Geometric feature-based knee joint cruciate ligament and collateral ligament dead center positioning method

Through computer graphics technology based on geometric features, positioning the center of the knee ligament dead center is solved, and the problem of insufficient positioning accuracy in traditional methods is achieved, high-precision and automated ligament dead center positioning is achieved, improving the surgical effect and patient recovery quality.

CN120510208APending Publication Date: 2025-08-19THE SECOND HOSPITAL OF DALIAN MEDICAL UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202510556835.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-04-29
Publication Date
2025-08-19

AI Technical Summary

Technical Problem

In clinical practice, the positioning of the knee ligament dead center has complex anatomical structure, large individual differences, and high positioning difficulty. Traditional methods rely on subjective experience to insufficient accuracy, which affects the stability and recovery effect of ligament reconstruction.

Method used

The geometric feature-based positioning method is adopted, combined with computer graphics technology, and VTK and OpenCV are used to process knee joint CT images. Through steps such as grayscale value conversion, binarization processing, least squares fitting, PCA fitting, three-dimensional reconstruction, etc., the geometric features of the bone are quantitatively analyzed to reduce subjective errors and improve positioning accuracy.

Benefits of technology

It improves the positioning accuracy and surgical efficiency of the knee ligament dead center, reduces the influence of doctors' subjective experience, is suitable for anatomical variations in different individuals, simplifies the operation process, and improves the success rate of ligament reconstruction surgery and the quality of patient recovery.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120510208A_ABST
    Figure CN120510208A_ABST
Patent Text Reader

Abstract

The invention discloses a geometric feature-based knee joint cruciate ligament and collateral ligament dead point center positioning method, which comprises the following steps of: acquiring knee joint CT images of a human knee joint and a cruciate ligament related skeleton, acquiring an intercondylar fossa vertex by combining a grey-scale map and binarization processing, screening an inner side point set by utilizing femur model features and combining a detection box, and positioning the knee joint cruciate ligament and collateral ligament dead point center. Determining a posterior cruciate ligament thighbone dead center; the femoral protuberance position is iteratively cut by utilizing the femur model characteristics and combining with medical priori knowledge, whether the medial collateral ligament dead center area is met or not is judged, and then the medial collateral ligament femoral dead center center is determined; combining medical priori knowledge and drawing characteristic lines, and determining the center of a medial collateral ligament tibia dead center; fitting a three-dimensional straight line by using a PCA (Principal Component Analysis) method by using femur model features, and determining the center of a lateral collateral ligament femur dead center through two feature lines in combination with medical priori knowledge; and determining the center of the fibula stop point of the lateral collateral ligament by utilizing the characteristics of the fibula model and combining the area of the fibula stop point of the lateral collateral ligament.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of medical image processing, and in particular to a method for locating the attachment centers of the cruciate ligament and collateral ligament of the knee joint based on geometric features. Background Art

[0002] The knee joint is the largest weight-bearing joint in the human body, and its stability depends on the coordinated action of multiple ligamentous structures. The anterior cruciate ligament (ACL) and posterior cruciate ligament (PCL) play a key role in the anterior-posterior stability of the knee, while the medial collateral ligament (MCL) and lateral collateral ligament (LCL) are responsible for maintaining medial and lateral stability. Knee ligaments are particularly susceptible to rupture in situations such as sports injuries and traffic accidents. After ligament injury, precise insertion center determination is crucial for successful ligament reconstruction. However, clinical practice still faces the following challenges: the complex anatomy of the knee joint makes positioning difficult, and individual variability in anatomical landmarks such as the intercondylar notch and tibial plateau complicates accurate positioning. Traditional surgical procedures often rely on surgeon experience, resulting in inaccurate positioning and compromised postoperative ligament restoration stability. Ligament insertions are often hidden in knee imaging, making anatomical landmarks difficult to identify, compromising accurate positioning. Standardized localization methods are lacking, and existing methods rely heavily on subjective experience. A unified, standardized workflow based on imaging and anatomical features has yet to be established, leading to significant individual variability. Summary of the Invention

[0003] In response to the problems existing in the prior art, the present invention discloses a method for locating the insertion centers of the cruciate ligament and collateral ligament of the knee joint based on geometric features. This method focuses on the image morphology of the bones themselves and locates the insertion centers by combining various computer graphics technologies. The method specifically includes the following steps:

[0004] Acquire CT images of the human knee joint and ligament-related bones, perform grayscale conversion and binarization on the knee CT images, determine the apex of the femoral intercondylar notch based on pixel points, fit the Blumensaat line using the least squares method, calculate the quadrant intersection using the perpendicular parallel relationship, and use this intersection as the insertion center of the two-dimensional sequence image. Use the VTK reconstruction model to determine the center position of the femoral insertion point of the anterior cruciate ligament.

[0005] The masking technology in OpenCV was used to superimpose knee CT images to obtain the tibial plateau image. The MinAreaRect function was used to determine the circumscribed rectangle of the tibial plateau image. The tibial insertion center was calculated by setting the ratio, and the tibial insertion center of the anterior cruciate ligament was located in three dimensions using ObbTree.

[0006] The intercondylar notch vertex was used as the cutting plane to obtain a half-femoral model. The centroid coordinates and boundaries of the detection box were determined based on the axis-aligned bounding box coordinates of the half-femoral model. The height and width of the detection box were determined based on the coordinate characteristics of the half-femoral model. This detection box was used to filter out a set of points on the medial femoral surface that met the criteria. The femoral insertion center of the posterior cruciate ligament was finally located based on the set ratio.

[0007] The posterior edge of the tibial plateau image is used to calibrate the insertion center, and the tibial plateau point closest to the insertion center is found by combining the closest point seeking function. This tibial plateau point is used as the final tibial insertion center of the posterior cruciate ligament.

[0008] The insertion area of the medial collateral ligament is found by using the geometric characteristics of the femoral bone surface and the area of the femoral insertion area of the medial collateral ligament, and the center point of the insertion area is used as the final center of the femoral insertion area of the medial collateral ligament;

[0009] Using the geometric features of the tibial bone surface, characteristic lines were drawn and combined with prior medical knowledge to determine the center position of the tibial insertion point of the medial collateral ligament;

[0010] The geometric features of the femoral bone surface are used to draw characteristic lines in combination with the straight line fitting of the three-dimensional point set to determine the center position of the femoral insertion point of the lateral collateral ligament;

[0011] The fibular head features were used in combination with the convex bone features of the fibular model to locate the center of the fibular insertion point of the lateral collateral ligament.

[0012] Furthermore, the femoral attachment center of the anterior cruciate ligament is obtained in the following manner: a 3D femoral model is reconstructed using the VTK surface rendering algorithm, and the image is cut into an image sequence along the sagittal plane, converted into a grayscale image and then binarized, a scanning strategy is formulated to obtain a point set on the intercondylar fossa based on white pixels, the point set is fitted with a linear fit using the least squares method to obtain a Blumensaat line, and its perpendicular line is calculated, and based on the medical theory of quartering, the four straight lines are interpolated at a 0.25 ratio intercept, and the intersection of two parallel straight lines is used as the femoral attachment center, the point is marked on the sagittal sequence image, and the VTK surface rendering technology is used to three-dimensionally reconstruct the attachment center and the femoral model to determine the three-dimensional position of the femoral attachment center of the anterior cruciate ligament.

[0013] Furthermore, the tibial insertion center of the anterior cruciate ligament is obtained by superimposing a series of CT cross-sectional images of the tibia and knee joint through OpenCV to obtain a complete tibial plateau image, using the MinAreaRect function to obtain the minimum circumscribed rectangle of the tibial plateau image and record the coordinates of its four vertices, calculating the midpoint coordinates of the four circumscribed rectangle vertices based on medical theory, and deducing the position of the tibial insertion center, and determining the intersection of the straight line and the tibial model through the ObbTree algorithm as the final tibial insertion center of the anterior cruciate ligament;

[0014] Furthermore, the femoral attachment center of the posterior cruciate ligament is obtained in the following manner: using the vertex of the intercondylar notch as the cutting plane, cutting along the middle of the femoral model, traversing all vertices of the femoral model, finding the point with the smallest Y value and recording its X coordinate as the width of the detection box, traversing downward along the Z axis, recording the value with the largest Y-axis difference as the height of the detection box, obtaining the medial surface of the femur based on the detection box, and determining the femoral attachment center of the posterior cruciate ligament through coordinate conversion and proportional calculation.

[0015] Furthermore, the tibial attachment center of the posterior cruciate ligament is obtained in the following manner: the attachment center is marked on each CT cross-sectional image, and the point on the tibial platform closest to the positioning point is obtained using the nearest point seeking function as the final tibial attachment center of the posterior cruciate ligament.

[0016] Furthermore, the femoral attachment center of the medial collateral ligament is obtained in the following manner: a three-dimensional model of the femur is reconstructed based on the CT image, and then, based on medical prior knowledge, the raised positions on the femoral surface are continuously cut until the area of the medial collateral ligament attachment point is reached, and then the midpoint of the area is determined as the final femoral attachment center of the medial collateral ligament.

[0017] Furthermore, the medial collateral ligament tibial insertion center is determined by performing a cut on the proximal and distal ends of the tibia based on a reconstructed 3D tibial model, obtaining the center points of the cut sub-models, and connecting them to obtain the tibial axis. The highest points on both sides of the tibial plateau are connected as the base axis. In conjunction with medical knowledge, the medial collateral ligament tibial insertion center is located frontally. Combining these two locations, the medial collateral ligament tibial insertion center is then determined.

[0018] Furthermore, the femoral attachment center of the lateral collateral ligament is obtained in the following manner: based on the reconstructed three-dimensional femoral model, a three-dimensional point set of the posterior cortex of the femur is collected, and a three-dimensional straight line is fitted as the femoral characteristic line according to PCA to determine the position of the lateral femoral collateral ligament on the side. Then, based on the horizontal placement of the two femoral condyles, the position of the lateral femoral collateral ligament on the front is determined, and the intersection of the two position information is combined as the femoral attachment center position of the lateral collateral ligament.

[0019] Furthermore, the center of the fibular attachment point of the lateral collateral ligament is obtained in the following manner: based on medical prior knowledge, the raised feature points of the fibular head are cut, and its area is iterated to reach the medical average area of the lateral collateral ligament at the fibular head, thereby determining the center position of the fibular attachment point of the lateral collateral ligament.

[0020] Due to the adoption of the above-mentioned technical solution, the present invention provides a method for locating the insertion centers of the cruciate ligament and collateral ligament of the knee joint based on geometric features. This method uses image processing technologies such as VTK and OpenCV to quantitatively analyze the insertion position using the geometric features of the bones (such as the vertex of the intercondylar notch, Blumensaat line, tibial plateau circumscribed rectangle, tibial axis, femoral base axis and insertion area, etc.), thereby reducing the errors in traditional surgery that rely on the doctor's subjective experience. In addition, the use of least squares method, iterative calculation, PCA and point-to-line distance formula ensures the mathematical accuracy of the positioning point, improves the reliability of the insertion center identification, and improves surgical efficiency. In addition, the computer program of this method occupies less memory, has low time complexity, and runs faster. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments recorded in this application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0022] Figure 1 Flowchart for realizing the method of the present invention

[0023] Figure 2 This is the effect diagram of the femoral intercondylar notch boundary in the present invention

[0024] Figure 3 Schematic diagram of points on the intercondylar fossa of the femur in the present invention

[0025] Figure 4 Schematic diagram of the femoral Blumensaat line in the present invention

[0026] Figure 5 Schematic diagram of three straight lines of the lateral femoral condyle in the present invention

[0027] Figure 6 Schematic diagram of the quartering method in the present invention

[0028] Figure 7 Schematic diagram of the quarter-point method for locating the anterior cruciate ligament insertion center in the present invention

[0029] Figure 8 Schematic diagram of the femoral insertion center of the anterior cruciate ligament in the present invention

[0030] Figure 9 Schematic diagram of the circumscribed rectangle of the tibia in the present invention

[0031] Figure 10 Schematic diagram of the tibial insertion center of the anterior cruciate ligament in the present invention

[0032] Figure 11 Schematic diagram of the femur detection box in the present invention

[0033] Figure 12 Schematic diagram of the extraction of the medial surface of the femur in the present invention

[0034] Figure 13 Schematic diagram of the femoral insertion center of the posterior cruciate ligament in the present invention

[0035] Figure 14 Schematic diagram of the tibial insertion center of the posterior cruciate ligament in the present invention

[0036] Figure 15 Schematic diagram of the initial cutting of the femoral insertion area of the medial collateral ligament in the present invention

[0037] Figure 16 Schematic diagram of the cutting of the femoral insertion area of the medial collateral ligament in the present invention

[0038] Figure 17 Schematic diagram of the femoral attachment center of the medial collateral ligament in the present invention

[0039] Figure 18 Schematic diagram of the medial collateral ligament tibial axis point in the present invention

[0040] Figure 19 Schematic diagram of the medial collateral ligament tibial axis in the present invention

[0041] Figure 20 Schematic diagram of the medial collateral ligament tibial base axis in the present invention

[0042] Figure 21 Schematic diagram of the front position of the medial collateral ligament of the tibia in the present invention

[0043] Figure 22 Schematic diagram of the tibial insertion center of the medial collateral ligament in the present invention

[0044] Figure 23 Schematic diagram of the femoral position of the lateral collateral ligament in the present invention

[0045] Figure 24 Schematic diagram of the femoral end point set of the lateral collateral ligament in the present invention

[0046] Figure 25 Schematic diagram of the femoral characteristic line of the lateral collateral ligament in the present invention

[0047] Figure 26 Schematic diagram of the femoral side position of the center of the lateral collateral ligament insertion point of the present invention

[0048] Figure 27 Schematic diagram of the femoral condyle base axis of the lateral collateral ligament of the present invention

[0049] Figure 28Schematic diagram of the femoral insertion center of the lateral collateral ligament of the present invention

[0050] Figure 29 Schematic diagram of the position of the lateral collateral ligament fibula in the present invention

[0051] Figure 30 Schematic diagram of the insertion point of the lateral collateral ligament of the present invention

[0052] Figure 31 Schematic diagram of the fibular insertion center of the lateral collateral ligament of the present invention DETAILED DESCRIPTION

[0053] To make the technical solutions and advantages of the present invention more clear, the technical solutions in the embodiments of the present invention are clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention:

[0054] like Figure 1 The method shown in the figure is a method for locating the insertion centers of the cruciate ligament and collateral ligament of the knee joint based on geometric features. This method is based on relevant medical literature and clinical requirements as theoretical support, and combines multiple computer graphics technologies to process knee joint images for algorithm design. In the specific implementation process, OpenCV technology is used to traverse the binary bone image to determine the boundary of the intercondylar fossa, such as Figure 2 As shown. Get the point set on the intercondylar fossa and use the least squares method to perform straight line fitting, as shown Figure 3-4 Based on the quantitative relationship between parallel and perpendicular straight lines in geometry, the four straight lines located by the four-grid method are obtained, and then the center of the anterior cruciate ligament insertion point in the two-dimensional sequence image is obtained according to the theory, as shown in FIG. Figure 5-7 As shown. The femoral model was reconstructed using VTK surface rendering to obtain the femoral insertion center of the anterior cruciate ligament, as shown Figure 8 The circumscribed rectangle and three-dimensional reconstruction technology are used to locate the tibial insertion center of the anterior cruciate ligament, as shown in Figure 9-10 As shown. Still using the coordinate features of the 3D model in VTK, construct the medial surface detection box and then locate the center of the femoral insertion point of the posterior cruciate ligament according to the proportion, as shown Figure 11-13 As shown. Based on the circumscribed rectangle and the closest point algorithm, the center of the tibial insertion point of the posterior cruciate ligament is located, as shown in Figure 14 In addition, Figure 15-17 As shown in the figure, the femoral bone surface convex features are used to iteratively cut the insertion area until the medical requirements are met, thereby determining the femoral insertion center of the medial collateral ligament. Figure 18-22 As shown in the figure, the center of the tibial insertion of the medial collateral ligament is located according to the tibial axis and the base axis. Figure 23-28 As shown in the figure, the femoral attachment center of the lateral collateral ligament is determined based on the femoral posterior cortex and base axis position information. Figures 29-31 As shown, the center position of the insertion point is determined based on the position of the fibula protrusion and combined with medical knowledge.

[0055] The specific steps of the method disclosed in the present invention are as follows:

[0056] S1: Obtain a CT image of the human femur and knee joint to determine the femoral insertion center of the anterior cruciate ligament. The specific method is as follows:

[0057] S11. Automatically obtain the height value of the apex of the intercondylar notch of the femur: Based on the morphological characteristics of the intercondylar notch of the femur, the pixel with the smallest height value when observed from the sagittal plane is the apex of the intercondylar notch. First, the three-dimensional model of the femur is reconstructed using the VTK surface rendering algorithm based on the serial images of the cross-section of the femoral knee joint CT. The sagittal plane sequence images of the femoral knee joint are cut according to the spacing when the CT images were taken, and then converted into grayscale images and binarized. Scan from the bottom of the image upward, traverse the image row coordinates, and scan from right to left in each row to detect whether the pixel value is 255 (white). Record the current image number and coordinates, compare the images in the sequence, find the row coordinates with the smallest height value, and then obtain the image and vertex of the intercondylar notch.

[0058] S12. After obtaining the intercondylar notch image and its vertex, scan from the right top of the intercondylar notch image to the left and record the point with the largest width value among the 255 pixel values. Figure 2 As shown, two perpendicular straight lines are drawn to mark the row where the apex of the intercondylar notch is located and the rightmost boundary, and points on the intercondylar notch are obtained within this range.

[0059] S13. Obtaining Blumensaat Line: Figure 2 Scan the pixels from right to left within the boundary shown. When a white pixel is detected, its coordinates are recorded and the scanning of the current row is terminated. The point set on the intercondylar fossa can be obtained, as shown in Figure 3 Then, the least square method was used to fit the intercondylar fossa point set to a straight line (as shown in Figure 4 This straight line is medically known as Blumensaat's line.

[0060] S14. Using OpenCV technology, the binary images of the lateral femoral condyle are superimposed, and the Blumensaat line is added to the superimposed lateral femoral condyle image to obtain the starting and ending point coordinates of the Blumensaat line (x A ,y A ), (x B ,y B Assuming that the equation of the Blumensaat line is: y = kx + b, we can then obtain two straight lines passing through the coordinates of the starting and ending points and perpendicular to the Blumensaat line, with slopes of like Figure 5Finally, the distance from the point to the straight line is calculated using the following formula. By iteratively changing the parameter b' and traversing all k straight lines with a slope of in the intercondylar fossa area, we calculate the straight line with the maximum distance, as shown in the figure below. Figure 6 As shown, four straight lines are drawn.

[0061]

[0062] S15. According to the medical theory of the quartering method, the anterior cruciate ligament is located at the intersection of the four quarters of the lateral femoral condyle near the Blumensaat line. Two parallel lines among the four lines are grouped together. Interpolation is performed between the two lines at a ratio of 0.25 and an intercept b' is used as the position of the quartering method positioning line. According to the following formula, k and b are the slope and intercept of the Blumensaat line, respectively. The intersection of the two lines is calculated, which is the center (x, y) of the femoral insertion point of the anterior cruciate ligament on the two-dimensional sequence image, as shown in the following example: Figure 7 shown.

[0063]

[0064] S16. Add the insertion center to each sequence of femoral 2D sagittal images and perform 3D reconstruction of the sagittal sequence images using VTK surface rendering technology. The insertion center is reconstructed as a straight line, and the femoral image is reconstructed as a 3D mesh model. The intersection of the straight line and the bone is obtained using ObbTree, which is the final femoral insertion center of the anterior cruciate ligament, as shown in the figure. Figure 8 shown.

[0065] S2: Locate the center of the tibial insertion of the anterior cruciate ligament using the circumscribed rectangle of the tibial plateau. The specific method is as follows:

[0066] S21. First, use OpenCV technology to overlay the complete tibial plateau on the CT cross-sectional images of the knee joint. Use the MinAreaRect function to find the minimum bounding rectangle of the overlaid image. The coordinates of the four boundary points of the bounding rectangle are stored in a two-dimensional array a with four rows and two columns, as shown in the following example: Figure 9 shown

[0067] S22. According to medical theory, the center of the tibial insertion of the anterior cruciate ligament is located 48% of the distance from the anterior end to the tibial plateau and 46% of the distance from the medial end to the tibial plateau. The center of the femoral insertion of the anterior cruciate ligament (w, h) on 2D sequential images can be calculated using the following formula.

[0068] w1=|a[0] x -a[1] x |×0.54+min(a[0] x ,a[1] x )

[0069] h1=|a[0]y -a[1] y |×0.54+min(a[0] y ,a[1] y )

[0070] w2=|a[2] x -a[3] x |×0.54+min(a[2] x ,a[3] x )

[0071] h2=|a[2] y -a[3] y |×0.54+min(a[2] y ,a[3] y )

[0072]

[0073] S23. As shown in S16, the insertion center is added to each sequence of 2D cross-sectional images of the tibia and the cross-sectional sequence images are reconstructed into three dimensions. The insertion center is reconstructed into a straight line, and the tibia image is reconstructed into a three-dimensional mesh model. The intersection of the straight line and the bone is the final tibial insertion center of the anterior cruciate ligament, as shown in FIG. Figure 10 shown.

[0074] S3: Using the detection box generated from the femoral 3D mesh model reconstructed using VTK, the femoral insertion center of the posterior cruciate ligament is located using the following method:

[0075] S31. Generate a cutting plane using the intercondylar notch vertex coordinates obtained in S11, cut the femoral 3D mesh model in half, observe the coordinate features of the femoral 3D mesh model, traverse all vertices, find the vertex with the smallest Y value, and use the X coordinate of this point as the width of the detection box; starting from the point with the smallest Y value, traverse downward along the Z axis and find the point with the largest Y axis difference as the height; calculate the bounding box of the femoral 3D mesh model to obtain the boundary range of the three coordinate axes, and then create a complete detection box, such as Figure 11 shown.

[0076] S32. Use the VtkSelectEnclosedPoints function to filter the points within the bounding box, then calculate the angle between the normal vector and the Z axis, retain the points facing a specific angle, further optimize the filtering effect, and obtain the medial surface of the femur, such as Figure 12 shown.

[0077] S33. According to medical theory, the center of the posterior cruciate ligament insertion is located 12 mm from the highest point and 7 mm from the anterior edge. The final center of the posterior cruciate ligament femoral insertion is located by converting the medial surface proportions, such as Figure 13 shown.

[0078] S4: Based on the cross-sectional CT image of the knee joint, locate the tibial insertion center of the posterior cruciate ligament:

[0079] According to the circumscribed rectangle of the superimposed posterior tibial plateau obtained in S21, according to medical theory, the insertion center is located at the posterior edge. After adding the insertion center to the CT image sequence of the tibial knee joint cross section and three-dimensional reconstruction, the closest point function is used to find the point on the tibial plateau closest to the positioning point as the final posterior cruciate ligament tibial insertion center, as shown in the figure below: Figure 14 As shown, avoid the insertion center in the two-dimensional sequence image being too peripheral and having no intersection with the tibial plateau.

[0080] S5: Using the 3D femoral mesh model obtained after VTK reconstruction, iteratively cut the lateral area of the femoral model to locate the femoral insertion center of the medial collateral ligament. The specific method is as follows:

[0081] S51: First, rotate the femoral model and adjust the position to the side view. According to medical theory, the medial collateral ligament femoral area is convex due to flexion movement. Therefore, cut from the highest point of the femoral bone surface, as shown in the figure. Figure 15 shown.

[0082] S52: Project the cut femoral sub-model on the xoy plane, calculate the projected area of the femoral sub-model, and compare it with the average value of the medial collateral ligament femoral attachment area in medical clinical knowledge. If the projected area reaches the average value range of the medial collateral ligament attachment area, then the area is determined to be the medial collateral ligament attachment area. Otherwise, the cutting is continuously iterated and the next depth cutting is performed according to the following formula, such as Figure 16 shown.

[0083] Depth next =Depth-Depth now

[0084] S54: When the average value of the clinical insertion area is reached, the average value of all point sets is calculated as the center point of the current insertion area. Then, through the kd tree, the point on the femoral bone surface closest to the calculated center point is located as the center position of the femoral insertion point of the medial collateral ligament, such as Figure 17 shown.

[0085] S6: Using the 3D tibial mesh model obtained after VTK reconstruction, we draw feature lines and, combined with clinical knowledge, locate the tibial insertion center of the medial collateral ligament using the following method:

[0086] S61: First, rotate the tibia laterally, then cut the proximal and distal tibia models, locate the center points of the two cut sub-models, and connect them as the tibial axis. According to medical clinical knowledge, the center of the tibial insertion point of the collateral ligament is located 4mm to the right of this axis, such as Figure 18-19 shown.

[0087] S62: Then rotate the model to the front, cut both sides of the tibial plateau, and then locate the highest points of the two sub-models and connect them as the base axis, such as Figure 20 shown.

[0088] S62: From the frontal view, the center of the medial collateral ligament tibial insertion is located 55 mm below the base axis. Figure 21 As shown in the figure, the lateral position information can be used to locate the center of the tibial insertion point of the medial collateral ligament, as shown in the figure. Figure 22 shown.

[0089] S7: Using the femoral 3D mesh model obtained after VTK reconstruction, draw feature lines and combine clinical medical knowledge to locate the femoral insertion center of the lateral collateral ligament. The specific method is as follows:

[0090] S71: First, rotate the femoral model to the side and cut the femoral end and femoral condyle. Figure 23 shown.

[0091] S72: Collect a three-dimensional point set of the posterior femoral cortex at the end of the femur, and fit a three-dimensional straight line according to the PCA method. The formula is as follows.

[0092]

[0093] Where C is the center of mass, P i is a point in the point set. After calculating the centroid, the point set is translated to the centroid. The formula is:

[0094] P′ i =P i -C

[0095] Among them, P′ i is the point set after translation, which is then used to calculate the covariance matrix of the point set after translation. The formula is as follows.

[0096]

[0097] Among them, S jk is an element in the covariance matrix, P′ ij is the coordinate of the i-th point in the j-dimension, is the mean of the j-dimension. The eigenvalues and eigenvectors of the covariance matrix are used to fit a three-dimensional straight line, such as Figures 24-26 shown.

[0098] S73: Then rotate the femoral model in the frontal position and cut the femoral condyle model into the inner and outer femoral condyles. Traverse the inner and outer femoral condyles, locate the lowest points of the two femoral condyles, and connect the two points as the base axis, such as Figure 27 shown.

[0099] S74: Based on the base axis position and combined with medical clinical knowledge, locate the center position of the femoral insertion point of the lateral collateral ligament in the front position, and determine the center position of the femoral insertion point of the lateral collateral ligament in combination with the lateral characteristic line, such as Figure 28 shown.

[0100] S8: Using the fibula 3D mesh model reconstructed using VTK and combined with clinical knowledge, the fibula insertion center of the lateral collateral ligament was located using the following method:

[0101] S81: First, rotate the fibula to the side. According to medical knowledge, the lateral collateral ligament pulling the bone tissue will cause the fibula to bulge. Therefore, locate the bulge of the fibula and cut it. Figures 29-30 shown.

[0102] S82: Cut the fibula until the area of the sub-model cut from the fibula reaches the average position of the medically defined lateral collateral ligament insertion area, then traverse all the point sets of the cut sub-model, and calculate the center point as the center position of the insertion point, such as Figure 31 shown.

[0103] The present invention is comprehensive and automated. It automatically locates the insertion center of the human knee ligament directly based on the corresponding algorithm by inputting the knee CT image. It adopts methods such as automatic scanning, parametric calculation and linear fitting, which significantly reduces manual intervention and improves operational efficiency, especially when the amount of image data is large. The algorithm is designed in combination with medical anatomical theory, has strong versatility, and is applicable to the anatomical variations of different individuals. While ensuring accuracy, it simplifies the operating process and has good clinical application prospects. It is expected to improve the success rate of knee ligament reconstruction surgery and the quality of postoperative recovery of patients.

[0104] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A method for locating the insertion centers of the cruciate ligament and collateral ligament of the knee joint based on geometric features, the characteristics of which include: Acquire CT images of the human knee joint and ligament-related bones, perform grayscale conversion and binarization on the knee CT images, determine the apex of the femoral intercondylar notch based on pixel points, fit the Blumensaat line using the least squares method, calculate the quadrant intersection using the perpendicular parallel relationship, and use this intersection as the insertion center of the two-dimensional sequence image. Use the VTK reconstruction model to determine the center position of the femoral insertion point of the anterior cruciate ligament. The masking technology in OpenCV was used to superimpose knee CT images to obtain the tibial plateau image. The MinAreaRect function was used to determine the circumscribed rectangle of the tibial plateau image. The tibial insertion center was calculated by setting the ratio, and the tibial insertion center of the anterior cruciate ligament was located in three dimensions using ObbTree. The intercondylar notch vertex is used as the cutting plane to obtain a half-femoral model. The centroid coordinates and boundaries of the detection box are determined based on the axis-aligned bounding box coordinates of the half-femoral model. The height and width of the detection box are determined based on the coordinate characteristics of the half-femoral model. The detection box is used to filter out the set of points on the medial surface of the femur that meet the requirements. The center of the femoral insertion point of the posterior cruciate ligament is finally located according to the set ratio. The posterior edge of the tibial plateau image is used to calibrate the insertion center, and the tibial plateau point closest to the insertion center is found by combining the closest point seeking function. This tibial plateau point is used as the final tibial insertion center of the posterior cruciate ligament. The insertion area of the medial collateral ligament is found by using the geometric characteristics of the femoral bone surface and the area of the femoral insertion area of the medial collateral ligament, and the center point of the insertion area is used as the final center of the femoral insertion area of the medial collateral ligament; Using the geometric features of the tibial bone surface, characteristic lines were drawn and combined with prior medical knowledge to determine the center position of the tibial insertion point of the medial collateral ligament; The geometric features of the femoral bone surface are used to draw characteristic lines in combination with the straight line fitting of the three-dimensional point set to determine the center position of the femoral insertion point of the lateral collateral ligament; The fibular head features were used in combination with the convex bone features of the fibular model to locate the center of the fibular insertion point of the lateral collateral ligament.

2. The method for locating the insertion centers of the cruciate ligament and collateral ligament of the knee joint based on geometric features according to claim 1, characterized in that: The VTK surface rendering algorithm was used to reconstruct the 3D model of the femur, and the image sequence was cut along the sagittal plane. The image sequence was converted into grayscale and then binarized. A scanning strategy was formulated to obtain the point set on the intercondylar fossa based on the white pixels. The Blumensaat line was obtained by linear fitting of the point set using the least squares method, and its perpendicular line was calculated. According to the medical theory of quartering, the four lines were interpolated with a 0.25 intercept ratio. The intersection of the two parallel lines was used as the femoral insertion center, and this point was marked on the sagittal plane sequence image. The insertion center and femoral model were reconstructed in 3D using VTK surface rendering technology to determine the 3D position of the femoral insertion center of the anterior cruciate ligament.

3. The method for locating the insertion centers of the cruciate ligament and collateral ligament of the knee joint based on geometric features according to claim 2, characterized in that: A complete tibial plateau image was obtained by superimposing CT cross-sectional image sequences of the tibia and knee joint using OpenCV. The minimum bounding rectangle of the tibial plateau image was obtained using the MinAreaRect function and the coordinates of its four vertices were recorded. The midpoint coordinates of the four vertices of the bounding rectangle were calculated based on medical theory, and the position of the tibial insertion center was deduced. The intersection of the straight line and the tibial model was determined using the ObbTree algorithm, which was used as the final tibial insertion center of the anterior cruciate ligament. Where k and b are the slope and intercept of the Blumensaat line, respectively, b′ is the line with the maximum distance, and (x, y) is the center of the femoral insertion of the anterior cruciate ligament on the two-dimensional sequence image.

4. The method for locating the insertion centers of the cruciate ligament and collateral ligament of the knee joint based on geometric features according to claim 3, characterized in that: The intercondylar notch vertex was used as the cutting plane, and the femoral model was cut along the middle. All vertices of the femoral model were traversed, and the point with the smallest Y value was found and its X coordinate was recorded as the width of the detection box. The detection box was traversed downward along the Z axis, and the value with the largest Y-axis difference was recorded as the height of the detection box. The medial surface of the femur was obtained based on the detection box, and the center of the femoral insertion point of the posterior cruciate ligament was determined through coordinate transformation and proportional calculation.

5. The method for locating the insertion centers of the cruciate ligament and collateral ligament of the knee joint based on geometric features according to claim 1, characterized in that: The insertion center was marked on each CT cross-sectional image, and the point on the tibial plateau closest to the positioning point was obtained using the nearest point seeking function as the final tibial insertion center of the posterior cruciate ligament.

6. The method for locating the insertion centers of the cruciate ligament and collateral ligament of the knee joint based on geometric features according to claim 5, characterized in that: The three-dimensional model of the femur is reconstructed based on the CT images. Based on medical prior knowledge, the raised areas on the femoral surface are continuously cut until the area of the femoral insertion of the medial collateral ligament is reached. The midpoint of the area is then determined as the final center of the femoral insertion of the medial collateral ligament.

7. The method for locating the insertion centers of the cruciate ligament and collateral ligament of the knee joint based on geometric features according to claim 6, characterized in that: Based on the reconstructed three-dimensional tibial model, the proximal and distal ends of the tibia were cut to obtain the center point of the cut sub-model, and the tibial axis was obtained by connecting the lines. The highest points on both sides of the tibial plateau were connected as the base axis and combined with medical knowledge to locate the center position of the tibial attachment point of the medial collateral ligament on the front. The two position information were combined to determine the center of the tibial attachment point of the medial collateral ligament.

8. The method for locating the insertion centers of the cruciate ligament and collateral ligament of the knee joint based on geometric features according to claim 7, characterized in that: Based on the reconstructed three-dimensional femoral model, a three-dimensional point set of the posterior femoral cortex was collected, and a three-dimensional straight line was fitted as the femoral characteristic line according to PCA to determine the lateral position of the lateral femoral collateral ligament. Then, based on the horizontal placement of the two femoral condyles, the frontal position of the lateral femoral collateral ligament was determined. The intersection of the two position information was combined as the center position of the femoral insertion point of the lateral collateral ligament.

9. The method for locating the insertion centers of the cruciate ligament and collateral ligament of the knee joint based on geometric features according to claim 8, characterized in that: According to medical prior knowledge, the raised feature points of the fibular head are cut, and its area is iterated to reach the average area of the lateral collateral ligament on the fibular head in medicine, thereby determining the center position of the fibular insertion point of the lateral collateral ligament.