radiation dose, and is the "gold standard" for diagnosing osteoporosis.

[0005] Quantitative computed tomography (QCT): Based on CT imaging technology, QCT can provide three-dimensional bone density information and measure the density of cancellous bone and cortical bone separately, which is suitable for the detection of early osteoporosis. However, the radiation dose of QCT is high, and its clinical application is subject to certain limitations.

[0006] However, existing bone density measurement technologies still have some limitations. Take DXA as an example. Although it has high measurement accuracy, its measurement results only provide two-dimensional projection information and cannot reflect the three-dimensional density distribution of bones. It is also easily affected by the patient's body shape, posture and other tissue structures, resulting in measurement errors. In addition, DXA equipment is expensive and has high requirements for the operating environment, which limits its popularization and application in primary medical institutions or areas with limited resources. Although QCT can provide three-dimensional density information of bones, its high radiation dose and expensive cost make it difficult to use as a routine screening method. Summary of the invention

[0007] In view of this, the purpose of the present invention is to provide a new bone density detection method, aiming to simplify the detection process, reduce equipment costs and improve detection accuracy, thereby effectively increasing the popularity of early diagnosis of osteoporosis.

[0008] To achieve the above objectives, the present invention adopts the following technical solutions:

[0009] A method for detecting bone density comprises the following steps:

[0010] Step 1) Image acquisition: a reference module with known equivalent bone density and the part of the human body to be tested are irradiated and imaged under the same conditions to obtain an X-ray image containing the lumbar spine and the reference module; the standard density value of the reference module is known, and its gray value distribution in the image is clear and distinguishable; (the reference module is a physical 3D model; the human lumbar spine and the reference module are in the same X-ray image)

[0011] Step 2) image segmentation, separating the lumbar vertebral body and the reference module from the background by using a preferred segmentation algorithm, and extracting the grayscale information of the segmented image;

[0012] Step 3) 3D surface reconstruction: based on the segmented image, a basic 3D surface model of the lumbar vertebral body and the reference module is established using an optimal 3D surface reconstruction algorithm, and then the established basic 3D surface model is continuously optimized according to the 2D contour of the segmented image;

[0013] Step 4) 3D volume distribution reconstruction: based on the preferred 3D volume distribution reconstruction algorithm, the grayscale information of the segmented image is mapped to the 3D space to obtain an initial 3D volume distribution model, and a series of optimizations are performed on the initial volume distribution model to finally obtain the 3D space density distribution information that matches the grayscale information;

[0014] Step 5) Calibration calculation, combining the optimized body distribution model, establishes the mapping relationship between the three-dimensional spatial density information and the grayscale information, and then converts the grayscale information into the bone density value and outputs it based on the mapping relationship between the three-dimensional spatial density distribution information and the bone density.

[0015] Furthermore,

[0016] S1.1 Use conventional X-ray equipment to collect X-ray images I containing the lumbar spine and a reference module. The image must include the target lumbar spine area and a reference module with a known physical density;

[0017] S1.2: The standard density value of the reference module is known, and its gray value distribution in the image must be clearly distinguishable.

[0018] Furthermore, the image segmentation includes a segmentation module and a grayscale extraction module;

[0019] The segmentation module is responsible for separating the lumbar vertebral body and the reference module from the background in the X-ray image containing the human body and the reference module, and completes this task through the preferred segmentation algorithm to ensure the accuracy and robustness of the segmentation, and saves the segmentation result;

[0020] The grayscale extraction module extracts and saves the grayscale information of the lumbar vertebral body and the reference module from the segmented image, providing basic data for subsequent work.

[0021] Furthermore,

[0022] S2.1: Input the acquired lumbar spine X-ray image into the pre-trained automatic segmentation model (the segmentation model is U-net++) to separate the lumbar spine region of interest and the reference module;

[0023] S2.2: extracting the grayscale information G(x, y) of the segmented area, where G(x, y) represents the grayscale distribution of the image on a two-dimensional plane;

[0024] S2.3: The output results are the lumbar segmentation image R(x,y) and the reference module segmentation image R ref (x,y).

[0025] Furthermore, the segmentation module includes:

[0026] Obtain the training data required for the segmentation model, including the original X-ray image and L1-L4 lumbar vertebra segmentation mask labels;

[0027] The original X-ray image is used as the input for model training, the corresponding lumbar vertebral segmentation mask label is used as the gold standard, and the difference between the model segmentation result and the gold standard is used as the supervised learning target to train the segmentation model and obtain a segmentation model that minimizes the difference between the model segmentation result and the gold standard.

[0028] Further, the three-dimensional surface reconstruction includes: a three-dimensional surface model reconstruction module and a three-dimensional surface model optimization module;

[0029] The three-dimensional surface model reconstruction module uses the lumbar segmentation image and adopts an optimal three-dimensional surface reconstruction algorithm based on the segmentation image to generate a basic three-dimensional model of the lumbar vertebral body and the reference module;

[0030] The three-dimensional surface model optimization module performs two-dimensional projection on the basic three-dimensional model, compares the projection image with the two-dimensional contour of the segmented image, and iteratively optimizes the basic three-dimensional surface model based on the contour difference, so that the optimized three-dimensional surface model more accurately matches the two-dimensional contour of the segmented image.

[0031] Furthermore,

[0032] S3.1: Lumbar region image R(x,y) based on segmentation and reference module segmentation image R ref (x, y), using the 3D reconstruction algorithm, generate the lumbar 3D surface model M(x, y, z) and the reference module 3D surface model M ref (x,y,z);

[0033] S3.2: Compare the generated lumbar 3D surface model M(x, y, z) with the reference module 3D surface model M ref (x, y, z) is projected onto a two-dimensional plane to generate a projection image I(x, y);

[0034] S3.3: Continuously optimize the three-dimensional surface model by reducing the difference between the projection I(x,y) and the two-dimensional contour of the real X-ray image I.

[0035] Furthermore, the 3D surface model reconstruction module and the 3D surface model optimization module include:

[0036] Read the segmented image, set a threshold to binarize the image to obtain the two-dimensional contours of the lumbar vertebral body and the reference module;

[0037] Based on the two-dimensional contours of the vertebral body and the reference module, combined with prior knowledge, a basic three-dimensional model of the lumbar vertebral body and the reference module is established using an optimal three-dimensional surface reconstruction algorithm;

[0038] The established basic three-dimensional surface model is projected two-dimensionally, and the difference between the generated projection image contour and the two-dimensional contours of the lumbar vertebral body and the reference module is compared, and the three-dimensional surface model is continuously optimized based on the difference.

[0039] Further, the three-dimensional volume distribution reconstruction includes: a volume distribution reconstruction module and an optimization module;

[0040] The volume distribution reconstruction uses an optimal three-dimensional volume distribution reconstruction algorithm to reconstruct the two-dimensional grayscale information of the segmented image into a three-dimensional space to obtain an initial three-dimensional volume distribution model;

[0041] The volume distribution reconstruction module comprises:

[0042] Extract the grayscale information of the target area from the image segmentation result;

[0043] Based on prior knowledge, the two-dimensional grayscale information is reconstructed into three-dimensional space using an optimal three-dimensional volume distribution reconstruction algorithm to obtain an initial three-dimensional volume distribution model containing three-dimensional space density information.

[0044] The optimization module limits the spatial density information mapped into the three-dimensional space within the established three-dimensional model, and continuously iteratively optimizes the three-dimensional volume distribution model by combining the difference in grayscale information between the simulated X-ray image and the segmented image obtained by performing simulated X-ray projection on the initial three-dimensional volume distribution model.

[0045] The optimization module is characterized by comprising: a projection module and a three-dimensional space density distribution information optimization module;

[0046] The projection module performs simulated X-ray projection on the three-dimensional spatial density distribution information to obtain a simulated X-ray image;

[0047] Projection module, including:

[0048] According to actual needs, set the position of the X-ray source and select the appropriate projection angle;

[0049] Along the propagation path of the ray, calculate the intersection of the ray and each spatial density information, and perform attenuation calculation;

[0050] The projection results are synthesized into a set of two-dimensional image data, and the projection images are subjected to necessary post-processing to simulate the real X-ray imaging effect.

[0051] The three-dimensional spatial density distribution information optimization module compares the differences between the corresponding areas of the obtained simulated X-ray image and the original lumbar spine X-ray segmentation image, and continuously optimizes the distribution of the three-dimensional spatial density information according to the differences.

[0052] Going further,

[0053] S4.1: Import the two-dimensional grayscale value information G(x,y) of the lumbar vertebrae and the reference module in the segmented image, and use the preferred three-dimensional volume distribution reconstruction algorithm to map the two-dimensional grayscale information to the three-dimensional space to obtain an initial three-dimensional volume distribution model μ(x,y,z). (The segmented image refers to the segmented image of the lumbar vertebra and the reference module, not just the segmented image of the lumbar vertebra)

[0054] S4.2: Using the established 3D surface model of the lumbar spine and reference module, perform boundary constraints on the 3D information distribution and remove abnormal voxels. The boundary constraint is expressed as:

[0055]

[0056] Among them, M represents the three-dimensional model boundary of the lumbar spine and the reference module, and μ′(x, y, z) is the three-dimensional density information after constraint.

[0057] S4.3: Import the constrained three-dimensional density information μ′(x, y, z), determine the position of the ray source, and select a suitable projection angle;

[0058] S4.4: Ray propagation simulation, simulating X-rays starting from the source point and passing through the three-dimensional space model. According to the Lambert-Beer law, during the ray propagation process, the attenuation effect of each three-dimensional density information on the ray is calculated;

[0059] S4.5: Synthesize the projection results into a set of two-dimensional image data I combined , and compare it with the two-dimensional grayscale information G(x,y) of the segmented image, and continuously optimize the distribution of three-dimensional spatial density information based on the difference between the two.

[0060] Further, calibration calculations include:

[0061] Combined with the optimized three-dimensional volume distribution model, a mapping relationship between two-dimensional grayscale information and three-dimensional spatial density distribution information is established;

[0062] Then, based on the conversion relationship between the three-dimensional spatial density distribution information and the bone density, the three-dimensional spatial density distribution information in the three-dimensional volume distribution model is converted into the corresponding bone density value.

[0063] Going further,

[0064] S5.1: Combined with the optimized three-dimensional volume distribution model, a mapping relationship between the three-dimensional spatial density distribution information μ′(x, y, z) and the two-dimensional grayscale information G(x, y) is established:

[0065] μ=a·G+b

[0066] Wherein, a is the proportional coefficient and b is the offset, and a and b are determined by fitting the measurement data of the reference module using the least squares method;

[0067] S5.2: Based on the conversion relationship between the three-dimensional spatial density distribution information and the bone density, the three-dimensional spatial density distribution information μ′(x, y, z) in the three-dimensional volume distribution model is converted into a corresponding bone density value.

[0068] The beneficial effects of the present invention are as follows:

[0069] The present invention proposes a bone density detection scheme based on ordinary X-ray images. By combining automatic segmentation algorithm, three-dimensional reconstruction algorithm and grayscale density mapping technology based on Lambert-Beer's law, it realizes rapid and accurate detection of lumbar vertebrae bone density, providing a reliable basis for the early diagnosis of osteoporosis.

[0070] The technical improvements of the present invention are reflected in the following aspects: First, by using the automatic segmentation model of deep learning, the areas of the lumbar spine and the reference module can be separated from the X-ray image efficiently and accurately, reducing the interference of human operation; second, by using a three-dimensional model generated based on the accurately segmented X-ray image, the three-dimensional model of the lumbar vertebrae can be accurately reconstructed from the two-dimensional image, greatly improving the reconstruction accuracy; finally, by combining the known density information of the reference module, a mapping relationship between the grayscale value and the bone density value is established, effectively improving the accuracy of bone density calculation.

[0071] In addition, the present invention is based on ordinary X-ray images, has low radiation dose and low equipment cost, is suitable for primary medical institutions and areas with limited resources, and has broad promotion value and application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 It is a flowchart of a bone density detection method disclosed in an embodiment of the present invention;

[0073] Figure 2 It is a schematic diagram of a process of performing lumbar vertebral image segmentation and establishing and optimizing a three-dimensional surface model of the lumbar vertebral body based on the segmentation result in a bone density detection method disclosed in an embodiment of the present invention;

[0074] Figure 3 It is a schematic diagram of a specific implementation process of mapping two-dimensional grayscale information into three-dimensional space and using an established three-dimensional model to perform boundary constraints in a bone density detection method disclosed in an embodiment of the present invention;

[0075] Figure 4 It is a schematic diagram of a specific implementation process of a bone density detection method disclosed in an embodiment of the present invention, which generates a projection image based on the three-dimensional spatial density information distribution, compares it with the original X-ray segmentation image, and continuously iterates and optimizes the three-dimensional spatial density distribution information based on the grayscale information difference. DETAILED DESCRIPTION

[0076] The present invention is described in detail below in conjunction with specific embodiments. The following examples are helpful for those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several changes and improvements can be made without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0077] The present invention discloses a bone density detection method, which is helpful to simplify the detection process compared with the prior art, and can perform bone density detection under conventional medical conditions, thereby improving the early diagnosis and popularization rate of osteoporosis. Detailed descriptions are given below.

[0078] Embodiment 1:

[0079] See also Figure 1 , Figure 1 The figure is a schematic diagram of the overall process of a bone density detection method disclosed in an embodiment of the present invention, which specifically includes the following steps:

[0080] S1. Acquire an X-ray image including a human lumbar spine and a reference module;

[0081] S1.1. Collect X-ray images of the lumbar spine of the individual to be tested and the reference module;

[0082] S1.2. Perform preprocessing such as standardization, normalization, and denoising on the images to ensure consistent image quality and remove unnecessary noise or artifacts.

[0083] S2. Separate the lumbar spine and reference module from the background through the automatic segmentation algorithm, and extract the grayscale information of the segmented image:

[0084] The construction of the lumbar spine automatic segmentation model is based on the deep convolutional neural network framework. Commonly used networks include u-net, U-Net++, SegNet, DeepLab, etc. The input of the model is the preprocessed lumbar spine X-ray image and the lumbar spine segmentation label annotated by experienced radiologists. The output is the binary segmentation result image of the lumbar vertebrae.

[0085] Optionally, the model training process is as follows:

[0086] S2.1. Collect the original X-ray image dataset required for model training and perform preprocessing such as standardization, normalization, and denoising on the dataset to ensure consistent image quality and remove unnecessary noise or artifacts;

[0087] S2.2. Under the guidance of professional technicians, manually or semi-automatically annotate the pre-processed X-ray images to generate the label image (ground truth) required for segmentation. The accuracy of the annotation directly affects the subsequent segmentation effect;

[0088] S2.3. The preprocessed lumbar spine X-ray images are used as the input for model training, the labeled images generated under the guidance of professional technicians are used as the gold standard, and the difference between the model segmentation result and the gold standard (loss function Ls) is used as the supervised learning target to train the lumbar spine segmentation model;

[0089] S2.4. Finally, an automated lumbar spine segmentation model is obtained that minimizes the difference between the model segmentation results and the gold standard.

[0090] The deep learning segmentation model used in the embodiment is an improved U-net++. During the model training process, the input is a lumbar spine X-ray image that has been data-augmented (flipped, translated, cropped, mirrored, noise added, etc.), and the output is the corresponding lumbar vertebral body segmentation image. Adam is used as the optimizer for back propagation in network training, the number of iterations (epoch) is 500 times, the batch size is 8, and the input image size is 256*256.

[0091] S3. Based on the segmented image, a three-dimensional model of the lumbar spine and the reference module is established using a three-dimensional reconstruction algorithm, and the established three-dimensional model is continuously optimized according to the two-dimensional contour of the segmented image:

[0092] The construction of the three-dimensional reconstruction model of the lumbar vertebrae is based on the segmented image, the binarization method is selected, and the appropriate threshold is set to obtain the two-dimensional contour of the lumbar vertebrae.

[0093] See also Figure 2 , Figure 2 The present invention discloses a bone density detection method for segmenting lumbar vertebral images and performing three-dimensional surface reconstruction on the segmentation results, which specifically includes the following steps:

[0094] S3.1. Import the segmented image, set a suitable threshold, and binarize the image to obtain the two-dimensional contours of the lumbar vertebral body and the reference module;

[0095] S3.2. Based on the two-dimensional contours of the lumbar vertebral body and the reference module, according to prior knowledge, a three-dimensional reconstruction algorithm is used (Step 1. Contour curve input: Input the two-dimensional contour curve data, which is given in the form of a set of discrete points, and the coordinates of each point are expressed as (x i, y i ); Step 2. Setting the rotation axis: 1) Set the rotation axis to the line connecting the upper and lower midpoints of the minimum circumscribed rectangle of the two-dimensional contour, and only keep the curve on the right side of the rotation axis; 2) Define the rotation angle range as 0 to 2π; Step 3. Generate a three-dimensional point cloud, 1) For each point (x i, y i ), calculate its three-dimensional coordinates (x i, y i ,z i ); 2) repeat the above operation to form a complete three-dimensional point cloud data; step 4. construct a triangular mesh based on the point cloud data to generate a three-dimensional surface), and establish a simple lumbar vertebral body and a reference module of the three-dimensional surface model;

[0096] S3.3. Project the established three-dimensional surface model onto a two-dimensional plane. In order to be as close to the actual situation as possible, a perspective projection method is used;

[0097] Optionally, perspective projection is performed as follows:

[0098] Perspective projection maps the depth of a three-dimensional object onto a two-dimensional plane.

[0099]

[0100] Among them, (x, y, z) are the coordinates of the point in three-dimensional space, f is the focal length, and (x′, y′) are the projection coordinates of the plane.

[0101] S3.4. Compare the result of the two-dimensional projection with the binarized segmented image, and continuously optimize the simple benchmark three-dimensional surface model established in step S3.2 based on the difference between the two. Set a fixed error standard, and output the result when the optimized result meets the error requirement. If it does not meet the requirement, continue to optimize.

[0102] S4. Using the preferred three-dimensional volume distribution reconstruction algorithm (step 1. input data preparation: 1) to obtain a set of X-ray image data, represented as a two-dimensional grayscale image set {I 1, I 2, …,I n}, the resolution of each image is W×H; 2) Based on prior knowledge, obtain the voxel data range of grayscale information in three-dimensional space; Step 2. Initialize the volume data matrix, establish a three-dimensional volume data matrix V(x, y, z) based on the surface three-dimensional model optimized by S3.4 and initialize it; Step 3. Fill the volume data matrix: fill the voxel data obtained in step 1 into the initialized volume data matrix to obtain an initial three-dimensional volume distribution model. ), map the grayscale information to three-dimensional space, obtain an initial three-dimensional volume distribution model, and perform a series of optimizations on it, including constraining the boundaries of the three-dimensional space density information distribution combined with the established three-dimensional volume distribution model and comparing the difference in grayscale information between the X-ray image simulated by the volume distribution model and the original X-ray segmentation image to continuously iteratively optimize the volume distribution model;

[0103] See also Figure 3 , Figure 3 This is a flowchart of a specific implementation process of a bone density detection method disclosed in an embodiment of the present invention for mapping two-dimensional grayscale information into a three-dimensional space and using an established three-dimensional model to perform boundary constraints, which specifically includes the following steps:

[0104] S4.1. According to Lambert-Beer's law, when X-rays pass through an object, the attenuation of the rays is related to the density of the object and the path length of the rays. The formula is as follows:

[0105] I=I0·e -μx

[0106] Where I0 is the intensity of the incident ray, I is the intensity of the ray after penetrating the object, μ is the linear attenuation coefficient of the object, and x is the path length of the ray through the object;

[0107] This formula shows the relationship between grayscale values ​​and the density distribution of an object. Grayscale values ​​are usually related to ray attenuation, so grayscale information can reflect the physical density at different locations. In order to map grayscale values ​​to three-dimensional space, a pre-trained mapping algorithm model can be used. This model has learned the relationship between grayscale information and density values. The key steps of the mapping process include:

[0108] S4.1.1. Extract the grayscale value of each pixel from the X-ray image;

[0109] S4.1.2. Use the trained mapping algorithm model to convert the grayscale value of each pixel into the corresponding three-dimensional spatial density value;

[0110] S4.1.2. Locate each converted density value to a corresponding position in the three-dimensional space using the established three-dimensional space coordinates;

[0111] S4.2. The three-dimensional surface model constrains the boundaries of the distribution of three-dimensional spatial density information:

[0112] The established 3D surface model not only provides the geometric structure information of the space, but also serves as a spatial constraint for density information. The shape and boundary of the 3D model can help determine which areas belong to the valid physical space and which areas should not have density information or the density should be zero. The specific steps include:

[0113] S4.2.1. Based on the reconstructed three-dimensional surface model, define its spatial boundary;

[0114] S4.2.2. Based on the boundaries of the 3D surface model, limit the density information to only exist inside the model or in the region of interest to avoid generating unreasonable density values ​​in blank areas;

[0115] S4.3. simulate X-ray projection of the three-dimensional spatial density distribution information, and iteratively optimize the three-dimensional spatial density distribution information according to the difference between the projection image and the segmented X-ray image to obtain the three-dimensional spatial density distribution information that matches the segmented X-ray image;

[0116] See also Figure 4 , Figure 4 The present invention discloses a method for detecting bone density, which generates a projection image based on the three-dimensional spatial density information distribution, compares the projection image with the original X-ray segmentation image, and continuously iterates and optimizes the three-dimensional spatial density distribution information based on the grayscale information difference. The method specifically includes the following steps:

[0117] The generation process of simulated X-ray images is as follows:

[0118] S4.3.1. Projection parameter setting, determine the projection direction, projection angle, focal position and other parameters of the X-ray. These parameters are consistent with the actual X-ray imaging geometric conditions so that the generated simulated X-ray image is as similar as possible to the actual X-ray image;

[0119] S4.3.2: Line integral calculation: X-rays are passed through the three-dimensional model from a specific angle to calculate the accumulated attenuation value of each voxel passed by each ray. The accumulated attenuation value reflects the absorption of the ray after passing through tissues of different densities in the body;

[0120] S4.3.3: Image generation, generating a corresponding grayscale image by using the calculated attenuation value;

[0121] S4.4. Compare the generated simulated X-ray image and the segmented image, and continuously optimize the density information distribution in the three-dimensional space based on the difference between the two.

[0122] S5. Combined with the known physical density information of the reference module, a mapping relationship between the physical density information and the three-dimensional space density information is established, and the density value of the lumbar spine is calculated and output;

[0123] S5.1. Combine the optimized three-dimensional volume distribution model to establish the mapping relationship between the three-dimensional spatial density distribution information and the two-dimensional grayscale information; use the known density reference to perform density distribution calibration and error correction, map the extracted spatial density information with the grayscale value through the density mapping function, and process and correct the density distribution in the three-dimensional model to ensure the accuracy of the final result;

[0124] S5.2. Then, based on the conversion relationship between the three-dimensional spatial density distribution information and the bone density, the three-dimensional spatial density distribution information in the three-dimensional body distribution model is converted into the corresponding bone density value, the bone density value of the lumbar vertebral body is output, and a report is generated.

[0125] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essence of the present invention. In the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other at will.

Claims

1. A method for detecting bone density, characterized in that: The steps include: Step 1) image acquisition, obtaining an X-ray image of a human lumbar spine and a reference module with known equivalent bone density; Step 2) image segmentation: using a segmentation algorithm to separate the lumbar vertebrae and reference modules in the image from the background, and extracting the grayscale information of the segmented image; Step 3) 3D surface reconstruction: based on the segmented image, a basic 3D surface model of the lumbar vertebral body and the reference module is established using a 3D surface reconstruction algorithm, and then the basic 3D surface model is established based on the 2D contour optimization of the segmented image; Step 4) 3D volume distribution reconstruction: based on the 3D volume distribution reconstruction algorithm, the grayscale information of the segmented image is mapped to the 3D space to obtain an initial 3D volume distribution model, and the initial 3D volume distribution model is optimized to finally obtain the 3D space density distribution information that matches the grayscale information; Step 5) calibration calculation, establish the mapping relationship between the three-dimensional spatial density information and the grayscale information, and then convert the grayscale information into a bone density value and output it based on the mapping relationship between the three-dimensional spatial density distribution information and the bone density.

2. A bone density detection method according to claim 1, characterized in that: The image acquisition in step 1) includes the following steps: 1.1) Acquire an original X-ray image I including a lumbar spine and a reference module, wherein the X-ray image I includes a target lumbar spine region and a reference module with a known equivalent bone density; 1.2) The original X-ray image I is standardized, normalized, denoised, and artifact removed to obtain an X-ray image I.

3. A bone density detection method according to claim 1, characterized in that: Image segmentation in step 2) includes the following steps: 2.1) Input the acquired lumbar spine X-ray image into the pre-trained segmentation model to separate the lumbar spine region of interest and the reference module; The segmentation model training process is as follows: Obtaining training data required for the segmentation model, including X-ray image I and lumbar vertebral segmentation mask labels; The X-ray image I is used as the input for model training, the corresponding lumbar vertebral segmentation mask label is used as GroundTruth, and the difference between the model segmentation result and GroundTruth is used as the supervised learning target. The segmentation model is trained and iteratively optimized to obtain a segmentation model that minimizes the difference between the model segmentation result and GroundTruth. 2.2) Extract the grayscale information G(x,y) of the segmented area, where G(x,y) represents the grayscale distribution of the image on a two-dimensional plane; 2.3) The output results are the lumbar segmentation image R(x,y) and the reference module segmentation image R ref (x,y).

4. A bone density detection method according to claim 1, characterized in that: The three-dimensional surface reconstruction in step 3) includes the following steps: 3.1) Read the lumbar segmentation image R(x,y), set a threshold to binarize the image to obtain the two-dimensional contours of the lumbar vertebral body and the reference module; 3.2) Based on the lumbar segmentation image R(x,y) and the reference module segmentation image R ref The 2D contour of (x, y) is used to generate the lumbar 3D surface model M(x, y, z) and the reference module 3D surface model M using the 3D surface reconstruction algorithm. ref (x,y,z); 3.3) Compare the generated lumbar 3D surface model M(x, y, z) with the reference module 3D surface model M ref (x, y, z) is projected onto a two-dimensional plane to generate a projection image I(x, y); 3.4) By reducing the difference between the two-dimensional contour of the projection image I (x, y) and the real X-ray image I, an error standard is set. When the optimized result meets the error requirement, the result is output. If it does not meet the error requirement, the optimization is continued.

5. A bone density detection method according to claim 1, characterized in that: The three-dimensional volume distribution reconstruction in step 4) includes the following steps: 4.1) Import the two-dimensional gray value information G(x,y) of the lumbar vertebrae and the reference module in the segmented image, and use the three-dimensional volume distribution reconstruction algorithm to map the two-dimensional gray information to the three-dimensional space to obtain an initial three-dimensional volume distribution model μ(x,y,z); 4.2) Using the established 3D surface model of the lumbar spine and reference module, the 3D information distribution is bounded and abnormal voxels are removed. The bounded constraint is expressed as: Where M represents the three-dimensional model boundary of the lumbar spine and the reference module, and μ′(x, y, z) is the three-dimensional density information after constraint; 4.3) Import the constrained three-dimensional density information μ′(x, y, z), determine the position of the ray source, and select the appropriate projection angle; 4.4) Ray propagation simulation: simulate X-rays starting from the source point and passing through the three-dimensional space model. According to Lambert-Beer's law, during the ray propagation process, along the ray propagation path, calculate the intersection of the ray and each spatial density information, perform attenuation calculation, and obtain the simulated X-ray projection result; 4.5) The simulated X-ray projection results obtained in 4.4) are synthesized into a set of two-dimensional image data I combined , and compare it with the grayscale information G(x,y) of the segmented image, and iteratively optimize the distribution of the three-dimensional spatial density information based on the difference between the two.

6. A bone density detection method according to claim 1, characterized in that: The calibration calculation in step 5) includes the following steps: 5.1) Combined with the optimized three-dimensional volume distribution model, the mapping relationship between the three-dimensional spatial density distribution information μ′(x, y, z) and the two-dimensional grayscale information G(x, y) is established: μ=a·G+b Wherein, a is the proportional coefficient and b is the offset, and a and b are determined by fitting the measurement data of the reference module using the least squares method; 5.2) Then, based on the conversion relationship between the three-dimensional spatial density distribution information and the bone density, the three-dimensional spatial density distribution information μ′(x, y, z) in the three-dimensional volume distribution model is converted into the corresponding bone density value.

Images