A method for detecting and analyzing lower limb deformity

Abstract

The present invention discloses a method for analyzing lower limb deformity, comprising the following steps: S1, obtaining lower limb images taken by an X-ray machine and splicing them to obtain a full image of the lower limb; S2, performing morphological processing based on color information, repairing the text area and reducing the watermark; S3, enhancing image contrast; S4, segmenting the enhanced image to remove muscle tissue and retain bone; S5, coarsely positioning the segmented image, obtaining a bounding box of key parts, and determining whether a prosthesis is implanted; S6, extracting edge information inside the bounding box, obtaining a closed contour, and constructing an edge set expressing bone characteristics; S7, calculating the mechanical axis based on the edge set, calculating the bone curvature, analyzing the lower limb deformity state, and visualizing the stress condition. Compared with the method based on the lower limb force line, the present invention uses the curvature of the bone itself for force analysis, which is more in line with the actual stress condition and can automatically mark the mechanical axis of the lower limb.

Description

Technical Field

[0001] The present invention relates to the field of medical technology, and in particular to a method for detecting and analyzing lower limb deformities. Background Art

[0002] Lower limb deformity is a common orthopedic disease in clinical practice. This deformity may be congenital, existing at birth, or acquired, such as due to disease, injury or bad living habits. Lower limb deformity can affect a person's walking ability, posture, comfort and overall health. In actual production, full-length lower limb films are usually used to evaluate the patient's lower limb force line, which plays an important role in orthopedic surgical planning, lower limb deformity correction, joint replacement and the diagnosis and treatment of lower limb force line-related diseases.

[0003] The most commonly used method for measuring lower limb force lines and angles in clinical practice still relies on manual operation by doctors. Doctors need to manually mark specific anatomical points on X-ray films, such as the hip joint center, knee joint center, and ankle joint center, and then connect these points to draw the lower limb force lines. Then, they measure various key angles (such as hip-knee-ankle angle, femoral angle, tibial angle, etc.). Although these operations are not complicated, they are tedious and time-consuming, and require high operational proficiency, attention, and psychological state of doctors. Therefore, they are easily affected by human factors, resulting in subjective errors in the measurement results. Automatic and intelligent lower limb deformity detection is of great significance. It automatically identifies key anatomical points and force lines, calculates relevant angles, improves the consistency of analysis, and provides advice to doctors, thereby improving the efficiency of the diagnosis and treatment process.

[0004] Existing methods usually determine the force lines of the lower limbs based on key point positioning, among which the regression method directly predicts the coordinates of the key points. This method usually uses a deep learning model to classify pixels in the input image and learn the mapping from the image to the key point coordinates, but the performance is poor when the key point position is blurred or occluded. The heat map method predicts the probability distribution of each key point, that is, predicts the probability of each pixel position in the image being a key point. The final key point position is usually the point with the highest probability on the heat map, which helps to deal with occlusion and blur, but the computational cost is higher.

[0005] For example, the existing Chinese patent technology CN112348892A point positioning method and related devices and equipment improve the accuracy of target point detection by combining coarse positioning heat maps and fine positioning heat maps. However, generating and analyzing coarse positioning and fine positioning heat maps may require higher computing resources, especially when processing high-resolution images. The accuracy of the fine positioning heat map is highly dependent on the accuracy of the coarse positioning heat map. If the coarse positioning is inaccurate, the improvement of the fine positioning is limited.

[0006] Existing Chinese patent technology CN113907774A is a method and device for measuring lower limb force lines. The full-length film of the lower limb is cut into independent images of the hip joint, knee joint and ankle joint. The corresponding key point model is used to extract key points from the hip joint, knee joint and ankle joint images respectively, including the center point of the femoral head, the center point of the intercondylar notch of the knee joint and the center point of the tibial plateau. The femoral mechanical axis is determined according to the center point of the femoral head and the center point of the intercondylar notch of the knee joint, the tibial mechanical axis is determined according to the center point of the intercondylar notch of the knee joint and the center point of the tibial plateau, the first angle and the second angle are determined according to the anatomical axis of the femur and the anatomical axis of the tibia, and the lower limb force line is determined according to the center point of the femur and the center point of the ankle joint, thereby realizing automatic calculation of the lower limb force line and angle.

[0007] CN114581469A uses straight line detection to determine the contour, ignoring the curvature of the bone itself, and the cropping area is manually set in advance for the cropping length, which cannot adapt to the diversity of deformed images.

[0008] The above scheme simplifies the limbs into points and lines. Although it can reflect the force line situation of the lower limbs to a certain extent, it ignores the contribution of muscles, ligaments and other soft tissues, which actually have an important influence on the stability and force distribution of the lower limbs. The above assumptions are different from the actual force situation of the lower limbs. To solve the above problems, the present invention proposes a lower limb deformity detection method with improved lower limb force lines, which realizes automatic force line annotation and deformity detection tasks of the entire X-ray lower limb. Summary of the Invention

[0009] The purpose of the present invention is to provide a method for analyzing lower limb deformities in order to solve the above problems. The mechanical axis is automatically marked based on contour information to make up for the current automatic extraction method of the mechanical axis based on the key point detection algorithm, which cannot truly reflect the stress situation of the lower limbs under special bone morphology and loses information, thereby solving the problems mentioned in the background technology.

[0010] In order to solve the above problems, the present invention provides a technical solution:

[0011] A method for detecting and analyzing lower limb deformity comprises the following steps:

[0012] S1. Use an X-ray machine to take segmented images of different parts of the lower limb to obtain images that include the morphology of the lower limb, and then stitch them together to obtain a full-length image of the lower limb;

[0013] S2. Repair the text area of ​​the entire film according to the color to reduce the watermark;

[0014] S3, converting into a binary image and enhancing the image contrast of the repaired image;

[0015] S4. Segment the enhanced image, remove the muscles in the whole image, and keep the bones;

[0016] S5. Perform coarse positioning on the segmented image, obtain the bounding box of the key parts, and determine whether the prosthesis is implanted;

[0017] S6. Extract edge information from the bounding box area and construct an edge set that expresses bone features;

[0018] S7. Calculate the robust mechanical axis based on the bone edge set, analyze the force conditions of the lower limbs, and perform deformity detection.

[0019] Preferably, the specific steps of determining whether a prosthesis is implanted in the lower limb in step S5 include:

[0020] S51, inputting the result of the rough positioning as ROI, setting a key part set to store the key parts representing the local features of the bone;

[0021] S52, dividing each key part into two subsets according to its position in the lower limb region, and calculating the mean pixel intensity within each bounding box according to the starting point and the ending point of each bounding box;

[0022] S53. Compare the intensity difference between the left and right corresponding areas, and use the threshold α to determine whether the current part is implanted with a prosthesis. If it is less than the threshold, it is considered that the left and right parts are in the same state. Otherwise, it is considered that the patient has been implanted with a prosthesis. Further compare it with the pixel mean of the non-critical part area. If it is less than the threshold, it is considered that the patient has not been implanted with a prosthesis.

[0023] Preferably, the processing and analysis of the skeleton edge set in step S7 includes the following steps:

[0024] S71. Use the Canny operator for edge detection and cv2.findContours to obtain bone contour information. For the two key parts at the top, namely the femoral area, count the pixel intensity in the x direction, and use the second highest peak and valley as the edge point of the greater trochanter. Similarly, use the first peak and valley as the starting point, count the pixel intensity in the y direction, and use the first peak and valley as the end point. Traverse the points on the contour and draw a perpendicular line along the contour, with the midpoint as the boundary. Accumulate the intersection points of the perpendicular lines and take the average value as the key force point of the femoral head.

[0025] S72. For the key intermediate parts, i.e., the tibia and femur, the tibia and femur are distinguished based on the size of the contour y-coordinate. For the femur ROI, the contour perimeter is calculated and then reduced to serve as the threshold for the next step. The contour is approximated using a polygon. The contour points are compared with the true line according to the set threshold. The screening criterion is the percentage threshold of the contour perimeter.

[0026] S73. Further subdivide the contour into three irregular curved edges by endpoints, among which the lower edge of the femur and the upper edge of the tibia are used as a pair of matching edges. Sample along the y-axis direction and take the midpoint of the intersection of the edge and the straight line to obtain the contact surface between the tibia and femur. Connect the center point of the femoral head and the knee joint to obtain the femoral mechanical axis. Similarly, connect the center of the talus and the center of the knee joint to obtain the tibial mechanical axis. Connect the center point of the femoral head and the center point of the ankle joint to obtain the lower limb mechanical axis.

[0027] Preferably, the method for generating a force curve for lower limb force analysis comprises the following steps:

[0028] A1. Calculate the shortest distance d from the center of gravity of the key part to the boundary. At the distance d from the center of gravity, draw a horizontal line along the x-axis to obtain two intersection points (x 11 ,y1),(x 21 ,y1), and the midpoint coordinates Calculate the angle between the midpoint and the center of gravity and the contact surface to reflect the degree of inclination of the bone at both ends. Determine the number of interpolation nodes n based on the degree of inclination and the patient's age.

[0029] A2. Starting from the y coordinate of the center of gravity of the key part, find the intersection point with the edges on both sides along the x direction, denoted as x 01 ,x 02 , then traverse along the y direction to obtain the coordinate sets on both sides {(x 11 ,y1),(x 21 ,y1),…,(x n1 ,y n ),(x n2 ,y n )}, for each pair of matching points (xi1,yi), (xi2,yi), calculate the midpoint, the coordinates of the midpoint are The interpolation node set is The curve C(y) describing the bone morphological characteristics is generated using the cubic spline interpolation method;

[0030] A3. Draw a perpendicular line to the generated curve C(y) for each y coordinate. The direction of the perpendicular line is obtained by calculating the slope C'(y) of the tangent line at each point: A series of arrows are generated to describe the force conditions at various locations on the bone.

[0031] The beneficial effects of the present invention are:

[0032] 1. Using curves instead of lower limb force lines for deformity detection, force analysis can more accurately reflect the physiological structure of the lower limbs and better simulate the force conditions of local parts of the lower limbs.

[0033] 2. Use YOLO for coarse positioning of key parts instead of key point detection, which reduces the manpower and time cost of fine annotation required for pixel-level key point detection.

[0034] 3. By comparing the mean pixel intensity of the left and right bounding boxes with the difference in other areas, it can automatically detect whether the patient's entire film is implanted with a prosthesis, avoiding the impact of foreign objects such as steel nails and steel plates that are obviously different from the bone structure on deformity detection, which is helpful for postoperative recovery analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] For ease of explanation, the present invention is described in detail with reference to the following specific implementations and accompanying drawings.

[0036] Figure 1 It is a schematic diagram of the workflow of the present invention;

[0037] Figure 2 This is a process flow chart for automatically extracting mechanical axes and detecting deformities according to the present invention;

[0038] Figure 3 This is a flow chart of simulation of the force conditions at various parts of the lower limbs of the present invention;

[0039] Figure 4 This is a comparison diagram before and after the image enhancement processing of the present invention;

[0040] Figure 5 The present invention roughly locates the key parts and extracts the edge set schematic diagram;

[0041] Figure 6 It is the backbone central axis sampling curve and force analysis diagram of the present invention;

[0042] Figure 7 This is a calculation and annotation diagram of the lower limb mechanical axis obtained by the present invention based on rough positioning. DETAILED DESCRIPTION

[0043] Exemplary embodiments will be described in detail herein, examples of which are illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present disclosure. Rather, they are merely examples of devices consistent with certain aspects of the present disclosure, as detailed in the appended claims.

[0044] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0045] like Figure 1-7 As shown, the specific implementation adopts the following technical solutions:

[0046] Example 1:

[0047] S1. Use an X-ray machine to take segmented images of different parts of the lower limb to obtain images that include the morphology of the lower limb, and then stitch them together to obtain a full-length image of the lower limb;

[0048] S2. If the whole film result contains text information from the machine, the text area of ​​the whole film is repaired according to the color to reduce the watermark;

[0049] S3, converting into a binary image and enhancing the image contrast of the repaired image;

[0050] S4. Segment the enhanced image, remove the muscles in the whole image, and retain the bones;

[0051] S5. Perform coarse positioning on the segmented image, obtain the bounding box of the key parts, and determine whether the prosthesis is implanted;

[0052] S6. Extract edge information from the bounding box area and construct an edge set that expresses bone features;

[0053] S7. Calculate the robust mechanical axis based on the bone edge set, analyze the force conditions of the lower limbs, and perform deformity detection.

[0054] The method for determining whether a prosthesis is implanted in the lower limb in step S5 is:

[0055] S51, using the bounding box output by YOLO as the ROI area, and dividing it into two left and right subsets according to the location of the area in the lower limb;

[0056] S52. For each bounding box, count the pixel intensities. When the number of pixels at a certain gray level exceeds a predefined threshold, crop the pixels and evenly distribute the excess pixels to other gray levels. Calculate the mean intensity.

[0057] S53. Compare the difference in the mean pixel intensities of the corresponding left and right areas. If it is less than the threshold, it is considered that both the left and right parts are implanted with prostheses or neither part is implanted with prostheses. Otherwise, it is considered that the patient has only one side implanted with a prosthesis. Further compare it with the pixel mean of the non-critical area. If it is less than the threshold, it is considered that the patient has not been implanted with a prosthesis.

[0058] The specific steps of processing and analyzing the skeleton edge set in step S7 are as follows:

[0059] S71. Use the Canny operator for edge detection and cv2.findContours to obtain bone contour information. For the two key parts at the top, i.e., the femoral area, calculate the pixel intensity in the x-direction, and use the second highest peak and valley as the edge point of the greater trochanter. Similarly, use the first peak and valley as the starting point, calculate the pixel intensity in the y-direction, and use the first peak and valley as the end point. Traverse the points on the contour and draw a perpendicular line along the contour, with the midpoint as the boundary. Accumulate the intersection points of the perpendicular lines and take the average value as the key force point of the femoral head.

[0060] S72. For the key intermediate parts, i.e., the tibia and femur, the tibia and femur are distinguished based on the size of the contour y-coordinate. For the femur ROI, the contour perimeter is calculated and then reduced. This perimeter is used as the threshold for the next step. The contour is approximated using a polygon. The contour points are compared with the true line according to the set threshold. The screening criterion is a percentage threshold of the contour perimeter. Points greater than this value are discarded, while points less than this value are retained. The resulting points are the region vertices. If the number of vertices is less than 4, the process returns to adjust the threshold size.

[0061] S73. Further subdivide the contour into three irregular curved edges by endpoints, among which the lower edge of the femur and the upper edge of the tibia are used as a pair of matching edges. Sample along the y-axis direction and take the midpoint of the intersection of the edge and the straight line to obtain the contact surface between the tibia and femur; connect the center point of the femoral head and the knee joint to obtain the femoral mechanical axis. Similarly, connect the center point of the talus and the center of the knee joint to obtain the tibial mechanical axis, and connect the center point of the femoral head and the center point of the ankle joint to obtain the lower limb mechanical axis.

[0062] Example 2:

[0063] Step 100: Obtain the full-length lower limb data and convert the DICOM format data into PNG format for easy reading by the program;

[0064] Step 200: Detect whether the image has a color mark, and use the neighboring information to repair it to reduce the impact of the watermark on subsequent processing;

[0065] Step 300: converting the image into a binary image, and performing local contrast enhancement on the processed image to highlight edge information;

[0066] Step 400: Segment the enhanced image to obtain an image with the muscle portion removed;

[0067] Step 500: Perform coarse positioning on the segmented image, obtain the ROI area, and detect whether the prosthesis is implanted;

[0068] Step 600: extract edge information from the key part bounding box and construct a bone edge set;

[0069] Step 700: Calculate the mechanical axis of the lower limb based on the bone edge set, sample the bone shaft as the interpolation point, obtain the bone interpolation curve starting from the center of gravity of the edge part, calculate the curvature of the two end points, and the tangent line at the intersection of the straight line where the center of gravity of the tibia and femur is located and the interpolation curve, so as to analyze the deformity of the lower limb.

[0070] The implementation method is as follows:

[0071] First, pydicom is used to read DICOM data, and the pixel intensity values ​​in different DICOM files are normalized to the range of (0, 255). Then, the image is detected to see if there is a color mark, and cv2.inpaint is used to repair it to reduce the impact of the mark watermark on subsequent analysis. The image is further converted into a binary image, and the processed image is enhanced. The target detection algorithm is used to obtain the coarse positioning of the key area, and a closed edge set is constructed for the ROI area to obtain the lower limb mechanical axis annotation.

[0072] Specifically:

[0073] The image contrast enhancement in step 300 includes the following operations:

[0074] 301: Divide the input image into multiple small, overlapping square sub-blocks, where the size of each sub-block can be adjusted according to the image resolution and patient age;

[0075] 302: For each sub-block, calculate its pixel grayscale frequency. If the number of pixels at a certain grayscale level exceeds a predefined threshold, crop it and evenly distribute the excess pixels to other grayscale levels. This can avoid over-enhancing the contrast of certain areas, thereby reducing noise amplification.

[0076] 303: converting the original grayscale distribution into a uniform distribution, thereby enhancing the contrast of the image;

[0077] The mechanical axis marking and deformity analysis in step 700 include the following operations:

[0078] 701: Use the Canny operator for edge detection and cv2.findContours to obtain bone contour information. For the two key areas at the top, namely the femoral region, calculate the pixel intensity in the x direction, and use the second highest peak and valley as the edge point of the greater trochanter. Similarly, use the first peak and valley as the starting point, calculate the pixel intensity in the y direction, and use the first peak and valley as the end point. Traverse the points on the contour and draw a perpendicular line along the contour, with the midpoint as the boundary. Accumulate the intersection points of the perpendicular lines and take the average value as the key force point of the femoral head;

[0079] 702: For the key intermediate areas, namely the tibia and femur, the tibia and femur are distinguished based on the size of the contour's y-coordinate. For the femoral ROI, the contour perimeter is calculated and then reduced to serve as the threshold for the next step. The contour is approximated using a polygon. The contour points are compared with the true line according to the set threshold. The screening criterion is the percentage threshold of the contour perimeter. Points greater than this value are discarded, and points less than this value are retained. The resulting points are the region vertices. If the number of vertices is less than 4, the threshold size is adjusted. The contour is further subdivided into three irregular curved edges by the endpoints. The lower edge of the femur and the upper edge of the tibia are used as a pair of matching edges. Sampling along the y-axis direction and taking the midpoint of the intersection of the edge and the straight line are respectively performed to obtain the contact surface between the tibia and femur.

[0080] 703: Connect the center point of the femoral head and the knee joint to obtain the femoral mechanical axis. Similarly, connect the center point of the talus and the center point of the knee joint to obtain the tibial mechanical axis. Connect the center point of the femoral head and the center point of the ankle joint to obtain the lower limb mechanical axis.

[0081] If the angle between the key force-bearing point of the femoral head and the edge of the greater trochanter and the mechanical axis of the femur is within 90°±5°, the angle between the tangent of the mechanical axis of the femur and the force-bearing surface of the knee joint is within 87.5°±2.5°, the angle between the tangent of the mechanical axis of the tibia and the force-bearing surface of the knee joint is within 87.5°±2.5°, and the angle between the tangent of the mechanical axis of the tibia and the force-bearing surface of the ankle joint is within 89°±3°, the lower limb is considered to be within the normal range. Otherwise, the lower limb is considered to have deformity and requires further treatment.

[0082] 704: Calculate the shortest distance d from the center of gravity of the key part to the boundary. At the position d away from the center of gravity, draw a horizontal line along the x-axis to obtain two intersection points (x 11 ,y1),(x 21 ,y1), and the midpoint coordinates Calculate the angle between the line connecting the midpoint and the center of gravity and the contact surface to reflect the degree of bone tilt at both ends. Determine the number of interpolation nodes, n, based on the degree of tilt and the patient's age. If no slope exists, the end is assumed to be perpendicular to the contact surface. Similarly, convert the bone edge set into a set of left and right edge points.

[0083] 705: Starting from the y coordinate of the center of gravity of the key part, find the intersection point with the edges on both sides along the x direction, recorded as x 01 ,x 02 , then traverse along the y direction to obtain the coordinate sets on both sides {(x 11 ,y1),(x 21 ,y1),…,(x n1 ,y n ),(x n2 ,y n)}, for each pair of matching points (xi1,yi), (xi2,yi), calculate the midpoint, the coordinates of the midpoint are The interpolation node set is The curve C(y) describing the bone morphological characteristics is generated using the cubic spline interpolation method;

[0084] Along the generated curve C(y), for each y coordinate, draw a perpendicular line perpendicular to the curve. The direction of the perpendicular line is obtained by calculating the slope C'(y) of the tangent line of the curve at each point: Generates a series of arrows to describe the forces acting on various parts of the bone.

[0085] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

[0086] Other embodiments of the present disclosure will readily occur to those skilled in the art after considering the specification and practicing the disclosure herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, and the true scope of the present disclosure is indicated by the following claims.

Claims

1. A method for analyzing lower limb deformity based on contour extraction, characterized in that: The following steps are involved: S1. Use an X-ray machine to take segmented images of different parts of the lower limb to obtain images that include the morphology of the lower limb, and then stitch them together to obtain a full-length image of the lower limb; S2. Repair the text area of ​​the entire film according to the color to reduce the watermark; S3, converting into a binary image and enhancing the image contrast of the repaired image; S4. Segment the enhanced image, remove the muscles in the whole image, and retain the bones; S5. Perform coarse positioning on the segmented image, obtain the bounding box of the key parts, and determine whether the prosthesis is implanted; S6. Extract edge information from the bounding box area and construct an edge set that expresses bone features; S7. Calculate the robust mechanical axis based on the bone edge set, analyze the force on the lower limbs, and perform deformity detection; The processing and analysis of the skeleton edge set in step S7 specifically includes the following steps: S71. Use the Canny operator for edge detection and cv2.findContours to obtain bone contour information. For the two key parts at the top, namely the femoral area, count the pixel intensity in the x direction, and use the second highest peak and valley as the edge point of the greater trochanter. Similarly, use the first peak and valley as the starting point, count the pixel intensity in the y direction, and use the first peak and valley as the end point. Traverse the points on the contour and draw a perpendicular line along the contour, with the midpoint as the boundary. Accumulate the intersection points of the perpendicular lines and take the average value as the key force point of the femoral head. S72. For the key intermediate parts, i.e., the tibia and femur, the tibia and femur are distinguished based on the size of the contour y-coordinate. For the femur ROI, the contour perimeter is calculated and then reduced to serve as the threshold for the next step. The contour is approximated using a polygon. The contour points are compared with the true line according to the set threshold. The screening criterion is the percentage threshold of the contour perimeter. S73. Further subdivide the contour into three irregular curved edges based on the endpoints. The lower edge of the femur and the upper edge of the tibia are used as a pair of matching edges. Sample along the y-axis and take the midpoint of the intersection of the edge and the straight line to obtain the contact surface between the tibia and femur. Connect the center point of the femoral head and the knee joint to obtain the mechanical axis of the femur. Similarly, connect the center point of the talus and the center point of the knee joint to obtain the mechanical axis of the tibia. Connect the center point of the femoral head and the center point of the ankle joint to obtain the mechanical axis of the lower limb. The method also includes lower limb force analysis, and the method for generating a force curve for the lower limb force analysis includes the following steps: A1. Calculate the shortest distance d from the center of gravity of the key part to the boundary. Draw a horizontal line along the x-axis at a distance d from the center of gravity to obtain two intersection points. , and the midpoint coordinates ( ), calculate the angle between the midpoint and the center of gravity and the contact surface to reflect the degree of inclination of the bone at both ends, and determine the number of interpolation nodes n according to the degree of inclination and the age of the patient; A2. Starting from the y coordinate of the center of gravity of the key part, find the intersection point with the edges on both sides along the x direction, which is recorded as , then traverse along the y direction to obtain the coordinate sets on both sides , for each pair of matching points (xi1,yi), (xi2,yi), calculate the midpoint, the midpoint coordinates are ( ), the interpolation node set is {( , ),( , ),…,( , )}, use the cubic spline interpolation method to generate the curve C(y) describing the bone morphological characteristics; A3. Draw a perpendicular line to the generated curve C(y) for each y coordinate. The direction of the perpendicular line is obtained by calculating the slope of the tangent line C'(y) at each point on the curve: the slope of the perpendicular line = , and generates a series of arrows to describe the force conditions at various parts of the bone.

2. The method for analyzing lower limb deformity based on contour extraction according to claim 1, characterized in that: The specific steps of determining whether a prosthesis is implanted in the lower limb in step S5 include: S51, inputting the result of the rough positioning as ROI, setting a key part set to store the key parts representing the local features of the bone; S52, dividing each key part into two subsets according to its position in the lower limb region, and calculating the mean pixel intensity within each bounding box according to the starting point and the ending point of each bounding box; S53. Compare the intensity difference between the left and right corresponding areas, and use the threshold α to determine whether the current part is implanted with a prosthesis. If it is less than the threshold, it is considered that the left and right parts are in the same state. Otherwise, it is considered that the patient has been implanted with a prosthesis. Further compare it with the pixel mean of the non-critical part area. If it is less than the threshold, it is considered that the patient has not been implanted with a prosthesis.

Images