Method for acquiring three-dimensional model of ankle skeleton of human body by simulating load-bearing CT (computed tomography) scanning

Abstract

The invention provides a method for acquiring a three-dimensional model of an ankle skeleton of a human body by simulating load-bearing CT scanning. The method comprises the following steps: acquiring CT scanning data; uploading the CT scanning data to medical image processing and three-dimensional reconstruction software to obtain an initial scanning image; constructing a three-dimensional model according to the initial scanning image; and carrying out model verification on the three-dimensional model to obtain a skeleton three-dimensional model. According to the method, the high-resolution three-dimensional model of the ankle skeleton can be obtained, and reliable support is provided for medical research, clinical diagnosis and treatment of the ankle and research and development of medical instruments. Compared with a traditional method, the data obtained by simulating the load is more practical, the model quality is remarkably improved through the advanced technology and algorithm, and development of related research and application in the medical field is promoted.

Description

Technical Field

[0001] The present invention relates to the field of medical three-dimensional modeling and imaging technology, and particularly to a method for obtaining a three-dimensional model of the human ankle and foot bones by simulating weight-bearing CT scanning. Background Art

[0002] As a widely used examination method in the orthopedic field, CT, with its excellent imaging ability, clearly presents the structures of bones and soft tissues, providing a key basis for disease diagnosis and treatment. Traditional CT examinations mostly use the supine position. However, in ankle and foot examinations, many subtle changes, such as the deviation of the lower limb force line and the injury of small joint surfaces, are difficult to detect in the non-weight-bearing state. Only through weight-bearing imaging can the influence of gravity on the biomechanics of the ankle and foot be accurately reflected. This not only helps doctors accurately diagnose diseases such as flat feet and high-arched feet, but also provides detailed data support for the formulation of personalized treatment plans.

[0003] In addition, due to the characteristics of the ankle and foot bones, such as small structure, irregular shape, and easy occurrence of osteoporosis, conventional CT scan parameters are difficult to meet the clear display of their fine structures. Too large a scan slice thickness may lead to the loss of bone detail information, affecting the accuracy of the model; while too small a scan slice thickness will increase the scan time and radiation dose, and may also introduce more noise. In addition, in the process of model reconstruction, there are also certain deficiencies in existing reconstruction algorithms and software processing methods. Some algorithms are prone to artifacts and edge blurring when dealing with complex bone structures, resulting in deviations between the reconstructed model and the real bone structure. Summary of the Invention

[0004] Based on this, it is necessary to provide a method for obtaining a three-dimensional model of the human ankle and foot bones by simulating weight-bearing CT scanning in view of the above technical problems.

[0005] A method for obtaining a three-dimensional model of the human ankle and foot bones by simulating weight-bearing CT scanning includes the following steps:

[0006] Obtain CT scan data;

[0007] Upload the CT scan data to medical image processing and three-dimensional reconstruction software to obtain an initial scan image;

[0008] Construct a three-dimensional model according to the initial scan image;

[0009] Perform model verification on the three-dimensional model to obtain a three-dimensional bone model.

[0010] In one embodiment, before obtaining the CT scan data, it further includes:

[0011] Obtain the basic weight data of the user to be measured;

[0012] Obtain the load data based on the said basic body weight data;

[0013] Obtain the CT scan data of the user to be measured in the standard body position according to the said load data.

[0014] In one embodiment, obtaining the CT scan data includes:

[0015] Obtain the CT scan parameters, and obtain the initial scan data according to the said CT scan parameters;

[0016] Perform data reconstruction on the said initial scan data to obtain the reconstructed data;

[0017] Perform post-processing operations on the said reconstructed data to obtain the CT scan data.

[0018] In one embodiment, performing data reconstruction on the said initial scan data to obtain the reconstructed data includes:

[0019] Perform data reconstruction on the said initial scan data based on the iterative reconstruction MLEM algorithm and the iDose 4 algorithm to obtain a noise suppression image;

[0020] Obtain the preset reconstruction layer thickness and the preset layer spacing, and reduce the volume effect according to the said preset reconstruction layer thickness and the said preset layer spacing to obtain the reconstructed data.

[0021] In one embodiment, performing post-processing operations on the said reconstructed data to obtain the CT scan data includes:

[0022] Obtain the virtual single energy keV value, and optimize the image display effect according to the said virtual single energy keV value to obtain the CT scan data.

[0023] In one embodiment, uploading the said CT scan data to the medical image processing and three-dimensional reconstruction software to obtain the initial scan image includes:

[0024] Upload the said CT scan data to the medical image processing and three-dimensional reconstruction software to obtain a pre-scan image;

[0025] Perform preprocessing on the said pre-scan image through Gaussian filtering to obtain a smoothed scan graph;

[0026] Optimize the image contrast based on the said smoothed scan graph through a contrast enhancement algorithm to obtain the initial scan image.

[0027] In one embodiment, constructing a three-dimensional model according to the said initial scan image includes:

[0028] Create a new mask based on the initial scan image, obtain a preset initial mask threshold, segment the ankle bones and surrounding tissues according to the preset initial mask threshold, and obtain an initial segmentation image;

[0029] Create seed points based on the initial segmentation image, obtain the CT value difference between the seed point part and the surrounding tissues according to the seed points, obtain a region growing similarity threshold, and optimize the model edge through erosion and dilation operations according to the CT value difference and the region growing similarity threshold to obtain an optimized image;

[0030] Construct a three-dimensional image based on the optimized image, perform an integrity check on the three-dimensional image, and in response to the three-dimensional image being correct, upload the three-dimensional image to reverse engineering software;

[0031] Obtain preset STL model optimization parameters, and optimize the three-dimensional image through the reverse engineering software to obtain a three-dimensional model.

[0032] In one embodiment, obtaining preset STL model optimization parameters and optimizing the three-dimensional image through the reverse engineering software to obtain a three-dimensional model includes:

[0033] Receive normal direction adjustment information, adjust the normal direction of the triangular facets according to the normal direction adjustment information to obtain a corrected model;

[0034] Obtain a preset maximum side length threshold and an angle difference threshold, and adjust the corrected model according to the preset maximum side length threshold and the angle difference threshold to obtain an adjusted model;

[0035] Based on the adjusted model, perform iterative subdivision through the quadrilateral mesh Loop subdivision algorithm to obtain a precision model;

[0036] Based on the precision model, adjust the smoothing factor and the number of iterations through the smoothing function to optimize the surface quality of the model and obtain a smoothed three-dimensional model;

[0037] Perform a defect check on the smoothed three-dimensional model, and in response to the smoothed three-dimensional model having no defects, output the smoothed three-dimensional model as the three-dimensional model.

[0038] In one embodiment, performing model verification on the three-dimensional model to obtain a bone three-dimensional model includes:

[0039] Obtain real anatomical structure data, compare the three-dimensional model and the real anatomical structure data according to a preset standard through a distance measurement tool, and in response to the error between the three-dimensional model and the real anatomical structure data being within a preset range, use the three-dimensional model as the bone three-dimensional model;

[0040] In response to the error between the three-dimensional model and the real anatomical structure data being outside the preset range, adjust the three-dimensional model until the error between the three-dimensional model and the real anatomical structure data is within the preset range.

[0041] In one embodiment, real anatomical structure data is obtained, and the three-dimensional model and the real anatomical structure data are compared according to a preset standard by a distance measurement tool. In response to the error between the three-dimensional model and the real anatomical structure data being within the preset range, taking the three-dimensional model as the bone three-dimensional model includes:

[0042] Judge whether the overall dimension error of the model exceeds the preset first allowable range;

[0043] In response to not exceeding the preset first allowable range, judge whether the local feature point dimension error exceeds the preset second allowable range;

[0044] In response to not exceeding the preset second allowable range, take the three-dimensional model as the bone three-dimensional model.

[0045] Compared with the prior art, the advantages and beneficial effects of the present invention are as follows: The present invention can obtain a high-resolution three-dimensional model of the foot and ankle bones, providing reliable support for foot and ankle medical research, clinical diagnosis and treatment, and medical device R & D. Compared with traditional methods, the data obtained by the present invention through simulated weight-bearing is more in line with the actual situation, and the advanced technology and algorithms significantly improve the model quality, promoting the development of related research and applications in the medical field. Brief Description of the Drawings

[0046] Figure 1 It is a flowchart of a method for obtaining a three-dimensional model of the foot and ankle bones of a human body by simulated weight-bearing CT scanning in one embodiment;

[0047] Figure 2 It is a schematic diagram of a standardized body position in one embodiment;

[0048] Figure 3 It is a three-dimensional reconstruction diagram of an STL model of bilateral foot and ankle weight-bearing CT examination in one embodiment;

[0049] Figure 4 It is a three-view drawing of the cross-section of an STL model of bilateral foot and ankle weight-bearing CT examination in one embodiment. Detailed Description of the Embodiment

[0050] Before describing the specific embodiments of the present invention, the overall concept of the present invention is described as follows:

[0051] The present invention is mainly developed for the diagnosis, treatment, and medical research of foot and ankle-related diseases. Currently, there are problems such as insufficient accuracy and inaccurate model reconstruction in obtaining three-dimensional models of foot and ankle bones.

[0052] Therefore, the present invention proposes a method for obtaining a three-dimensional model of the human ankle and foot bones by simulating weight-bearing CT scanning. By simulating the weight-bearing state of the human body, the bone data of the ankle and foot closer to the real physiological state can be obtained, and combined with advanced CT scanning technology and professional software processing algorithms, an accurate construction of a high-resolution three-dimensional model of the ankle and foot bones can be achieved.

[0053] After introducing the overall concept of the present invention, in order to make the purpose, technical solution and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below through specific embodiments in combination with the accompanying drawings.

[0054] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in one or more embodiments of this specification should be the ordinary meanings understood by those with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar terms used in one or more embodiments of this specification do not represent any order, quantity or importance, but are only used to distinguish different components. Words such as "including" or "comprising" mean that the elements or objects appearing before this word cover the elements or objects listed after this word and their equivalents, without excluding other elements or objects. Words such as "connected" or "linked" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right", etc. are only used to represent relative position relationships. When the absolute position of the object being described changes, the relative position relationship may also change accordingly.

[0055] In one embodiment, as Figure 1 shown, a method for obtaining a three-dimensional model of the human ankle and foot bones by simulating weight-bearing CT scanning is provided, including the following steps:

[0056] Step S101, obtaining CT scan data.

[0057] Specifically, obtain the CT scan data of the weight-bearing human ankle and foot to be modeled.

[0058] On this basis, before obtaining the CT scan data, it further includes:

[0059] Obtain the basic weight data of the user to be tested;

[0060] Obtain the weight-bearing data according to the basic weight data;

[0061] Obtain the CT scan data of the user to be tested in the standard position according to the weight-bearing data.

[0062] Specifically, a weighing scale is used to obtain the weight data of the user to be measured. During the measurement, the user to be measured should wear light clothing and stand upright and steadily on the scale. After the reading stabilizes, record the weight. Accurate weight data is the basis for subsequent simulated weight-bearing.

[0063] As Figure 2 shown, the user to be measured lies supine on the CT examination table, with both hands placed on both sides of the body. By touching the bilateral patellas, guide the user to rotate the legs so that the patellas are located at the center of the coronal plane of the knee joint. The bilateral ankles are slightly separated to the same width as the pelvis, the toes are upward and kept on the same horizontal line to avoid height differences. The knees are slightly bent and padded and fixed with a foam pad or clothing at the back. Observed from the sagittal plane, the center of the knee joint and the center of the sole of the foot are on the same horizontal line, so that the mechanical conduction axis of the lower limb passes through the center of the sole of the foot, rather than the back of the sole of the foot.

[0064] Place a wooden box at the foot end of the user to be measured, and a mechanical scale is included in the box. Select rubber elastic bands with different elastic specifications. One end is connected to the box, and the other end is connected to the shoulder of the user to be measured. The first simulated weight-bearing is carried out through the contraction force of the rubber elastic band. Observe the reading of the mechanical scale. If there is a little error, fine-tuning can be carried out by the following method: Select rubber elastic bands with different elastic specifications. One end is connected to the box, and the other end is held by the subject being examined. According to the reading shown on the mechanical scale, the user to be measured can flex and extend the arm to flexibly adjust the tightness of the connection between the elastic band and the box for fine-tuning of the counterweight. Observe the reading of the mechanical scale until it is the same as the weight of the user to be measured, simulate the stress state of the ankle during human weight-bearing, and obtain the weight-bearing data. The mechanical scale can be designed in the box to ensure that the part in contact with the foot is made of wood to prevent metal from interfering with the CT scan image by generating artifacts. Obtain the CT scan data of the user to be measured in the standard position according to the weight-bearing data.

[0065] On this basis, obtaining the CT scan data includes:

[0066] Obtain the CT scan parameters, and obtain the initial scan data according to the CT scan parameters;

[0067] Perform data reconstruction on the initial scan data to obtain reconstructed data;

[0068] Perform post-processing operations on the reconstructed data to obtain the CT scan data.

[0069] Specifically, first obtain the CT scanning parameters, and scan according to these parameters to obtain the initial scanning data. Set the tube voltage to 120kVp and the tube current to 200mA to reduce image noise while ensuring X-ray penetration and radiation dose. Set the rotation speed to 0.5s / rot to ensure data quality while taking into account scanning efficiency. Set the pitch to 0.561:1. Using a smaller pitch can reduce image distortion. Set the collimator width to 0.6×64mm and the scanning matrix to 512×512 to ensure image resolution. According to the body shape of the user to be tested, flexibly adjust the scanning range to ensure that the foot and ankle are fully included.

[0070] After the initial scanning data is obtained, the original projection data is calculated and the image is reconstructed to obtain the reconstructed data.

[0071] The reconstructed data is then post-processed to obtain CT scan data.

[0072] On this basis, the initial scan data is reconstructed to obtain reconstructed data including:

[0073] Reconstructing the data based on the iterative reconstruction MLEM algorithm and the iDose 4 algorithm according to the initial scan data to obtain a noise suppressed image;

[0074] A preset reconstruction layer thickness and a preset layer spacing are obtained, and a partial volume effect is reduced according to the preset reconstruction layer thickness and the preset layer spacing to obtain reconstruction data.

[0075] Specifically, the initial scan data is reconstructed using the iterative reconstruction MLEM algorithm combined with the iDose 4 algorithm. This algorithm combination can effectively suppress noise. The principle of this algorithm combination effectively suppressing noise is mainly achieved through the statistical iterative characteristics of the MLEM algorithm and the intelligent filtering processing of the iDose 4 algorithm.

[0076] The MLEM algorithm uses an iterative approach to image reconstruction. The traditional filtered back projection (FBP) algorithm completes image reconstruction in one go, while the iterative reconstruction algorithm gradually optimizes the image through multiple iterations. In each iteration, the algorithm adjusts and corrects the image based on the current reconstruction result and the noise model. Specifically, the estimated noise is subtracted from the original data, and then the image is reconstructed again, and this process is repeated until satisfactory image quality is achieved. Through multiple iterations, the algorithm can gradually reduce the impact of noise while retaining the details and features of the image.

[0077] The iDose 4 algorithm has an adaptive filtering function, which adaptively adjusts the filtering parameters according to the local features of the image. In the flat area of ​​the image, the noise is relatively obvious, and the algorithm will use stronger filtering to suppress the noise; while in the edge and detail areas of the image, such as the bone edges of the ankle, in order to avoid blurring these important features, the algorithm will appropriately weaken the filtering strength. This adaptive filtering strategy enables the algorithm to flexibly adjust the degree of noise suppression according to the specific situation of the image, thereby achieving a better balance between noise suppression and image fidelity. Specifically, the Level value is set to 3 in the iDose 4 algorithm.

[0078] The MLEM algorithm is combined with the iDose 4 algorithm. The MLEM algorithm optimizes the overall data from a statistical iteration perspective, gradually converges to a more accurate image estimate, and suppresses noise at a macro level. The iDose 4 algorithm starts from the local features of the image and performs intelligent filtering according to the characteristics of different areas of the foot and ankle, accurately removing noise and retaining details. The two work together to more effectively suppress noise in foot and ankle CT scan data, provide high-quality image data for the subsequent construction of a three-dimensional model of the foot and ankle bones, and improve the accuracy and reliability of the model.

[0079] In order to preserve bone details as much as possible and reduce the partial volume effect, the thinnest layer thickness (preset reconstruction layer thickness) and minimum layer spacing (preset layer spacing) of the device are obtained, and the reconstruction layer thickness and layer spacing are set to the thinnest layer thickness and minimum layer spacing. In this embodiment, the reconstruction layer thickness is set to 0.6 mm and the layer spacing is set to 0.3 mm.

[0080] On this basis, the reconstructed data is post-processed to obtain CT scan data including:

[0081] A virtual single energy KeV value is obtained, and an image display effect is optimized according to the virtual single energy KeV value to obtain CT scanning data.

[0082] Specifically, the virtual low single-energy CTA technology of double-layer detector spectral CT is used to further process the SBI data obtained from the scan. In view of the characteristics of small bones and relatively porous bone in the foot and ankle, the range of virtual single-energy KeV values ​​is obtained at the level of the region of interest. The generally adjustable range is 40-120KeV. The virtual single-energy KeV value is adjusted to find the single-energy KeV value with the clearest bone display, and the value is used to reconstruct the DICOM data, that is, the CT scan data, to provide high-quality image data for subsequent modeling.

[0083] Step S102, uploading the CT scan data to medical image processing and three-dimensional reconstruction software to obtain an initial scan image.

[0084] Specifically, the reconstructed single-energy keV CT scan data is accurately imported into the medical image processing and three-dimensional reconstruction software. During the import process, the integrity and accuracy of the data are strictly checked to ensure that the software correctly identifies the data information. If an import error occurs, problems such as data format, file corruption, or software compatibility are promptly investigated and resolved. In this embodiment, the medical image processing and three-dimensional reconstruction software is Mimics software. The present invention is not limited to only this software, and all medical image processing and three-dimensional reconstruction software capable of implementing this function are within the scope of the present invention.

[0085] On this basis, the CT scan data is uploaded to the medical image processing and three-dimensional reconstruction software, and the initial scan images obtained include:

[0086] The CT scan data is uploaded to the medical image processing and three-dimensional reconstruction software to obtain a pre-scan image;

[0087] The pre-scan image is pre-processed by Gaussian filtering to obtain a smoothed scan graph;

[0088] Based on the smoothed scan graph, the image contrast is optimized by a contrast enhancement algorithm to obtain an initial scan image.

[0089] Specifically, Gaussian filtering is used to pre-process the pre-scan image. Gaussian filtering is a linear smoothing filter commonly used to remove image noise. The specific formula is as follows:

[0090]

[0091] Where (x, y) represents the coordinates of a pixel point in the image. In a two-dimensional image, each pixel has corresponding horizontal and vertical coordinates, which are used to determine the position of the pixel in the image. When performing a filtering operation, each pixel is processed based on these coordinates. (x0, y0) represents the center coordinates of the Gaussian function. In actual image filtering, the center position of the filter is usually taken as (x0, y0). When filtering a certain pixel, this pixel point is equivalent to the center of the current Gaussian function, and the weights of the surrounding pixels are calculated based on it. σ represents the standard deviation, which is a key parameter in Gaussian filtering. The standard deviation determines the distribution shape of the Gaussian function, thereby affecting the filtering effect. A smaller σ value makes the Gaussian function curve steeper, meaning that only pixels very close to the central pixel will have a larger weight. When filtering, mainly the pixels near the central pixel are smoothed, and the image details are better preserved, but the noise reduction effect is relatively weak; a larger σ value makes the Gaussian function curve flatter, and pixels farther from the central pixel will also have a higher weight. When filtering, a larger range of pixels will be smoothed, and the noise reduction effect is obvious, but it may cause the loss of image details. G(x, y) represents the Gaussian weight value at the coordinates (x, y), and this value reflects the importance of the corresponding pixel in the filtering process. The larger the weight value, the greater the influence of the pixel on the filtering result of the central pixel. When calculating the filtered pixel value, the surrounding pixel values are multiplied by the corresponding Gaussian weight values and accumulated.

[0092] When performing Gaussian filtering on the pre-scanned image, for each pixel in the image, a Gaussian template (also called a convolution kernel) is constructed with it as the center. The size of the template is usually odd, such as 3×3, 5×5, etc., which can ensure that there is a clear central pixel. Taking a 3×3 Gaussian template as an example, the calculation process is as follows: Assume that the current pixel to be processed is P(x, y), and its surrounding pixels are P(x - 1, y - 1), P(x - 1, y), P(x - 1, y + 1) …… a total of 9 pixels (including itself). According to the Gaussian filtering formula, the corresponding Gaussian weight values G(x - 1, y - 1), G(x - 1, y) …… of these 9 pixels are calculated. Then multiply the gray value of each pixel by the corresponding Gaussian weight value, that is, P(x - 1, y - 1) × G(x - 1, y - 1), P(x - 1, y) × G(x - 1, y) …… Finally, add these products together, and the result is the new gray value of the pixel P(x, y) after filtering. By performing such operations on each pixel in the image, the Gaussian filtering of the entire image is completed, achieving the purpose of removing noise and smoothing the image.

[0093] After that, contrast enhancement algorithms such as histogram equalization are used to optimize the image contrast and highlight the ankle bone structure to obtain the initial scanned image.

[0094] Step S103: Construct a 3D model based on the initial scanned image.

[0095] Specifically, construct a 3D model through Mimics software based on the initial scanned image.

[0096] On this basis, constructing a 3D model based on the initial scanned image includes:

[0097] Create a new mask based on the initial scanned image, obtain a preset initial mask threshold, and segment the ankle bones and surrounding tissues according to the preset initial mask threshold to obtain an initial segmentation image.

[0098] Create seed points based on the initial segmentation image, obtain the CT value difference between the seed point part and the surrounding tissues according to the seed points, obtain a region growing similarity threshold, and optimize the model edge through erosion and dilation operations according to the CT value difference and the region growing similarity threshold to obtain an optimized image.

[0099] Construct a 3D image according to the optimized image, perform an integrity check on the 3D image, and in response to the 3D image being correct, upload the 3D image to reverse engineering software.

[0100] Obtain preset STL model optimization parameters, and optimize the 3D image through the reverse engineering software to obtain a 3D model.

[0101] Specifically, create a new mask in Mimics software, obtain a preset initial mask threshold, and use the threshold function to segment the ankle bones and surrounding tissues to obtain an initial segmentation image. In this embodiment, the preset initial mask threshold is 220 - 1450 HU.

[0102] Mimics software provides multi - azimuth view functions, such as axial, sagittal, and coronal views. Start from the initial mask threshold and adjust the threshold in steps of +10 HU until the boundary situation of the ankle bones and surrounding tissues in the three - view observation reaches complete fitting separation. In the axial view, determine the boundary range of the bones at this layer according to the cross - sectional shape of the ankle bones; in the sagittal view, determine the structure and boundary of the bones in the front - to - back direction; in the coronal view, determine the boundary of the bones in the left - to - right direction.

[0103] Then, using the region growing function of Mimics software, multiple seed points are selected in the core regions of the bones of the ankle and foot, such as at the central parts of the calcaneus, talus, and navicular bone, to ensure that the internal tissues of the bones are completely selected. The CT value difference between the bone and the surrounding tissues is obtained, and the preset region growing similarity threshold is 10%-20% of the CT value difference. Erosion and dilation in morphological operations are used to clean the edges and optimize the model edges to obtain an optimized image. The erosion operation removes the tiny discontinuous parts and isolated pixels at the mask edges, making the boundary clearer; the dilation operation compensates for the boundary shrinkage that may be caused by erosion to ensure the integrity of the bone boundary. Select appropriate operation kernel sizes and shapes. In this embodiment, a 3×3 square kernel is selected for 2 times of erosion and 2 times of dilation to clean the edges.

[0104] Use the 3D reconstruction function of Mimics software to construct the 3D images of the bones of the ankle and foot, and check the integrity and accuracy of the 3D images. If there are problems such as missing, deformed, or discontinuous parts, return to adjust the relevant parameters.

[0105] After confirmation, export the STL format file, which is the 3D model.

[0106] On this basis, obtain the preset STL model optimization parameters, and optimize the 3D image through the reverse engineering software to obtain a 3D model including:

[0107] Receive the normal direction adjustment information, and adjust the normal direction of the triangular facets according to the normal direction adjustment information to obtain a corrected model;

[0108] Obtain the preset maximum side length threshold and angle difference threshold, and adjust the corrected model according to the preset maximum side length threshold and the angle difference threshold to obtain an adjusted model;

[0109] Based on the adjusted model, perform iterative subdivision through the quadrilateral mesh Loop subdivision algorithm to obtain a precision model;

[0110] Based on the precision model, adjust the smoothing factor and the number of iterations through the smoothing function to optimize the surface quality of the model to obtain a smoothed 3D model;

[0111] Perform defect inspection on the smoothed 3D model. In response to the smoothed 3D model having no defects, output the smoothed 3D model as the 3D model.

[0112] Specifically, import the 3D model STL file into the reverse engineering software, adjust the normal direction of the triangular facets, set the precision parameters of the maximum side length and the angle difference, and balance the model precision and the file size. In this embodiment, the reverse engineering software is 3-Matic software. The present invention is not limited to only this software, and all reverse engineering software that can achieve this function is within the scope of the present invention.

[0113] The normal line is a vector perpendicular to the surface of the triangular patch, which determines the surface direction of the model. The 3-Matic software has the function of automatically adjusting the normal line direction. By analyzing the relationship between adjacent triangular patches, it keeps the normal line direction consistent. If the automatic adjustment effect is not ideal, manually select the patches to generate normal line direction adjustment information, and accurately correct the normal line direction according to the actual geometric structure of the model. For example, in the ankle model, if the normal line directions in some areas are chaotic, the model surface will appear uneven. After manual correction, the smoothness and realism of the model surface are improved.

[0114] Obtain the preset maximum side length threshold and angle difference threshold. The maximum side length limits the maximum size of the triangular patch. When optimizing the ankle model, for the small bone structures in the ankle, such as sesamoid bones, set the maximum side length to 0.8 mm to generate more small patches, accurately depict the model details and improve the accuracy; for relatively smooth areas, such as part of the surface of the calcaneus, set the maximum side length to 1.5 mm to control the file size while ensuring the overall quality of the model.

[0115] The angle difference is used to control the angle change between adjacent triangular patches. When processing the ankle model, for areas that require smooth transition, such as the ankle joint, set the angle difference to 12° to accurately present the physiological shape of the joint; for relatively flat bone surfaces, such as part of the flat area of the navicular bone, set the angle difference to 20°.

[0116] Refine the triangles. Using the Loop subdivision algorithm for quadrilateral meshes, insert new vertices on the original triangle edges to subdivide a triangle into multiple smaller triangles. In the ankle model, apply this algorithm to the overall model and perform 2 iterations of subdivision. Each subdivision increases the number of patches and improves the model accuracy. After 2 iterations, the roughness of the model surface can be significantly improved, enriching the model details, while avoiding excessive refinement that may cause the file to be too large.

[0117] After that, use the smoothing function to optimize the model surface quality by adjusting the smoothing factor and the number of iterations. In the optimization of the ankle model, set the smoothing factor to 0.5 and the number of iterations to 10. When the smoothing factor is 0.5, it can moderately smooth the surface while retaining the key details of the ankle bones; after 10 iterations, it can effectively remove surface defects and unevenness, improving the overall quality of the model. After completing the above operations, carefully check whether the model has geometric problems such as holes and cracks, use the built-in repair tool of the software for repair, and finally export the optimized STL file, which is the three-dimensional model.

[0118] In step S104, perform model verification on the three-dimensional model to obtain the bone three-dimensional model.

[0119] Specifically, use the distance measurement tool to perform model verification on the three-dimensional model. If the verification is passed, the bone three-dimensional model is obtained.

[0120] On this basis, model verification is performed on the three-dimensional model, and the obtained bone three-dimensional model includes:

[0121] Obtain real anatomical structure data, compare the three-dimensional model and the real anatomical structure data according to a preset standard through a distance measurement tool, and in response to the error between the three-dimensional model and the real anatomical structure data being within a preset range, use the three-dimensional model as the bone three-dimensional model;

[0122] In response to the error between the three-dimensional model and the real anatomical structure data being outside the preset range, adjust the three-dimensional model until the error between the three-dimensional model and the real anatomical structure data is within the preset range.

[0123] Specifically, use a distance measurement tool to measure the size of the reconstructed model, compare it with the real ankle bone data, and evaluate the similarity between the model and the real anatomical structure. If the size deviation is large, analyze the reasons (such as scanning errors, incorrect threshold setting, improper model optimization, etc.), and adjust the three-dimensional model parameters accordingly, reprocess the data until the size of the three-dimensional model meets the requirements. In this embodiment, the distance measurement tool is the Hausdorff distance measurement tool.

[0124] On this basis, obtain real anatomical structure data, compare the three-dimensional model and the real anatomical structure data according to a preset standard through a distance measurement tool, and in response to the error between the three-dimensional model and the real anatomical structure data being within a preset range, using the three-dimensional model as the bone three-dimensional model includes:

[0125] Judge whether the overall size error of the model exceeds a preset first allowable range;

[0126] In response to not exceeding the preset first allowable range, judge whether the local feature point size error exceeds a preset second allowable range;

[0127] In response to not exceeding the preset second allowable range, use the three-dimensional model as the bone three-dimensional model.

[0128] Specifically, set the allowable range of the overall size error of the model. For the ankle bone model, the first allowable range of the overall linear size error (such as the length and width of the foot and the key diameters of the ankle joint, etc.) should be controlled within ±1 mm. If the overall size error of the model exceeds the first allowable range, it is determined that the size deviation is large.

[0129] Then, local feature point size error assessment is carried out. For the key feature points of the ankle and foot, such as the calcaneal tuberosity, the head of the talus, the tuberosity of the navicular bone, etc., measure the distance error between them and the corresponding points of the real anatomical structure. The second allowable range of the error of a single key feature point is set to ±0.5 mm. If the distance error of more than 5% of the key feature points exceeds this range, it is also regarded that the model has a large size deviation.

[0130] Model error calibration method, scanning error calibration: If it is determined that the error is caused by the scanning link, re-check the calibration of the CT scanning device. Check whether the scanning parameter settings, such as tube voltage, tube current, pitch, etc., are correct. If it is found that the parameters deviate from the standard values, re-calibrate the parameters according to the device operation manual and re-scan the subject.

[0131] Threshold setting error calibration, review the threshold range set when creating the mask in the Mimics software. If the inaccurate segmentation of the bone area is caused by too wide or too narrow threshold setting, which in turn causes model size deviation, re-determine the appropriate threshold range by carefully observing the bone boundary through multi-directional views according to the CT values of the ankle and foot bones and the surrounding tissues. For example, if the initial threshold range was set to 200 - 1500 HU before, after re-evaluation and adjustment to 220 - 1450 HU, re-perform operations such as mask creation and region growing to construct a new model.

[0132] Improper model optimization calibration: Check the parameter settings for model optimization in the 3-Matic software. If the precision parameters such as the maximum side length and the angle difference are set unreasonably, re-adjust the parameters. For example, if the previously set maximum side length was too large resulting in loss of model details, adjust the maximum side length from 2 mm to 1 mm and re-perform the refinement and smoothing processing of the model. For the algorithms used, such as the Loop subdivision algorithm, if the number of iterations is insufficient or excessive, adjust the number of iterations. For example, if the previous number of iterations was 1 time, adjust it to 2 - 3 times and re-optimize the model until the model size meets the requirements of the evaluation criteria.

[0133] The method for obtaining a three-dimensional model of the human ankle and foot bones by simulated weight-bearing CT scanning provided by the present invention can accurately reflect the bone deformation under physiological load. The bone alignment, joint space and ligament tension of the ankle and foot in the weight-bearing state (such as standing or walking) are significantly different from those in the non-weight-bearing state (such as lying flat). Simulated weight-bearing CT can capture the three-dimensional shape of the bones under real stress and avoid the "non-physiological" data deviation of traditional lying flat CT. It can obtain a high-resolution three-dimensional model of the ankle and foot bones, providing reliable support for medical research, clinical diagnosis and treatment of the ankle and foot, as well as the research and development of medical devices. Compared with traditional methods, the data obtained by the simulated weight-bearing of the present invention is more in line with the actual situation. Advanced technologies and algorithms significantly improve the model quality and promote the development of related research and applications in the medical field.

[0134] It should be noted that the method of the embodiment of the present invention can be executed by a single device, such as a computer or a server. The method of this embodiment can also be applied to a distributed scenario and completed by multiple devices cooperating with each other. In this case of a distributed scenario, one of the multiple devices can only execute one or more steps of the method of the embodiment of the present invention, and these multiple devices will interact with each other to complete the described method.

[0135] It should be noted that some embodiments of the present invention have been described above. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in a different order than in the above embodiments and still achieve the desired result. Additionally, the processes depicted in the figures do not necessarily require the particular order or sequential order shown to achieve the desired result. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0136] Next, the technical solutions of the present invention will be clearly and completely described in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0137] Embodiment 1

[0138] For an adult female weighing 45 kg, 6 elastic bands were tried, and the connection position was adjusted to make the mechanical scale reading reach 45 kg to simulate the stress state of the ankle when the human body is bearing weight, and a three-dimensional model simulation was carried out accordingly.

[0139] First, set the CT scan parameters: adjust the scan parameters according to the foot type, FOV 28 cm × 28 cm, tube voltage 120 kVp, tube current 200 mA, rotation speed 0.5 s / rot, pitch 0.561:1, matrix 512 × 512, collimator width 0.6 × 64 mm.

[0140] Perform threshold adjustment. Preset the initial mask threshold range of 220 - 1450 HU, and fine-tune it to 260 - 1490 HU through multi-directional views to achieve accurate segmentation.

[0141] Set the region growing parameters. Set the similarity threshold to 15%, select seed points at the key bones (such as the center parts of the calcaneus, talus, and navicular bone) to ensure the complete selection of tissues. Select a 3 × 3 square kernel for morphological operations, and perform 2 times of erosion and 2 times of dilation to clean the edges.

[0142] Perform model optimization. In the 3-Matic software, set the maximum side length of triangular facets to 0.8 mm (for small bone structure areas, such as sesamoid bones) and 1.5 mm (for relatively smooth areas, such as part of the surface of the calcaneus), set the angle difference to 12° in joint areas (such as the ankle joint) and 20° for relatively flat bone surfaces (such as the flat area of the navicular bone). Use the Loop subdivision algorithm to subdivide triangles, set the smoothing factor to 0.5, and perform 10 iterations to optimize the surface quality.

[0143] Model verification: Use the Hausdorff distance measurement tool to compare the model with the real data. Control the overall linear dimension error within ±1 mm, and control the error of individual key feature points (such as the calcaneal tuberosity, head of the talus, tuberosity of the navicular bone, etc.) within ±0.5 mm, and ensure that the distance error of no more than 5% of the key feature points exceeds this range.

[0144] As Figure 3 shown, it is a three-dimensional reconstruction diagram of the STL model for bilateral foot and ankle weight-bearing CT examination. As Figure 4 shown, it is the three-view sectional view of the STL model for bilateral foot and ankle weight-bearing CT examination.

[0145] Those of ordinary skill in the art should understand that: The discussion of any of the above embodiments is only exemplary and is not intended to imply that the scope of the present invention (including the claims) is limited to these examples; Under the concept of the present invention, the technical features between the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations in different aspects of the embodiments of the present invention as described above, and they are not provided in detail for the sake of brevity.

[0146] Any process or method description shown in the flowchart or described in other ways herein can be understood as representing a module, segment, or part of code including one or more executable instructions for implementing a specific logical function or process, and the scope of the preferred embodiments of the present invention includes additional implementations, where the functions can be executed in a manner that is not shown or discussed in order, including in a substantially simultaneous manner or in the reverse order according to the functions involved, which should be understood by those skilled in the technical field to which the embodiments of the present invention belong.

[0147] In the case where specific details are set forth to describe exemplary embodiments of the present invention, it will be apparent to those skilled in the art that the present invention can be practiced without these specific details or with variations of these specific details. Therefore, these descriptions should be considered illustrative rather than restrictive. Although the present invention has been described in conjunction with specific embodiments of the present invention, many alternatives, modifications, and variations of these embodiments will be apparent to those of ordinary skill in the art based on the foregoing description.

[0148] Embodiments of the present invention are intended to cover all such substitutions, modifications, and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present invention shall be included within the protection scope of the present invention.

Claims

1. Method for obtaining three-dimensional model of human ankle bone by simulated weight-bearing CT scan, characterized in that, Including: Obtain CT scan data; Upload the CT scan data to medical image processing and 3D reconstruction software to obtain an initial scan image; Construct a 3D model based on the initial scan image; Perform model verification on the 3D model to obtain a bone 3D model.

2. The method for obtaining a three-dimensional model of the human ankle bone by simulated weight-bearing CT scanning according to claim 1, wherein, Before obtaining the CT scan data, it further includes: Obtain the basic body weight data of the user to be measured; Obtain the load-bearing data based on the basic body weight data; Obtain the CT scan data of the user to be measured in a standard body position according to the load-bearing data.

3. The method for obtaining a three-dimensional model of the human ankle bone by simulated weight-bearing CT scanning according to claim 1, wherein The obtaining of the CT scan data includes: Obtain CT scan parameters, and obtain initial scan data according to the CT scan parameters; Perform data reconstruction on the initial scan data to obtain reconstructed data; Perform post-processing operations on the reconstructed data to obtain CT scan data.

4. The method for obtaining a three-dimensional model of the human ankle bone by simulated weight-bearing CT scanning according to claim 3, wherein, The performing of data reconstruction on the initial scan data to obtain reconstructed data includes: Perform data reconstruction on the initial scan data based on the iterative reconstruction MLEM algorithm and the iDose 4 algorithm to obtain a noise-suppressed image; Obtain a preset reconstruction layer thickness and a preset layer spacing, and reduce the volume effect according to the preset reconstruction layer thickness and the preset layer spacing to obtain reconstructed data.

5. The method for obtaining a three-dimensional model of the human ankle bone by simulated weight-bearing CT scanning according to claim 3, characterized in that, The performing of post-processing operations on the reconstructed data to obtain CT scan data includes: Obtain a virtual single-energy keV value, and optimize the image display effect according to the virtual single-energy keV value to obtain CT scan data.

6. The method for obtaining a three-dimensional model of the human ankle bone by simulated weight-bearing CT scanning according to claim 1, characterized in that, The uploading of the CT scan data to medical image processing and 3D reconstruction software to obtain an initial scan image includes: Upload the CT scan data to medical image processing and 3D reconstruction software to obtain a pre-scan image; Perform preprocessing on the pre-scan image through Gaussian filtering to obtain a smoothed scan graph; Optimize the image contrast based on the smoothed scan graph through a contrast enhancement algorithm to obtain an initial scan image.

7. The method for obtaining a three-dimensional model of the human ankle bone by simulated weight-bearing CT scanning according to claim 1, characterized in that, The constructing of a 3D model based on the initial scan image includes: Create a new mask based on the initial scan image, obtain a preset initial mask threshold, and segment the ankle bone and surrounding tissues according to the preset initial mask threshold to obtain an initial segmentation image; Create seed points based on the initial segmentation image, obtain the CT value difference between the seed point part and the surrounding tissues according to the seed points, obtain a region growing similarity threshold, and optimize the model edge through erosion operation and dilation operation according to the CT value difference and the region growing similarity threshold to obtain an optimized image; Construct a 3D image according to the optimized image, perform integrity check on the 3D image, and in response to the 3D image being correct, upload the 3D image to reverse engineering software; Obtain preset STL model optimization parameters, and optimize the 3D image through the reverse engineering software to obtain a 3D model.

8. The method for obtaining a three-dimensional model of the human ankle bone by simulated weight-bearing CT scanning according to claim 1, wherein, The obtaining of preset STL model optimization parameters and optimizing the 3D image through the reverse engineering software to obtain a 3D model includes: Receive normal direction adjustment information, and adjust the normal direction of the triangular facets according to the normal direction adjustment information to obtain a corrected model; Obtain a preset maximum side length threshold and an angle difference threshold, and adjust the correction model according to the preset maximum side length threshold and the angle difference threshold to obtain an adjusted model; Based on the adjusted model, perform iterative subdivision through the quadrilateral mesh Loop subdivision algorithm to obtain a precision model; Based on the precision model, through the fairing function, adjust the smoothing factor and the number of iterations to optimize the surface quality of the model to obtain a smooth three-dimensional model; Perform defect inspection on the smooth three-dimensional model, and in response to the absence of defects in the smooth three-dimensional model, output the smooth three-dimensional model as a three-dimensional model.

9. The method for obtaining a three-dimensional model of the human ankle bone by simulated weight-bearing CT scanning according to claim 1, wherein The model verification of the three-dimensional model to obtain a skeletal three-dimensional model includes: Obtain real anatomical structure data, compare the three-dimensional model and the real anatomical structure data according to a preset standard through a distance measurement tool, and in response to the error between the three-dimensional model and the real anatomical structure data being within a preset range, use the three-dimensional model as a skeletal three-dimensional model; In response to the error between the three-dimensional model and the real anatomical structure data being outside the preset range, adjust the three-dimensional model until the error between the three-dimensional model and the real anatomical structure data is within the preset range.

10. The method for obtaining a three-dimensional model of the human ankle bone by simulated weight-bearing CT scanning according to claim 9, characterized in that The obtaining of the real anatomical structure data, comparing the three-dimensional model and the real anatomical structure data according to a preset standard through a distance measurement tool, and in response to the error between the three-dimensional model and the real anatomical structure data being within a preset range, using the three-dimensional model as a skeletal three-dimensional model includes: Judge whether the overall dimension error of the model exceeds a preset first allowable range; In response to not exceeding the preset first allowable range, judge whether the local feature point dimension error exceeds a preset second allowable range; In response to not exceeding the preset second allowable range, use the three-dimensional model as a skeletal three-dimensional model.

Images