CN114191075A

CN114191075A - Rapid construction method and system of personalized knee joint prosthesis model

Info

Publication number: CN114191075A
Application number: CN202111547245.2A
Authority: CN
Inventors: 刘斌; 李鹏熙
Original assignee: Dalian University of Technology
Current assignee: Dalian University of Technology
Priority date: 2021-12-16
Filing date: 2021-12-16
Publication date: 2022-03-18
Anticipated expiration: 2041-12-16
Also published as: CN114191075B

Abstract

The invention discloses a method and a system for quickly constructing an individualized knee joint prosthesis model, wherein the method comprises the following steps: preprocessing all CT images to obtain denoised binary images, and performing surface drawing reconstruction based on the binary images to obtain a plurality of knee joint integral models; cutting the separated femur model and tibia model with the required fixed length; respectively and approximately superposing the healthy models in the femur and tibia models in a registration mode, performing retention adjustment based on a set retention ratio threshold value to obtain an average femur model and an average tibia model which do not contain any healthy individual characteristic, performing registration alignment on the average femur model close to the size of a case and the femur model of the case, displaying and analyzing the condition of the case, determining a diseased part needing to be resected in the case femur model, resecting the diseased area, performing cutting operation at the same position on the average femur model, and correcting a meniscus part in the prosthesis model by combining the accurate position of the femur model at each flexion and extension angle, thereby obtaining the knee joint prosthesis model which accords with the motion characteristic of a human body at each flexion and extension angle.

Description

Rapid construction method and system of personalized knee joint prosthesis model

Technical Field

The invention relates to the field of medical treatment, in particular to a method and a system for quickly constructing an individualized knee joint prosthesis model.

Background

With the development of economy and the advancement of society, safer, faster, more accurate and more humanized medical services are urgently needed by human beings. Therefore, the individual difference of the patient is fully emphasized, and the personalized and precise medical design is combined with the advanced graphic image processing technology, so that the current important scientific research direction is formed. The knee joint is the largest, most complex and more frequent in damage, and the damage of the knee joint will seriously affect the normal life of people. At present, the most common treatment mode aiming at the serious injury of the knee joint is total knee joint replacement, and the artificial knee joint prosthesis is used for replacing the original joint biomechanical function. A large number of knee joint prostheses are needed every year all over the world, but under the influence of genetic factors, living habits and working environments, the bone morphology of each patient has large individual difference, so that the expectation of the patient on the precise operation of the knee joint replacement is stronger and stronger, and the common part type prosthesis is difficult to adapt to the individual requirements of different patients, so that the design of the knee joint replacement prosthesis is inevitably developed towards the customized direction. In the future, a new customized orthopedic surgery mode of ' patient-patient → hospital-image data → personalized prosthesis design unit-image parameter → prosthesis manufacturing unit-personalized prosthesis → hospital-replacement operation → patient ' will be formed, wherein the personalized prosthesis design unit ' is a technological blank to be filled at present.

Disclosure of Invention

According to the problems in the prior art, the invention discloses a rapid construction method of a personalized knee joint prosthesis model, which specifically comprises the following steps:

acquiring CT images of the knee joint, preprocessing all the CT images to obtain denoised binary images, and performing surface drawing reconstruction based on the binary images to obtain a plurality of integral models of the knee joint;

separating a femur model and a tibia model in a plurality of knee joint integral models by adopting a grid connectivity analysis mode, and respectively cutting the separated femur model and tibia model by reserving required fixed lengths by means of a least square method and a coordinate system conversion mode;

selecting a group of femur model data of a healthy human body with BMI indexes close to cases, approximately superposing femur models in a registration mode, setting a retention ratio threshold, determining a mesh model vertex needing to be retained or adjusted based on the retention ratio threshold, performing retention or adjustment operation on the vertex, finally obtaining an average femur model without any healthy case characteristics, and obtaining an average tibia model in the same mode;

simulating a plane tangent diagram based on space vertex projection and a least square method and combining the position relation of a plurality of reference lines of the knee joint rotating shaft to obtain an approximate rotating shaft of the average femur model;

according to the rotation results of the approximate rotation axis simulation average femur model at each angle under the flexion and extension angles, correcting and adjusting the rough simulation rotation results at each flexion and extension angle based on the characteristic that the position relationship between the femoral posterior condylar connecting line axis and the postoperative condylar axis is approximately unchanged in the flexion and extension movement process of the human knee joint, so as to obtain the accurate position of the average femur model at each flexion and extension angle;

registering and aligning the average femur model close to the size of the case with the femur model of the case, displaying and analyzing the disease condition of the case, determining the diseased part needing to be resected in the case femur model and resecting the diseased area, executing cutting operation at the same position on the average femur model, generating an initial femur prosthesis model based on Boolean operation, and obtaining an initial meniscus prosthesis model and a tibial plateau model in the same way;

and correcting the meniscus part in the prosthesis model by combining the accurate positions of the average femur model at all flexion and extension angles, thereby obtaining the knee joint prosthesis model which accords with the motion characteristics of the human body at all flexion and extension angles.

And reading the denoised binary image, carrying out maximum contour screening, corroding and expanding the region encircled by the maximum contour so as to remove scattered points in the CT image and possible interference of a missing part, and finally carrying out surface drawing reconstruction to obtain the whole knee joint model.

The method comprises the steps of obtaining a bottom plane equation of a femur model by adopting least square fitting, establishing a new coordinate system by taking the plane as an xoy plane and a normal vector of the plane as a z axis, determining the position of the top end of the femur model by finding the maximum z value in the coordinate system, and reserving the top end downwards at a required fixed length, so that each group of subsequent femur models participating in operation are approximately equal in length, and reserving the tibia model for cutting at the required fixed length in the same mode.

When obtaining the mean femur model: for each healthy femur model of the selected BMI index approaching case, all the healthy femur models are moved to the same position in space by using an improved ICP (inductively coupled plasma) registration method, and the vertexes of the mesh model are determined to be reserved and adjusted based on a set reservation ratio threshold, wherein the specific mode is as follows: assuming that the number of models is n, the set retention ratio threshold is k, if the number of vertexes near a certain vertex is greater than k × n, the vertex adjusts the position according to the mean value by combining all other vertexes near the vertex, otherwise, the vertex is regarded as a special point and is not retained; the resulting set of mean femoral models did not contain any healthy case features.

Obtaining the approximate axis of rotation of the mean femoral model: fitting the bottom surface of the average femur model based on a least square method, setting a certain simulated thickness, taking the bottom surface as a reference surface and a normal vector thereof as a direction, performing simulated cutting on the average femur model and projecting to obtain a simulated planar tangent diagram of the average femur model, finding an image with the largest area in all the simulated planar tangent diagrams, obtaining a reference line femur posterior malleolus connecting line of a femur rotation simulated shaft by analyzing the vertex position of a minimum convex hull in the image, and determining the position of the postoperation condyle axis by means of the position relation that the reference line femur posterior malleolus connecting line axis is approximately parallel to the postoperation condyle axis, wherein the axis is an approximate rotating shaft of the average femur model.

When the rough simulation rotation result under each flexion and extension angle is corrected and adjusted: the method comprises the steps of preliminarily simulating the rotation positions of different angles of the femur by means of an approximate rotation shaft, marking and analyzing the pre-rotation position and the post-rotation position of two characteristic points of the femur, which are initially approximately parallel to a tibial plateau plane, and calculating a rotation correction shaft, a rotation correction angle and a correction displacement amount on the basis of the angle relation between the longest shafts of the femur and the tibial coronal plane and the two characteristics that the minimum distance between the femur and the tibia is approximately unchanged in the human body flexion-extension motion process after the femur rotates by a certain angle along the approximate rotation shaft.

Obtaining an initial knee prosthesis model: for the tibia model, selecting the affected area to be resected, fitting all vertexes of the tibia section by least square, setting the vertex set as A, and respectively performing distance s on A along the normal vector direction of a plane₁、s₂To obtain a set B and a set C of vertices after translation, where s₁<s₂Generating a tibial plateau model in the knee joint prosthesis model close to the size of the completely fitted case incision by combining the set A and the set B, and generating an initial state of a meniscus prosthesis model in the knee joint prosthesis model by combining the set B and the set C; for the femoral model, the remaining parts are spliced into the femoral prosthesis model by selecting the areas needing to be reserved; and performing difference operation in Boolean operation on the obtained femoral prosthesis model and the initial state of the meniscus prosthesis, so as to remove the part which can be overlapped with the femoral prosthesis model in the initial state of the meniscus prosthesis model.

A system for constructing a personalized knee prosthesis model, comprising:

the preprocessing module is used for reading the CT images of the knee joint, preprocessing all the CT images to obtain denoised binary images, and performing surface drawing reconstruction based on the binary images to obtain a plurality of integral models of the knee joint;

the separation cutting module is used for separating a femur model and a tibia model in a plurality of knee joint integral models in a grid connectivity analysis mode, and simultaneously cutting the separated femur model and tibia model in a required fixed length by means of a least square method and a coordinate system conversion mode;

the health average model calculation module selects a group of femur model data of a healthy human body with BMI indexes close to cases, approximately superposes all femur models in a registration mode, sets a retention ratio threshold, determines the peaks of a retention and adjustment grid model based on the retention ratio threshold, finally obtains an average femur model without any health case characteristics, and obtains a healthy average tibia model in the same mode;

the approximate rotating shaft acquisition module is used for simulating a plane tangent diagram based on space vertex projection and a least square method and obtaining an approximate rotating shaft of a healthy average femur model by combining the position relation of a plurality of reference lines of a knee joint rotating shaft;

the rough rotation correction adjustment module is used for simulating the rotation result of each angle of the femur model at the flexion and extension angles by the approximate rotation shaft, and correcting and adjusting the rough simulation rotation result at each flexion and extension angle based on the characteristic that the position relationship between the femoral posterior ankle connecting line axis and the surgical posterior condyle axis is approximately unchanged in the flexion and extension movement process of the knee joint of the human body, so that the accurate position of the average femur model at each flexion and extension angle is obtained;

the personalized prosthesis initialization module is used for registering and aligning the average femur model close to the size of the case with the femur model of the case, displaying and analyzing the state of the case, determining the diseased part needing to be resected in the case femur model and resecting the diseased area, executing cutting operation at the same position on the average femur model, generating an initial femur prosthesis model based on Boolean operation, and obtaining an initial meniscus prosthesis model and a tibial plateau model in the same way;

and the individualized prosthesis optimization output module corrects the meniscus part in the prosthesis model by combining the accurate positions of the femur model at all flexion and extension angles, so as to output the knee joint prosthesis model which accords with the motion characteristics of the human body at all flexion and extension angles.

By adopting the technical scheme, the method and the system for quickly constructing the personalized knee joint prosthesis model provided by the invention can obtain the needed personalized knee joint prosthesis model data which completely adapts to a certain patient only by giving a plurality of groups of healthy CT images and CT images of cases by a user in the implementation process and even only needing one set of CT images of cases after the later database is perfected, thereby realizing the quick design of the personalized knee joint prosthesis model aiming at a specific case.

Drawings

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments described in the present application, and other drawings can be obtained by those skilled in the art without creative efforts.

FIG. 1 is a flow chart of the implementation of the method of the present invention

FIG. 2 is a CT image of human knee joint inputted in the present invention

FIG. 3 is a diagram illustrating the effect of the knee joint CT image after denoising in the present invention

FIG. 4 is a schematic representation of the femur of the knee joint model of the present invention after separation

FIG. 5 is a view of the tibia after the knee joint model of the present invention has been separated

FIG. 6 is a diagram of a patella after separation of a knee joint model according to the present invention

FIG. 7 is a diagram illustrating the effect of cutting the femur according to the present invention

FIG. 8 is a diagram illustrating the cutting effect of the tibia according to the present invention

FIG. 9 is a graph of mean femur model results of the present invention

FIG. 10 is a graph of mean tibial model results for the present invention

FIG. 11 is a diagram illustrating the result of the preliminary rotation at various angles according to the present invention

FIG. 12 is a diagram showing the results of the rotation correction at various angles in the present invention

FIG. 13 is a diagram of the effect of the tibial plateau portion of the knee prosthesis model of the present invention

FIG. 14 is a diagram illustrating the effect of meniscus on a prosthetic knee joint model of the present invention

FIG. 15 is a diagram of the effect of the femoral part of the knee joint prosthesis model of the present invention

FIG. 16 is a graph showing the effect of meniscus formation in a prosthetic knee joint model of the present invention

FIG. 17 is a graph of the effect of meniscus adjustment in the present invention

FIG. 18 is a final effect diagram of the knee prosthesis of the present invention

FIG. 19 is a view showing a system configuration of the present invention

Detailed Description

In order to make the technical solutions and advantages of the present invention clearer, the following describes the technical solutions in the embodiments of the present invention clearly and completely with reference to the drawings in the embodiments of the present invention:

in the rapid construction method of the personalized knee joint prosthesis model shown in fig. 1, a plurality of groups of separated femur and tibia models are obtained by grouping, denoising, reconstructing and separating original data in the implementation process, and the result is shown in fig. 4-6. In a plurality of three-dimensional femur models and tibia models in a group, the same length of remaining cuts are respectively made, and after the registration, an average bone model which is finally retained is calculated according to a set retention ratio threshold, and the result is shown in fig. 9 and fig. 10. The obtained average femur model was subjected to positioning of the approximate rotation axis of the femur by means of the method of simulating planar slicing and the physiological characteristic line, and then the rotation result was corrected based on the characteristics of the human knee joint motion, and the result is shown in fig. 12. Based on the initial position of the tibia, the position of the femur at different rotation angles, and the case, it is determined to generate a knee prosthesis model matching therewith, and the result is shown in fig. 18. The method disclosed by the invention comprises the following specific steps:

s1: inputting CT images of the healthy human knee joints, grouping, denoising and three-dimensional reconstructing the images to obtain a plurality of groups of healthy knee joint integral models:

s11, firstly, the human knee joint CT images are grouped according to BMI (height/(weight square) and gender (according to statistics, the factors are main factors influencing the size of the knee joint) in the patient information, so that a plurality of groups of knee joint CT images with similar sizes are obtained, and the CT images are converted into gray level images, namely pixel values (r) of each pixel point of the traversal images_k,g_k,b_k) Num, num is the number of pixel points, and the calculated gray value is:

gray_k＝0.299*r_k+0.587*g_k+0.114*b_k

r_k＝gray_k，g_k＝gray_k，b_k＝gray_k

by processing each pixel point, the image can be converted into a gray image.

And carrying out binarization processing on the image, and setting the value of the pixel point larger than the threshold value to be 255 and the value of the pixel point smaller than the threshold value to be 0 by adopting a self-adaptive threshold value mode.

And S12, carrying out contour positioning on the binary image, reserving a region contained in the maximum contour in the image, setting the values of pixel points in other regions to be 0, and then carrying out expansion and corrosion operations with operation cores of 3 x 3 on the image in sequence so as to fill up small holes and small cracks in the image, wherein the result is shown in figure 3.

S13, performing surface rendering on the denoised image for reconstruction, extracting an isosurface by using vtkContourFilter, and setting a value range to be (0, 170); using vtkSmoothPolyDataFilter to carry out grid smoothing, and setting the iteration number to be 300; using vtkk polytatanormals to calculate the normal vector, the eigenangles are set to 30; the triangular meshes are connected using vtkslipper. Finally, rendering is carried out through a pipeline of the VTK to obtain a reconstruction result of the human knee joint, an integral model (comprising three parts of a femur, a tibia and a patella) of the knee joint is obtained, and the three parts are separated into three independent models by confirming the connectivity of each part grid in the model, so that the subsequent operation is facilitated, and the result is shown in fig. 4, 5 and 6.

S2: the knee joint models in each group are controlled to be the same length in the following way:

s21, fitting a bottom surface equation of the femur model by using a least square method, taking the plane as an xoy plane of a new coordinate system, taking a normal vector of the xoy plane as the positive direction of a z axis of the new coordinate system, and converting the femur model into a new coordinate system through coordinate system conversion:

Ax+By+Cz+D＝0

wherein [ A, B, C ] is the normal vector direction of the plane obtained by fitting.

S22, after the femur model is transferred to a new coordinate system, finding a point with the maximum z value in the femur model, wherein the point is the vertex of the femur model, setting the unified needed length of the femur model to be 80mm, and reserving model parts of all the vertices in the range of (max _ z-80) -max _ z.

max_z＝max[z_i],i＝1,2…n

Wherein [ z ] is_i]Is the z-coordinate value of a certain point in the single femoral model.

S23, the tibia model is controlled in the same manner and in the same length, and the result is shown in fig. 7 and 8.

S3: obtaining an average model of each group of bone models by the following method:

s31, assuming that the number of the femoral models participating in budget is m, and taking the first femoral model as a reference, respectively calculating the sum of the minimum distance values from each point to the first femoral model for the 2 nd to the m th femoral models based on an improved ICP algorithm:

wherein d is_iThe distance between the ith point in a femur (2-m) and the first femur model is shown, and n is the total number of the midpoints of the femurs (2-m).

The other femoral (2-m) models can be aligned to achieve the maximum degree of overlap with the first femoral model.

S32, searching in all superimposed model spaces by taking each vertex of the first model as a center and the minimum distance between the vertex of the first model and other points in the model as a radius:

p∈P_i,d<r

wherein P is any point in the superimposed model, d is the distance from the point to the ith vertex in the first model, and P_iIs the set of points searched for by the ith vertex in the first model.

S33, assuming that the set retention ratio threshold is 0.9 (can be increased or decreased according to the precision requirement),if the above-mentioned point set P is searched for a certain one_iIf the number of included points is greater than 0.9 × m, the vertices of the point set are retained in the form of coordinate mean, otherwise, the point set is discarded to obtain a mean femur model, and a mean tibia model is obtained in the same manner, with the results shown in fig. 9 and 10.

S4: the approximate rotating shaft of the average femur model in the human body flexion and extension motion is obtained by the following method:

s41, fitting an average femur model bottom surface equation by using a least square method, taking the plane as an xoy plane of a new coordinate system, taking a normal vector of the xoy plane as the z-axis direction of the new coordinate system, and converting the average femur model into the following conditions in the new coordinate system through coordinate system conversion:

Ax+By+Cz+D＝0

where [ A, B, C ] is the normal vector direction of the plane.

S42, dividing the model into 1mm thick sections (0-1 mm, 1-2 mm … …) according to the z value, projecting the space points of each layer onto different planes along the negative direction of the z axis to obtain simulated plane tangent diagrams of the average femur model, and then selecting one of all the simulated plane tangent diagrams with the largest outline area as a target tangent diagram.

S43, firstly finding out the centroid p of the maximum outline in the target tangent diagram₁Then find out the sum p in the contour₁Nearest point p₂Then, the minimum bounding convex hull of the maximum contour is calculated, and the vertex distance p in the minimum bounding convex hull is calculated₁Two nearest points p₃、p₄According to the definition of the medical reference line of the knee joint, p₃、p₄The two points are the reference line PCA (Posterior femoral Axis) of the femoral rotation simulation shaft, after the reference line PCA of the femoral rotation simulation shaft is obtained, the position of the sTEA is determined by means of the position relation that the Axis of the PCA is approximately parallel to the Axis of the sTEA (Posterior Condylar Axis) so as to obtain the approximate rotation shaft of the femoral model in the human body flexion and extension motion.

S5: the rotation result of the femur model is corrected in the following way:

s51, preliminarily simulating strands according to the approximate rotating shaft obtained in S4The different angular rotational positions of the bone, the result of which is shown in fig. 11. Record the position line of the PCA axis in S4 during rotation₁And the position line of the PCA axis after rotation₂。

m₁₁＝u*u+(v*v+w*w)*cost

m_i2＝u*v*(1-cost)-w*sint

m₁₃＝u*w*(1-cost)+v*sint

m₁₄＝a*(v*v+w*w)-u*(b*v+c*w))*(1-cost)+(b*w-c*v)*sint

m₂₁＝u*v*(1-cost)+w*sint

m₂₂＝v*v+(u*u+w*w)*cost

m₂₃＝v*w*(1-cost)+u*sint

m₂₄＝b*(u*u+w*w)-v*(a*u+c*w))*(1-cost)+(c*u-a*w)*sint

m₃₁＝u*w*(1-cost)-v*sint

m₃₂＝v*w*(1-cost)+u*sint

m₃₃＝w*w+(u*u+v*v)*cost

m₃₄＝c*(u*u+v*v)-w*(a*u+b*v))*(1-cost)+(a*v-b*u)*sint

Wherein, [ x ]⁰ y⁰ z⁰]Is the spatial coordinate of a point prior to rotation, [ x y z ]]Is the spatial coordinate of the point after rotation, [ ab c]Is the coordinate of any point on the rotation axis, [ uv w]Is the direction vector of the rotation axis.

S52, analyzing line based on the characteristic that the connecting line of the rear ankles of the thighbone should be kept approximately parallel in the process of human body flexion and extension movement₁And line₂If line₁、line₂If they are not parallel, then the correction is not requiredline₁、line₂The included angle is the rotation correction angle, the rotation correction axis is the straight line of the normal vector of the plane determined by the two straight lines, and the straight line passes through line₁、line₂Determining the rotation correction axis and the rotation correction angle, performing a second rotation on the femur model, and recording the position of the PCA axis after the second rotation as a line₃。

S53, calculating line based on the characteristic that the minimum distance between the femur and the tibia is approximately unchanged in the process of human body flexion and extension movement₃And line₁Determining the correction displacement amount, if the shortest distance between the two is 0, then no correction is needed, otherwise, performing translation correction on the femur model after the second rotation, wherein the correction distance is a shortest distance line segment, and the direction is line₃Pointing to a line₁Finally, accurate positions of the femoral model at various flexion and extension angles are obtained, and the result is shown in fig. 12.

Wherein, [ x ]⁰ y⁰ z⁰]Is the spatial coordinate of a point prior to translation, [ x ]^- y^- z^-]Is line₃And line₁Determined displacement vector, [ x y z ]]Is the spatial coordinate of the point after translation.

S6: the initial model of the knee joint prosthesis is obtained by the following method:

s61, processing except grouping in S1 and S2 is carried out on the CT image of the knee joint of the patient to obtain a femur model and a tibia model which are separated from each other, then the femur model of the patient and the corresponding healthy femur model are respectively led into a computer, the area needing to be resected is determined according to the diseased condition of the patient, and simultaneously, the cutting operation with the same position is carried out on the healthy model while the cutting of the femur model of the patient is simulated by means of alignment.

S62, selecting a diseased area to be resected for the tibia model, fitting all vertexes of the tibia section through least square, setting the vertex set as A, and enabling the A to be along the direction of a plane normal vectorAre respectively carried out for a distance s₁、s₂To obtain a set B and a set C of vertices after translation, where s₁＜s₂The tibial plateau model in the knee prosthesis model that closely fits the size of the case incision was generated by combining set a and set B, and the initial state of the meniscal prosthesis model in the knee prosthesis model was generated by combining set B and set C, and the tibial plateau model results are shown in fig. 13 for the initial state of the meniscal prosthesis model as shown in fig. 14.

S63, for the femoral part, because the curved surface of the prosthesis model part is complex, the remaining part is spliced into the femoral prosthesis model by selecting the area needing to be reserved; and performing difference operation in Boolean operation on the obtained femoral prosthesis model and the initial state of the meniscus prosthesis, so as to remove the part which can be overlapped with the femoral prosthesis model in the initial state of the meniscus prosthesis model, and obtain the meniscus prosthesis model. The femoral prosthesis model is shown in fig. 15 and the meniscal prosthesis model is shown in fig. 16.

Where x is any vertex in the menisci of the set, set a represents the initial model of the meniscus, and set B represents the femoral prosthesis model.

S7: the method for acquiring the personalized knee joint prosthesis model which accords with the motion characteristics of the human body under each flexion and extension angle comprises the following steps: and correcting the meniscus model in the prosthesis part obtained in the step S6 by combining the accurate position of the femur model at each flexion and extension angle obtained in the step S5, performing difference operation in Boolean operation on the meniscus model and the femur model at each position in sequence, and combining the adjusted meniscus part, the femur prosthesis part and the tibia prosthesis part obtained in the previous step to obtain the personalized knee joint prosthesis model which accords with the motion characteristics of the human body at each flexion and extension angle. The correction effect is shown in fig. 17, and the final result is shown in fig. 18.

As shown in fig. 19, a construction system of a personalized knee joint prosthesis model includes:

The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and the technical solutions and the inventive concepts thereof according to the present invention should be equivalent or changed within the scope of the present invention.

Claims

1. a kind of rapid construction method of individualized knee joint prosthesis model is characterized in that comprising:

Obtaining CT images of knee joints, preprocessing all CT images to obtain denoised binary images, and performing surface rendering and reconstruction based on binary images to obtain multiple overall models of knee joints;

The femur model and tibia model in multiple knee joint overall models are separated by grid connectivity analysis method, and the separated femur model and tibia model are respectively reserved for the required fixed length by means of least squares method and coordinate system transformation. cut;

Select a group of femoral model data of healthy people with a BMI index close to the case, use the registration method to approximately superimpose all the femoral models, set the retention ratio threshold, and determine the mesh model vertices that need to be retained or adjusted based on the retention ratio threshold. The vertices are preserved or adjusted, and finally the average femur model that does not contain any healthy individual case characteristics is obtained, and the average tibia model is obtained in the same way;

The approximate rotation axis of the average femur model is obtained based on the spatial vertex projection, the least squares method to simulate the plane tangent graph and the positional relationship of multiple reference lines of the knee joint rotation axis;

According to the approximate rotation axis to simulate the rotation results of the average femoral model at various angles of flexion and extension, based on the characteristics that the positional relationship between the axis of the posterior femoral malleolus and the axis of the condyle after surgery is approximately unchanged during the flexion and extension of the human knee Roughly simulate the rotation results for correction and adjustment, so as to obtain the accurate position of the average femoral model under each flexion and extension angle;

Align the average femur model close to the case size with the case's femur model, display and analyze the case condition, determine the diseased part of the case femur model to be resected and resect the diseased area, perform the same position on the average femur model The cutting operation is based on the Boolean operation to generate the initial femoral prosthesis model, and the initial meniscus prosthesis model and tibial plateau model are obtained in the same way;

The meniscus part in the prosthesis model is corrected based on the precise position of the average femoral model under each flexion and extension angle, so as to obtain a knee joint prosthesis model that conforms to the motion characteristics of the human body under each flexion and extension angle.

2. The method for rapidly constructing a personalized knee joint prosthesis model according to claim 1, wherein the denoised binary image is read and the maximum contour is screened, and the area circled by the maximum contour is corroded , expansion to remove the scatter in the CT image and the possible interference of missing parts, and finally the surface rendering and reconstruction is performed to obtain the overall model of the knee joint.

3. the rapid construction method of individualized knee joint prosthesis model according to claim 1, it is characterized in that: adopt least squares fitting to obtain the bottom plane equation of femur model, take this plane as xoy plane again, its normal vector A new coordinate system is established for the z-axis, in which the position of the top of the femur model is determined by finding the maximum z-value, and after the top is found, the required fixed length is reserved downward to ensure that each subsequent group participates in the calculation. The femoral models are all approximately equal in length, and the tibia models are cut in the same way to retain the desired fixed length.

4. the rapid construction method of individualized knee joint prosthesis model according to claim 1, is characterized in that: when obtaining the average femur model: for each healthy femur model of the selected BMI index approaching case, utilize the improved ICP registration The method moves all of them to the same position in space, and determines the retention and adjustment of mesh model vertices based on the set retention ratio threshold. The specific method is: assuming that the number of models is n, the set retention ratio threshold is k, if a If the number of vertices near a vertex is greater than k*n, then the vertex is combined with all other nearby vertices to adjust the position according to the mean value, otherwise it is regarded as a singular point and will not be retained; finally, the average femur model of the group that does not contain any healthy individual case characteristics is obtained.

5. The method for rapidly constructing a personalized knee joint prosthesis model according to claim 1, wherein: when obtaining the approximate rotation axis of the average femur model: fitting the bottom surface of the average femur model based on the least squares method, and setting a certain simulation Thickness, take the bottom surface as the reference plane and its normal vector as the direction, perform simulated cutting and projection of the average femoral model to obtain a simulated plane cut of the average femur model, and find the image with the largest area in all the simulated plane cuts , in this image, by analyzing the vertex position of the smallest convex hull, the reference line of the femoral rotation simulation axis is obtained by connecting the posterior femoral malleolus, which is approximately parallel to the axis of the posterior femoral malleolus after the operation. The location of the condyle axis, which is the approximate axis of rotation of the average femoral model.

6. The rapid construction method of the personalized knee joint prosthesis model according to claim 1, characterized in that: when correcting and adjusting the rough simulated rotation results under each flexion and extension angle: preliminarily simulate different angles of the femur by means of an approximate rotation axis After the femur rotates by a certain angle along the approximate rotation axis, based on the angular relationship between the longest axis of the coronal plane of the femur and the tibia and the minimum distance between the femur and the tibia during the flexion and extension of the human body are approximately unchanged. Two characteristics, mark and analyze the pre-rotational position and post-rotational position of two feature points of the femur that are initially nearly parallel to the tibial plateau plane, and calculate the rotation correction axis, rotation correction angle and correction displacement.

7. The method for rapidly constructing a personalized knee joint prosthesis model according to claim 1, characterized in that: when acquiring the initial knee joint prosthesis model: for the tibia model, select the diseased area that needs to be resected, and pass the least two Multiply all the vertices of the fitted tibial section, set the vertex set as A, and translate A along the plane normal vector direction with distances s ₁ and s ₂ respectively, and obtain the translated vertex set B and set C, where s ₁ < s ₂ , combine set A and set B to generate the tibial plateau model in the knee joint prosthesis model that is close to the size of the incision of the case, and combine set B and set C to generate the initial state of the meniscus prosthesis model in the knee joint prosthesis model ; For the femoral model, select the area that needs to be reserved to splicing the remaining parts into a femoral prosthesis model; perform the Boolean difference operation between the obtained femoral prosthesis model and the initial state of the meniscal prosthesis, so as to combine the meniscal prosthesis The part of the model that will overlap with the femoral prosthesis model in the initial state is removed.

8. A system for constructing a personalized knee joint prosthesis model, comprising:

The preprocessing module reads knee CT images, preprocesses all CT images to obtain denoised binary images, and performs surface rendering and reconstruction based on the binary images to obtain multiple overall knee joint models;

Separate the cutting module, use the grid connectivity analysis method to separate the femur model and tibia model in multiple knee joint overall models, and use the least squares method and coordinate system transformation to keep the separated femur model and tibia model respectively. required fixed length cuts;

The healthy average model calculation module selects a group of femoral model data of healthy people with a BMI index close to the case, uses the registration method to approximately superimpose all the femoral models, sets the retention ratio threshold, and determines the retention and adjusts the grid model based on the retention ratio threshold. vertices, and finally get the average femur model that does not contain any characteristics of healthy individual cases, and obtain the healthy average tibia model in the same way;

The approximate rotation axis acquisition module, based on the spatial vertex projection, the least squares method to simulate the plane tangent graph, and combined with the positional relationship of multiple reference lines of the knee joint rotation axis to obtain the approximate rotation axis of the healthy average femur model;

Coarse rotation correction adjustment module, the approximate rotation axis simulates the rotation results of each angle of the femoral model under the flexion and extension angle, based on the characteristics that the positional relationship between the axis of the posterior malleolus of the femur and the axis of the postoperative condyle is approximately unchanged during the flexion and extension of the human knee. The rough simulated rotation results under the flexion and extension angle are corrected and adjusted, so as to obtain the precise position of the average femoral model under each flexion and extension angle;

The personalized prosthesis initialization module registers and aligns the average femoral model close to the size of the case with the femoral model of the case, displays and analyzes the condition of the case, determines the diseased part of the femoral model of the case that needs to be resected, and resects the diseased area. Perform the same cutting operation on the average femur model, generate the initial femoral prosthesis model based on Boolean operation, and obtain the initial meniscus prosthesis model and tibial plateau model in the same way;

The personalized prosthesis optimization output module, combined with the precise position of the femur model under each flexion and extension angle, corrects the meniscus part in the prosthesis model, so as to output the knee joint prosthesis model that conforms to the human motion characteristics under each flexion and extension angle.