Lower limb projection image generation method and terminal

Abstract

The invention discloses a lower limb projection image generation method and a terminal. The method comprises the following steps: acquiring a full lower limb three-dimensional model, and processing the full lower limb three-dimensional model to obtain a single-side lower limb long bone model and a single-side lower limb three-dimensional model; extracting a surface data point set of the single-side lower limb long bone model, and extracting a projection reference point of the single-side lower limb long bone model from the surface data point set based on bone anatomical features; and determining a standard projection direction of the single-side lower limb three-dimensional model according to the projection reference point, and projecting the single-side lower limb three-dimensional model based on the standard projection direction to obtain a single-side lower limb projection image. According to the invention, the projection direction can be automatically calculated, and the subjectivity of manual adjustment is avoided, so that the accuracy of the projection image is ensured, and the imaging efficiency of the projection image is improved.

Description

Technical Field

[0001] This invention relates to the field of medical imaging technology, and in particular to a method and terminal for generating lower limb projection images. Background Technology

[0002] In the field of medical imaging, whole-lower limb imaging techniques mainly include two-dimensional X-ray imaging and three-dimensional imaging (such as computed tomography (CT) and magnetic resonance imaging (MRI)). Two-dimensional X-ray imaging requires taking anteroposterior and lateral views. The anteroposterior view requires the patella to be centered on the femoral condyle, but the accuracy of the positioning depends on the technician's experience, and errors are easily introduced when the patella is dislocated or the patient's movement is restricted. In addition, the lateral projection of two-dimensional X-ray imaging has the problem of overlapping images of the left and right lower limbs, affecting the clarity. Although three-dimensional imaging can generate two-dimensional projections through maximum intensity projection (MIP) or average intensity projection (AIP), it requires manual adjustment of the projection direction of both lower limbs, which is cumbersome and prone to subjective errors, affecting imaging efficiency and image accuracy. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a method and terminal for generating lower limb projection images, which can ensure the accuracy of the projection images and improve the imaging efficiency of the projection images.

[0004] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A method for generating a projection image of the lower limbs, comprising: A three-dimensional model of the entire lower limb is obtained, and the three-dimensional model of the entire lower limb is processed to obtain a long bone model of one lower limb and a three-dimensional model of one lower limb. Extract the surface data point set of the unilateral lower limb long bone model, and extract the projection reference point of the unilateral lower limb long bone model from the surface data point set based on the skeletal anatomical features; The standard projection direction of the unilateral lower limb 3D model is determined based on the projection reference point, and the unilateral lower limb 3D model is projected based on the standard projection direction to obtain a unilateral lower limb projection image.

[0005] To solve the above-mentioned technical problems, another technical solution adopted by the present invention is as follows: A terminal for generating lower limb projection images includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements each step of the lower limb projection image generation method described above.

[0006] The beneficial effects of this invention are as follows: By segmenting the entire lower limb 3D model, the overall model is decomposed into a unilateral lower limb model, avoiding the problem of overlapping images of the left and right lower limbs. By extracting the surface data point set of the unilateral lower limb long bone model and automatically locating the projection reference point based on skeletal anatomical features, it replaces the traditional method of relying on manual positioning to determine the projection direction. Calculating the standard projection direction based on the projection reference point eliminates subjective judgment errors and ensures that the projection direction is consistent with the spatial relationship of the skeletal anatomical structure, thereby generating standardized lower limb projection images. This invention provides an independent data foundation for subsequent unilateral projection through segmentation; the extraction of the surface data point set combined with the selection of reference points based on anatomical features ensures that the projection direction is aligned with the skeletal physiological structure; at the same time, the automatic calculation of the projection direction based on the projection reference point avoids the inefficiency and subjectivity of manual adjustment, ultimately achieving the automatic generation of standardized projection images without positioning intervention. Attached Figure Description

[0007] Figure 1 This is a flowchart of a method for generating a lower limb projection image according to the present invention; Figure 2 This is a frontal view of the posterior point of the femoral condyle according to the present invention; Figure 3 This is a top view of the posterior point of the femoral condyle according to the present invention; Figure 4 This is the orthogonal projection image of the present invention; Figure 5 This is a side-view projection image of the present invention; Figure 6 This is a schematic diagram of the structure of a terminal for generating lower limb projection images according to the present invention; Label Explanation: 100. Terminal for generating lower limb projection images; 101. Memory; 102. Processor. Detailed Implementation

[0008] To explain in detail the technical content, objectives, and effects of the present invention, the following description is provided in conjunction with the embodiments and accompanying drawings.

[0009] In existing technologies, when using two-dimensional X-ray imaging to project images of the entire lower limb, both anteroposterior and lateral views are typically required. The key quality standard for the anteroposterior view is ensuring the patella is centered between the femoral condyles, a standard highly dependent on the patient's lower limb positioning. However, the accuracy of positioning is limited by the technician's experience, often requiring repeated adjustments, increasing operational complexity. Furthermore, in cases of patellar dislocation, forcibly placing the patella in a centered position may lead to external or internal rotation of the lower limb, negatively impacting the accuracy of measurements. For severely mobile patients, adjusting the lower limb to a standard position is difficult, further increasing the difficulty of imaging. In addition, two-dimensional X-ray imaging suffers from overlapping images of the left and right lower limbs, especially in lateral projection, where tissues from both lower limbs in a normal standing position are easily blurred, leading to decreased image clarity and affecting diagnostic results. Although three-dimensional imaging technology eliminates the need to distinguish the projection directions of the left and right lower limbs, the projection directions of both lower limbs must be manually adjusted separately to ensure they are unaffected by positioning and to maintain a unified projection direction. This process not only increases the workload and affects imaging efficiency, but also introduces subjective errors due to the reliance on subjective judgment for projection direction, thus affecting the accuracy of the projected image.

[0010] To address the above problems, the present invention provides a method for generating lower limb projection images, comprising: A three-dimensional model of the entire lower limb is obtained, and the three-dimensional model of the entire lower limb is processed to obtain a long bone model of one lower limb and a three-dimensional model of one lower limb. Extract the surface data point set of the unilateral lower limb long bone model, and extract the projection reference point of the unilateral lower limb long bone model from the surface data point set based on the skeletal anatomical features; The standard projection direction of the unilateral lower limb 3D model is determined based on the projection reference point, and the unilateral lower limb 3D model is projected based on the standard projection direction to obtain a unilateral lower limb projection image.

[0011] As described above, the beneficial effects of this invention are as follows: By segmenting the entire lower limb 3D model, the overall model is decomposed into a unilateral lower limb model, avoiding the problem of overlapping images of the left and right lower limbs. By extracting the surface data point set of the unilateral lower limb long bone model and automatically locating the projection reference point based on skeletal anatomical features, the traditional method of relying on manual positioning to determine the projection direction is replaced. Calculating the standard projection direction based on the projection reference point can eliminate subjective judgment errors and ensure that the projection direction is consistent with the spatial relationship of the skeletal anatomical structure, thereby generating a standardized lower limb projection image. This invention provides an independent data foundation for subsequent unilateral projection through segmentation processing; the extraction of the surface data point set combined with the selection of reference points based on anatomical features ensures that the projection direction is aligned with the skeletal physiological structure; at the same time, the automatic calculation of the projection direction based on the projection reference point avoids the inefficiency and subjectivity of manual adjustment, ultimately achieving the automatic generation of standardized projection images without positioning intervention.

[0012] Furthermore, the unilateral lower limb long bone model includes a unilateral femur model; The projection reference points include the posterior point of the lateral femoral condyle and the posterior point of the medial femoral condyle; Based on skeletal anatomical features, the projection reference points of the unilateral lower limb long bone model extracted from the surface data point set include: Based on the spatial distribution characteristics of the surface data point set and the skeletal anatomical characteristics, the posterior points of the lateral and medial femoral condyles of the femur are extracted from the unilateral femoral model.

[0013] As described above, the femur, as the largest long bone of the lower limb, possesses significant and stable morphological features in its distal condyle, providing a reliable anatomical basis for selecting projection reference points. Specifically, the projection reference points are the posterior points of the lateral and medial femoral condyles. These two bony landmarks have clear anatomical significance and can form stable spatial connections in three-dimensional space. By extracting feature points based on the spatial distribution characteristics of the surface data point set, the subjectivity of traditional manual identification can be effectively overcome, and the automatic identification of anatomical landmarks can be achieved using the spatial topological relationships of the surface points of the three-dimensional model. Furthermore, feature extraction based on the spatial distribution characteristics of the surface data point set can adapt to the differences in skeletal morphology among individuals, automatically identifying key anatomical points through mathematical methods, avoiding errors caused by manual intervention, and providing accurate reference point data for the subsequent automatic calculation of the projection direction.

[0014] Further, extracting the posterior points of the lateral and medial femoral condyles from the unilateral femoral model based on the spatial distribution characteristics and skeletal anatomical features of the surface data point set includes: Based on the first spatial distribution characteristics of the surface data point set in the human body height direction, the surface data points located at the distal end of the unilateral femur model are marked as the knee joint data point set. Based on the second spatial distribution characteristics of the knee joint data point set in the anterior-posterior direction of the human body, the knee joint data point closest to the posterior side of the human body is marked as the first femoral condyle posterior point. Calculate the distance between the first femoral condyle posterior point and each knee joint data point in the knee joint data point set, and mark the knee joint data points whose distance is greater than a preset threshold as a candidate femoral condyle posterior point set; Based on the second spatial distribution characteristics of the candidate femoral condyle posterior point set, the candidate femoral condyle posterior point closest to the posterior side of the human body is marked as the second femoral condyle posterior point; The lateral and medial femoral condyles are determined based on the third spatial distribution characteristics of the first and second femoral condyles in the left-right direction of the human body.

[0015] As described above, precise localization of anatomical landmarks is achieved through multi-dimensional spatial feature analysis. First, data points in the knee joint region are selected based on the spatial distribution characteristics of the human body's height, effectively eliminating interference from non-target data points. Then, the spatial distribution characteristics of the human body's anterior-posterior direction are used to pinpoint the first femoral condyle posterior point closest to the posterior end, establishing an initial localization benchmark. A candidate point set is then filtered using a preset distance threshold to exclude adjacent points on the same condyle. Next, based on the anterior-posterior characteristics of the human body, the second femoral condyle posterior point is determined from the candidate set, ensuring the anatomical distance between the two posterior condyles. Finally, the spatial distribution characteristics in the left-right direction are combined to accurately distinguish the medial and lateral posterior condyles. This phased, multi-dimensional feature selection mechanism, through hierarchical progressive analysis of spatial distribution characteristics, overcomes the subjectivity and error accumulation problems of traditional manual localization, providing a precise anatomical benchmark for the subsequent automatic determination of the projection direction.

[0016] Furthermore, the standard projection direction includes a side projection direction and a frontal projection direction; Determining the standard projection direction of the unilateral lower limb 3D model based on the projection reference point includes: The lateral projection direction of the unilateral lower limb three-dimensional model is determined based on the direction of the line connecting the posterior point of the lateral femoral condyle and the posterior point of the medial femoral condyle. The orthogonal projection direction of the unilateral lower limb 3D model is determined based on the lateral projection direction and the preset vertical direction.

[0017] As described above, the lateral projection direction is determined by the line connecting the posterior points of the lateral and medial femoral condyles, utilizing the anatomical symmetry of the femoral condyles as a spatial reference to ensure alignment of the projection direction with the actual anatomical axis of the human body. The anterior projection direction is determined by pre-setting the vector orthogonality between the vertical direction and the lateral projection direction, combining the objective physical reference of gravity with the anatomical reference to construct an orthogonal projection coordinate system that conforms to medical imaging standards. This composite projection direction determination method based on the spatial relationship of anatomical landmarks and physical references avoids subjective errors from manual adjustments and ensures the standardization and repeatability of the projection direction in three-dimensional space.

[0018] Furthermore, the unilateral lower limb projection image includes a lateral projection image and an orthogonal projection image; The unilateral lower limb 3D model includes a unilateral lower limb element model; Projecting the unilateral lower limb 3D model based on the standard projection direction to obtain a unilateral lower limb projection image includes: Calculate the average density of voxel units in the lateral projection direction and the frontal projection direction of the unilateral lower limb voxel model, respectively. The lateral projection image and the orthogonal projection image are generated based on the average density values ​​of the lateral projection direction and the orthogonal projection direction, respectively.

[0019] As described above, the 3D model of a unilateral lower limb is a voxel model, which can directly utilize the 3D spatial distribution characteristics of voxel data to provide structured data support for subsequent projection calculations. By calculating the average density values ​​of voxel units under two standard projection directions—lateral and anteroposterior—a standard projection parameter system based on anatomical features was established, eliminating the need for repeated manual adjustments to the projection direction. Corresponding projection images are generated based on the average density values ​​of different projection directions, preserving the anatomical structure information of the 3D model while effectively eliminating voxel data noise through density averaging. This ensures that the generated 2D projection images meet both the standardized requirements of clinical diagnosis and possess good image quality. The coordinated calculation of the lateral and anteroposterior projection directions enables automatic matching of multi-angle projection parameters, solving the workload problem of having to process different projection directions separately in traditional methods.

[0020] Furthermore, processing the full lower limb 3D model to obtain a unilateral lower limb long bone model and a unilateral lower limb 3D model includes: The three-dimensional model of the entire lower limb was segmented into voxels to obtain a voxel model of the long bones of the entire lower limb. Extract isosurface patches from the full lower limb long bone voxel model, and construct a full lower limb long bone mesh model based on the isosurface patches; A connected component is constructed based on the grid data points in the aforementioned full lower limb long bone mesh model; The full lower limb long bone mesh model is divided into a single lower limb long bone model on the left side of the human body and a single lower limb long bone model on the right side of the human body based on the center position of the connected domain. Calculate the minimum bounding box of the unilateral lower limb long bone model, and extract the unilateral lower limb 3D model corresponding to the unilateral lower limb long bone model from the full lower limb 3D model based on the minimum bounding box.

[0021] As described above, the entire lower limb 3D model is segmented into voxels to accurately obtain the long bone voxel model. A mesh model is generated by constructing isosurface patches, providing a precise geometric foundation for subsequent connected component analysis. Left and right lateral segmentation is performed based on the center position of the connected components, utilizing anatomical symmetry features to achieve accurate lateralization. Finally, minimum bounding box matching of the original 3D model ensures the spatial correspondence between the unilateral lower limb 3D model and the long bone model, avoiding segmentation errors. The application of the pre-trained model improves segmentation speed, the combination of the mesh model and connected components enhances segmentation robustness, and the geometric feature matching of the minimum bounding box guarantees the anatomical accuracy of the segmentation results.

[0022] Further, extracting the surface data point set of the unilateral lower limb long bone model includes: The number of voxels in each connected region is calculated based on the unilateral lower limb long bone model and the full lower limb long bone voxel model. Connected components whose number of voxels is greater than a preset voxel threshold are marked as valid connected components; Connected regions whose number of voxels is less than or equal to the preset voxel threshold are marked as noise; The surface data point set is obtained by extracting all grid data points within the effective connected domain.

[0023] As described above, firstly, the number of voxels contained in each connected region is calculated by combining the unilateral lower limb long bone model and the whole lower limb long bone voxel model, and the effectiveness of each connected region is objectively evaluated through quantitative indicators. Secondly, binary classification is performed by setting a preset voxel threshold, and regions with a sufficient voxel size are marked as effective connected regions, excluding noise regions with sparse voxels. Finally, only the grid data points within the effective connected regions are extracted to form a surface data point set, ensuring that the geometric feature data upon which the subsequent projection reference point extraction depends all originates from effective anatomical structure regions, avoiding interference from noise points on the localization of key feature points.

[0024] Further, determining the lateral projection direction of the unilateral lower limb three-dimensional model based on the direction of the line connecting the posterior point of the lateral femoral condyle and the posterior point of the medial femoral condyle includes: Calculate the standard vector from the posterior point of the medial femoral condyle to the posterior point of the lateral femoral condyle, and normalize the standard vector to obtain a standard unit vector; The standard unit vector is determined as the lateral projection direction of the unilateral lower limb 3D model.

[0025] As described above, a standard vector is first generated based on the direction from the posterior point of the medial femoral condyle to the posterior point of the lateral femoral condyle. This vector directly reflects the physiological structural characteristics of the human femoral condyle and can accurately characterize the horizontal reference direction required for lateral projection. By normalizing the standard vector, the influence of differences in lower limb size among different patients on the projection direction is eliminated, ensuring the consistency of projection direction calculation among different individuals. Finally, the normalized standard unit vector is directly used as the lateral projection direction, realizing fully automatic direction determination based on anatomical feature points, avoiding operational errors caused by manual intervention, and significantly improving the efficiency of projection image generation.

[0026] Further, determining the orthogonal projection direction of the unilateral lower limb 3D model based on the lateral projection direction and the preset vertical direction includes: Obtain the vertical vector corresponding to the preset vertical direction; Calculate the target vector that is perpendicular to the standard unit vector and the vertical vector, and normalize the target vector to obtain the target unit vector; The target unit vector is determined as the orthogonal projection direction of the unilateral lower limb 3D model.

[0027] As described above, obtaining the plumb line vector transforms the standard anatomical position of the human body into a mathematical vector reference, establishing a connection between the projection direction and the direction of gravity. A target vector perpendicular to the standard unit vector and the plumb line vector is calculated. The orthogonality of vectors ensures that the orthogonal projection direction, lateral projection direction, and the plumb line form a spatial rectangular coordinate system, eliminating directional deviations. The target vector is normalized to generate a unit vector, guaranteeing the standardization of the projection direction parameters. Finally, the target unit vector is directly used as the orthogonal projection direction. The deterministic nature of vector operations replaces subjective human judgment, achieving automated calculation of the projection direction parameters. This scheme, through the mathematical method of spatial vector orthogonal decomposition, transforms the standard medical imaging positioning requirements into calculable geometric constraints, generating an orthogonal projection direction that conforms to clinical standards without human intervention.

[0028] The present invention also provides a terminal for generating lower limb projection images, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the various steps in the lower limb projection image generation method described above.

[0029] As described above, the beneficial effects of this invention are as follows: By segmenting the entire lower limb 3D model, the overall model is decomposed into a unilateral lower limb model, avoiding the problem of overlapping images of the left and right lower limbs. By extracting the surface data point set of the unilateral lower limb long bone model and automatically locating the projection reference point based on skeletal anatomical features, the traditional method of relying on manual positioning to determine the projection direction is replaced. Calculating the standard projection direction based on the projection reference point can eliminate subjective judgment errors and ensure that the projection direction is consistent with the spatial relationship of the skeletal anatomical structure, thereby generating a standardized lower limb projection image. This invention provides an independent data foundation for subsequent unilateral projection through segmentation processing; the extraction of the surface data point set combined with the selection of reference points based on anatomical features ensures that the projection direction is aligned with the skeletal physiological structure; at the same time, the automatic calculation of the projection direction based on the projection reference point avoids the inefficiency and subjectivity of manual adjustment, ultimately achieving the automatic generation of standardized projection images without positioning intervention.

[0030] The embodiments of the present invention provide a method and terminal for generating lower limb projection images applicable to the field of medical imaging. It can automatically determine the projection direction, avoiding the subjectivity of manual adjustment, thereby ensuring the accuracy of the projection image and improving the imaging efficiency. Specific embodiments are described below: Please refer to Figures 1 to 5One embodiment of the present invention is as follows: like Figure 1 As shown, a method for generating a projection image of the lower limb includes steps S1-S3.

[0031] S1. Obtain the full lower limb 3D model and process the full lower limb 3D model to obtain the unilateral lower limb long bone model and the unilateral lower limb 3D model.

[0032] The full lower limb 3D model refers to a 3D data model reconstructed from CT or MRI scans, containing the bones and soft tissues of both lower limbs. It can be stored using voxel meshes or curved meshes, and its purpose is to provide a complete anatomical data foundation for segmentation. The unilateral lower limb long bone model refers to a 3D data model containing only the bones of one lower limb. The unilateral lower limb 3D model refers to a 3D data model containing only the bones and soft tissues of one lower limb.

[0033] S2. Extract the surface data point set of the unilateral lower limb long bone model, and extract the projection reference point of the unilateral lower limb long bone model from the surface data point set based on the skeletal anatomical features.

[0034] The surface data point set is a collection of three-dimensional coordinates of the surface of a unilateral lower limb long bone model, specifically obtained through isosurface extraction algorithms. Its function is to provide spatial data for anatomical feature analysis. Skeletal anatomical features refer to the typical landmarks or characteristics of bones in terms of morphology, structure, location, and relationship with other tissues. The unilateral lower limb long bone model includes a unilateral femur model. The main long bones of the human lower limb include the femur and tibia. As the largest and most morphologically distinctive long bone of the lower limb, the femur's condyle has a stable geometric structure in three-dimensional space. Projection reference points refer to key landmarks selected based on skeletal anatomical features. In this embodiment, the projection reference points include the posterior point of the lateral femoral condyle and the posterior point of the medial femoral condyle, which provide geometric reference for the projection direction.

[0035] Specifically, in step S2, extracting the projection reference points of the unilateral lower limb long bone model from the surface data point set based on skeletal anatomical features includes step S250.

[0036] S250. Based on the spatial distribution characteristics of the surface data point set and the skeletal anatomical characteristics, extract the posterior points of the lateral and medial femoral condyles from the unilateral femoral model.

[0037] In step S2, the surface data point set is a three-dimensional coordinate set of the surface of a unilateral femoral model. The spatial distribution characteristics of the surface data point set refer to the geometric distribution pattern of the vertices of the three-dimensional model surface in the spatial coordinate system. Specifically, the position of the vertices posterior to the medial and lateral condyles can be accurately located by analyzing the distribution density and curvature changes of the surface data point set in the coronal, sagittal, and transverse planes. The coronal plane is the plane that divides the human body in the anterior-posterior direction (anterior-posterior), the sagittal plane is the plane that divides the human body in the lateral direction (left-right), and the transverse plane is the plane that divides the human body in the horizontal direction (up-down). Figure 2 and Figure 3 As shown, the posterior femoral condyle point refers to the most posterior protruding point on the distal femur (femoral condyle); the posterior femoral condyle point refers to the most posteriorly protruding bony landmark of the lateral femoral condyle in the sagittal plane, located on the lateral side of the knee joint; the posterior femoral condyle point refers to the most posteriorly protruding bony landmark of the medial femoral condyle in the sagittal plane, located on the medial side of the knee joint.

[0038] In one alternative implementation, step S250 includes steps S2501-S2505.

[0039] S2501. Based on the first spatial distribution characteristics of the surface data point set in the human body height direction, the surface data points located at the distal end of the unilateral femur model are marked as the knee joint data point set.

[0040] In this model, the distal end of the unilateral femur refers to the end furthest from the center of the body, i.e., the lower part of the femur. The first spatial distribution feature refers to the distribution of coordinate points along the body's height axis. Specifically, the knee joint region can be quickly located by comparing the coordinate values ​​along the body's height axis and filtering the set of points in the distal region.

[0041] S2502. Based on the second spatial distribution characteristics of the knee joint data point set in the anterior-posterior direction of the human body, the knee joint data point closest to the posterior side of the human body is marked as the posterior point of the first femoral condyle.

[0042] The second spatial distribution feature refers to the distribution of coordinate points along the front-back direction of the human body. Specifically, the initial position of the posterior condyle point can be determined by comparing the coordinate values ​​in the front-back direction of the human body.

[0043] S2503. Calculate the distance between the first femoral condyle posterior point and each knee joint data point in the knee joint data point set, and mark the knee joint data points with a distance greater than a preset threshold as candidate femoral condyle posterior point sets.

[0044] The preset threshold refers to the anatomically defined distance between the medial and lateral femoral condyles. By calculating the distance between the posterior point of the first femoral condyle and all knee joint data points, data points that conform to the distance range between the medial and lateral femoral condyles are selected (i.e., adjacent points located on the same condyle are excluded), thus ensuring that the two posterior points of the femoral condyles originate from different condyles.

[0045] S2504. Based on the second spatial distribution characteristics of the candidate femoral condyle posterior point set, the candidate femoral condyle posterior point closest to the back of the human body is marked as the second femoral condyle posterior point.

[0046] S2505. Determine the lateral femoral condyle posterior point and the medial femoral condyle posterior point based on the third spatial distribution characteristics of the first and second femoral condyles in the left-right direction of the human body.

[0047] Among them, the third spatial distribution feature refers to the distribution of coordinate points along the left and right directions of the human body. Specifically, the posterior points of the femoral condyles on the medial and lateral sides can be distinguished by comparing the coordinate values ​​in the left and right directions of the human body.

[0048] In a specific application scenario, taking a standing right femur model as an example, the three-dimensional coordinate system of the right femur model is based on the front-back direction of the human body as the Y-axis, the height direction of the human body as the Z-axis, and the left-right direction of the human body as the X-axis. The positive direction of the Y-axis is the front of the human body, the positive direction of the Z-axis is the top of the human body, and the positive direction of the X-axis is the left side of the human body. The coordinates of the i-th data point in the surface data point set D0 are [x i ,y i ,z i At this point, the specific process of step S250 includes step ae.

[0049] Step a: Obtain the maximum value of the Z-axis in the surface data point set D0 as the maximum height in the right femur model, i.e., the maximum height H = max z. i The minimum Z-axis value obtained from the surface data point set is the minimum height in the right femur model, i.e., the minimum height h = min z. i ;where [x i ,y i ,z i ] The Z-axis coordinate values ​​must satisfy the condition z i Surface data points with a value less than (h+(Hh) / 5) are labeled as knee joint data points, resulting in knee joint data point set D1. Condition z i <(h+(Hh) / 5) indicates that the Z-axis coordinate value is below 1 / 5 of the total femoral height (Hh) calculated from the lowest point (h) upwards.

[0050] Step b: Mark the knee joint data point with the smallest Y-axis coordinate value in the knee joint data point set D1 as the posterior point of the first femoral condyle P1. That is, if the Y-axis coordinate value of the k-th data point is equal to argmin y k , where [x k ,y k ,z k ] argmin y represents finding the value that minimizes the Y-axis coordinate value, and argmin y represents the k-th data point as the posterior point P1 of the first femoral condyle.

[0051] Step c: Calculate the distance from the j-th knee joint data point in the knee joint data point set D1 to the posterior point P1 of the first femoral condyle. , where [x j ,y j ,z j ] When D j When the preset threshold D' is reached, the j-th knee joint data point belongs to the candidate femoral condyle posterior point set D2.

[0052] Step d: Mark the candidate posterior femoral condyle point with the smallest Y-axis coordinate value in the candidate posterior femoral condyle point set D2 as the second posterior femoral condyle point P2. That is, if the Y-axis coordinate value of the m-th data point is equal to argmin y m , where [x m ,y m ,z m ] If , then it means that the m-th data point is the posterior point P2 of the second femoral condyle.

[0053] Step e: When the X-axis coordinate value of the posterior point P1 of the first femoral condyle is greater than the X-axis coordinate value of the posterior point P2 of the second femoral condyle, i.e., x k >x m At that time, the first femoral condyle posterior point P1 is the posterior point of the medial femoral condyle, and the second femoral condyle posterior point P2 is the posterior point of the lateral femoral condyle; when the X-axis coordinate value of the first femoral condyle posterior point P1 is less than the X-axis coordinate value of the second femoral condyle posterior point P2, that is, x k <x m At that time, the first femoral condyle posterior point P1 is the posterior point of the lateral femoral condyle, and the second femoral condyle posterior point P2 is the posterior point of the medial femoral condyle.

[0054] S3. Determine the standard projection direction of the 3D model of the unilateral lower limb based on the projection reference point, and project the 3D model of the unilateral lower limb based on the standard projection direction to obtain the projection image of the unilateral lower limb.

[0055] The standard projection direction includes lateral projection and orthographic projection. The lateral projection direction is the projection axis perpendicular to the sagittal plane, used to eliminate projection deviations caused by lower limb rotation. The orthographic projection direction is the projection axis perpendicular to the coronal plane, ensuring that the projection direction conforms to the human standing posture.

[0056] Specifically, in step S3, determining the standard projection direction of the unilateral lower limb 3D model based on the projection reference point includes steps S310-S320.

[0057] S310. Determine the lateral projection direction of the unilateral lower limb 3D model based on the direction of the line connecting the posterior point of the lateral femoral condyle and the posterior point of the medial femoral condyle.

[0058] In one optional implementation, step S310 includes: calculating a standard vector pointing from the posterior point of the medial femoral condyle to the posterior point of the lateral femoral condyle, normalizing the standard vector to obtain a standard unit vector, and determining the standard unit vector as the lateral projection direction of the unilateral lower limb three-dimensional model.

[0059] In a specific application scenario, taking the calculation of the lateral projection direction of a 3D model of the right lower limb as an example, the coordinates of the posterior point of the lateral femoral condyle are [x1, y1, z1], and the coordinates of the posterior point of the medial femoral condyle are [x2, y2, z2]. In this case, the lateral projection direction is a standard unit vector. : .

[0060] S320. Determine the orthogonal projection direction of the unilateral lower limb 3D model based on the lateral projection direction and the preset vertical direction.

[0061] The vertical direction refers to the vector representation of the direction of gravity in three-dimensional space. Specifically, it can be represented by the vertical axis vector in a three-dimensional coordinate system, serving as the physical reference for calculating the orthogonal projection direction.

[0062] In an optional implementation, step S320 includes: obtaining a vertical vector corresponding to a preset vertical direction, calculating a target vector perpendicular to the standard unit vector and the vertical vector, normalizing the target vector to obtain a target unit vector, and determining the target unit vector as the orthogonal projection direction of the unilateral lower limb 3D model.

[0063] In a specific application scenario, taking the calculation of the orthogonal projection direction of a 3D model of the right lower limb as an example, the vertical vector corresponding to the vertical direction is: The standard unit vector is At this point, the orthogonal projection direction is the target unit vector. : .

[0064] The unilateral lower limb projection image includes lateral projection image and frontal projection image; the unilateral lower limb 3D model includes the unilateral lower limb pixel model, and the single-layer lower limb pixel model is a unilateral lower limb model built based on voxel units.

[0065] In one optional implementation, step S3, which involves projecting a unilateral lower limb 3D model onto a standard projection direction to obtain a unilateral lower limb projection image, includes steps S330-S340.

[0066] S330. Calculate the average density of voxel units in the lateral and orthogonal projection directions of the unilateral lower limb voxel model.

[0067] S340. Generate lateral projection images and orthographic projection images based on the average density values ​​of the lateral projection direction and the orthographic projection direction, respectively.

[0068] In a specific application scenario, the projected image is specifically represented as ;in,( u,v ) represents the pixel coordinates of the projected image. AIP ( u,v ) represents the projected image at pixel coordinates ( u,v The pixel value at () G ( x,y,z ) indicates that the unilateral lower limb model is in ( x,y,z Density values ​​of voxel units on coordinates, operators Mean This represents the average density value of the voxel unit along the standard projection direction. Therefore, the coordinates of each pixel on the projection plane ( u,v ) corresponds only to a line parallel to the standard projection direction and passing through ( u,v A ray (i.e., projection line) extends in three-dimensional space, traversing the entire voxel model of the unilateral lower limb. Spatial sampling is performed on the voxel units in the unilateral lower limb model along this ray, discretizing the continuous ray into a series of sampling points (x, y, z), thereby realizing the calculation of the average density value.

[0069] In another alternative implementation, a projected image is generated by calculating the maximum or minimum density value of voxel units in a unilateral lower limb voxel model under a standard projection direction.

[0070] like Figure 4 and Figure 5 As shown, the above method for generating lower limb projection images can be used to obtain both frontal and lateral projection images of the right lower limb.

[0071] In summary, this invention generates unilateral lower limb projection images from a full lower limb 3D model and automatically calculates the standard projection directions for both lower limbs, achieving fully automated projection, which is more efficient than manually adjusting the projection direction. Furthermore, this invention determines the standard projection direction by identifying the posterior points of the medial and lateral femoral condyles, unaffected by the patella, making it suitable for severe patients with patellar dislocation or limited lower limb movement. This ensures that the projection image quality is unaffected by patient positioning, solving the problem of high positioning requirements in previous full lower limb anteroposterior radiographs, which were unsuitable for severe patients. In addition, this invention can automatically determine a unilateral lower limb model and project only one lower limb based on the model, solving the problem of overlapping images of the left and right lower limbs in previous full lower limb lateral radiographs.

[0072] Another embodiment of the present invention is as follows: A method for generating a projection image of the lower limb, which differs from Embodiment 1 in that it specifies the specific implementation of extracting the surface data point set of the unilateral lower limb long bone model in steps S1 and S2.

[0073] In this embodiment, the obtained full lower limb three-dimensional model is a CT three-dimensional model.

[0074] Specifically, in step S1, processing the full lower limb three-dimensional model to obtain a unilateral lower limb long bone model and a unilateral lower limb three-dimensional model includes steps S110-S150.

[0075] S110. Perform voxel segmentation on the three-dimensional model of the entire lower limb to obtain the voxel model of the long bones of the entire lower limb.

[0076] In one optional implementation, a pre-trained deep learning segmentation model is used to perform voxel segmentation on the entire lower limb 3D model. Specifically, in a particular application scenario, the deep learning segmentation model classifies each voxel unit in the entire lower limb 3D model according to its region. The classification result for each voxel unit is assigned a value of 0, 1, or 2, where 1 represents the femur region, 2 represents the tibia region, and 0 represents other regions. The resulting segmentation is a full lower limb long bone voxel model containing femur and tibia voxel models.

[0077] In another alternative implementation, the voxel model of the entire lower limb is obtained by voxel segmentation of the three-dimensional model of the entire lower limb using the threshold segmentation method.

[0078] S120. Extract isosurface patches from the voxel model of the entire lower limb long bones, and construct a mesh model of the entire lower limb long bones based on the isosurface patches.

[0079] In a specific application scenario, assuming the threshold for the isosurface to be extracted is 0.5, multiple 2×2×2 cube voxels are constructed with each voxel unit in the full lower limb long bone voxel model as the center. The voxel values ​​at each vertex of the voxel are compared with the given threshold. First, the voxels that intersect with the isosurface are found. Then, the intersection points of the isosurface and the edges of the voxels are found by interpolation. Finally, the intersection points are connected to form triangles to form isosurface patches. The set of triangles in all voxels constitutes the full lower limb long bone mesh model.

[0080] In one alternative implementation, a model registration method can be used to construct a mesh model of the entire lower limb long bones.

[0081] S130. Construct connected components based on grid data points in the whole lower limb long bone mesh model.

[0082] In a specific application scenario, surface data points that have direct or indirect connections to each other are marked as the same connected component, ultimately resulting in multiple connected components that are not connected to each other.

[0083] S140. Based on the center position of the connected domain, the whole lower limb long bone mesh model is divided into a single lower limb long bone model on the left side of the human body and a single lower limb long bone model on the right side of the human body.

[0084] In one optional implementation, step S140 specifically involves: classifying each voxel unit within the full lower limb long bone voxel model into two categories based on whether it is located inside the connected component; counting the number of voxel units located inside the connected component, which is the volume of the connected component. A volume threshold is set, and connected components exceeding the threshold are identified as connected components of the lower limb long bones. At this point, the unilateral lower limb long bone model can be determined by comparing the coordinates of the center position of the connected component in the left-right direction of the human body. In a specific application scenario, if the center position of the first connected component is represented as (x... c1 ,y c1 ,z c1 The center position of the second connected component is represented as (x). c2 ,y c2 ,z c2 ), then when x c1 >x c2 When x is the first connected component, the first connected component is the long bone model of the left lower limb, and the second connected component is the long bone model of the right lower limb; when x c1 <x c2 At that time, the first connected component is the long bone model of the right lower limb, and the second connected component is the long bone model of the left lower limb.

[0085] S150. Calculate the minimum bounding box of the unilateral lower limb long bone model, and extract the unilateral lower limb 3D model corresponding to the unilateral lower limb long bone model from the full lower limb 3D model based on the minimum bounding box.

[0086] In a specific application scenario, the minimum bounding boxes of the left femur mesh model and the left tibia mesh model are calculated. After appropriately expanding the range, the left lower limb mesh model is obtained, ensuring that all voxel units of the left lower limb bones and soft tissues are included in the left lower limb mesh model. The same principle applies to the right lower limb mesh model, which will not be elaborated here.

[0087] Specifically, in step S2, extracting the surface data point set of the unilateral lower limb long bone model includes steps S210-S240.

[0088] S210. Calculate the number of voxels in each connected region based on the unilateral lower limb long bone model and the full lower limb long bone voxel model.

[0089] S220. Mark connected components with a voxel count greater than a preset voxel threshold as valid connected components.

[0090] S230. Mark connected components with a number of voxels less than or equal to a preset voxel threshold as noise.

[0091] S240. Extract all grid data points within the effective connected domain to obtain the surface data point set.

[0092] In an optional implementation, the method for identifying valid connected components in steps S210-S220 is the same as the method for identifying connected components of the long bones of the lower limb in step S140, and will not be described again here.

[0093] Please refer to Figure 6 Another embodiment of the present invention is as follows: A lower limb projection image generation terminal 100 includes a memory 101, a processor 102, and a computer program stored in the memory 101 and running on the processor 102. When the processor 102 executes the computer program, it implements each step of the lower limb projection image generation method of the above embodiments.

[0094] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent modifications made based on the content of the present invention specification and drawings, or direct or indirect applications in related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for generating a projection image of the lower limbs, characterized in that, include: A three-dimensional model of the entire lower limb is obtained, and the three-dimensional model of the entire lower limb is processed to obtain a long bone model of one lower limb and a three-dimensional model of one lower limb. Extract the surface data point set of the unilateral lower limb long bone model, and extract the projection reference point of the unilateral lower limb long bone model from the surface data point set based on the skeletal anatomical features; The standard projection direction of the unilateral lower limb 3D model is determined based on the projection reference point, and the unilateral lower limb 3D model is projected based on the standard projection direction to obtain a unilateral lower limb projection image.

2. The method for generating lower limb projection images according to claim 1, characterized in that, The unilateral lower limb long bone model includes a unilateral femur model; The projection reference points include the posterior point of the lateral femoral condyle and the posterior point of the medial femoral condyle; Based on skeletal anatomical features, the projection reference points of the unilateral lower limb long bone model extracted from the surface data point set include: Based on the spatial distribution characteristics of the surface data point set and the skeletal anatomical characteristics, the posterior points of the lateral and medial femoral condyles of the femur are extracted from the unilateral femoral model.

3. The method for generating lower limb projection images according to claim 2, characterized in that, Based on the spatial distribution characteristics and skeletal anatomical features of the surface data point set, the posterior points of the lateral and medial femoral condyles of the femur are extracted from the unilateral femoral model, including: Based on the first spatial distribution characteristics of the surface data point set in the human body height direction, the surface data points located at the distal end of the unilateral femur model are marked as the knee joint data point set. Based on the second spatial distribution characteristics of the knee joint data point set in the anterior-posterior direction of the human body, the knee joint data point closest to the posterior side of the human body is marked as the first femoral condyle posterior point. Calculate the distance between the first femoral condyle posterior point and each knee joint data point in the knee joint data point set, and mark the knee joint data points whose distance is greater than a preset threshold as a candidate femoral condyle posterior point set; Based on the second spatial distribution characteristics of the candidate femoral condyle posterior point set, the candidate femoral condyle posterior point closest to the posterior side of the human body is marked as the second femoral condyle posterior point; The lateral and medial femoral condyles are determined based on the third spatial distribution characteristics of the first and second femoral condyles in the left-right direction of the human body.

4. The method for generating lower limb projection images according to claim 2, characterized in that, The standard projection direction includes the side projection direction and the frontal projection direction; Determining the standard projection direction of the unilateral lower limb 3D model based on the projection reference point includes: The lateral projection direction of the unilateral lower limb three-dimensional model is determined based on the direction of the line connecting the posterior point of the lateral femoral condyle and the posterior point of the medial femoral condyle. The orthogonal projection direction of the unilateral lower limb 3D model is determined based on the lateral projection direction and the preset vertical direction.

5. The method for generating lower limb projection images according to claim 4, characterized in that, The unilateral lower limb projection image includes a lateral projection image and an orthogonal projection image; The unilateral lower limb 3D model includes a unilateral lower limb element model; Projecting the unilateral lower limb 3D model based on the standard projection direction to obtain a unilateral lower limb projection image includes: Calculate the average density of voxel units in the lateral projection direction and the frontal projection direction of the unilateral lower limb voxel model, respectively. The lateral projection image and the orthogonal projection image are generated based on the average density values ​​of the lateral projection direction and the orthogonal projection direction, respectively.

6. The method for generating lower limb projection images according to claim 1, characterized in that, The three-dimensional model of the entire lower limb is processed to obtain a long bone model of the unilateral lower limb and a three-dimensional model of the unilateral lower limb, including: The three-dimensional model of the entire lower limb was segmented into voxels to obtain a voxel model of the long bones of the entire lower limb. Extract isosurface patches from the full lower limb long bone voxel model, and construct a full lower limb long bone mesh model based on the isosurface patches; A connected component is constructed based on the grid data points in the aforementioned full lower limb long bone mesh model; The full lower limb long bone mesh model is divided into a single lower limb long bone model on the left side of the human body and a single lower limb long bone model on the right side of the human body based on the center position of the connected domain. Calculate the minimum bounding box of the unilateral lower limb long bone model, and extract the unilateral lower limb 3D model corresponding to the unilateral lower limb long bone model from the full lower limb 3D model based on the minimum bounding box.

7. The method for generating lower limb projection images according to claim 6, characterized in that, Extracting the surface data point set of the unilateral lower limb long bone model includes: The number of voxels in each connected region is calculated based on the unilateral lower limb long bone model and the full lower limb long bone voxel model. Connected components whose number of voxels is greater than a preset voxel threshold are marked as valid connected components; Connected regions whose number of voxels is less than or equal to the preset voxel threshold are marked as noise; The surface data point set is obtained by extracting all grid data points within the effective connected domain.

8. The method for generating lower limb projection images according to claim 4, characterized in that, The lateral projection direction of the unilateral lower limb 3D model is determined based on the direction of the line connecting the posterior point of the lateral femoral condyle and the posterior point of the medial femoral condyle, including: Calculate the standard vector from the posterior point of the medial femoral condyle to the posterior point of the lateral femoral condyle, and normalize the standard vector to obtain a standard unit vector; The standard unit vector is determined as the lateral projection direction of the unilateral lower limb 3D model.

9. The method for generating lower limb projection images according to claim 8, characterized in that, Determining the frontal projection direction of the unilateral lower limb 3D model based on the lateral projection direction and the preset vertical direction includes: Obtain the vertical vector corresponding to the preset vertical direction; Calculate the target vector that is perpendicular to the standard unit vector and the vertical vector, and normalize the target vector to obtain the target unit vector; The target unit vector is determined as the orthogonal projection direction of the unilateral lower limb 3D model.

10. A terminal for generating lower limb projection images, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the method for generating lower limb projection images as described in any one of claims 1-9.

Images