A knee replacement surgery planning method, device and equipment based on X-ray images and a storage medium

Abstract

The application provides a knee joint replacement surgery planning method and device based on X-ray images, equipment and a storage medium. The method comprises the following steps: acquiring two orthogonal X-ray images of a knee joint; using a statistical shape model method to three-dimensionally reconstruct a three-dimensional model of a femur and a tibia of the knee joint from the two input X-ray images, and outputting key point positions; recommending a prosthesis model and a pose according to the three-dimensionally reconstructed model and the key point positions; confirming the inspection result by a doctor by using the spatial relationship constructed by the key points, and further adjusting the prosthesis pose according to needs, and finally outputting a knee joint prosthesis replacement surgery planning scheme. The method disclosed by the application uses the statistical shape model method to three-dimensionally reconstruct a two-dimensional X-ray image into a three-dimensional bone model and visualize the three-dimensional bone model, which is beneficial to the doctor to more intuitively observe the prosthesis positioning condition, so that the surgery scheme can be better outputted. Moreover, the method disclosed by the application has a relatively low cost and can cover more people.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence image processing technology, and more particularly to the field of intelligent processing technology of medical images, specifically to a method, device, equipment and storage medium for knee replacement surgery planning based on X-ray images. Background Technology

[0002] Total knee arthroplasty (TKA) is one of the most effective treatments for severe knee osteoarthritis and other related conditions. Successful TKA surgery relies on accurate, comprehensive, and detailed surgical planning, which typically involves determining the size and appropriate positioning of the replacement prosthesis. Currently, there are generally two methods for planning knee replacement surgery: three-dimensional planning based on CT images and two-dimensional planning based on X-ray images.

[0003] CT-based prosthesis planning methods can provide three-dimensional information, which can be applied to robot-assisted surgery. However, CT images are expensive and not suitable for the general population. X-ray images are relatively inexpensive and can cover a wider range of people. Doctors can develop planning schemes on X-ray images based on prosthesis templates or two-dimensional template simulation software provided by manufacturers. However, the process requires doctors to repeatedly compare until a suitable prosthesis is matched. This process is cumbersome, time-consuming, and inefficient. In addition, two-dimensional planning based on X-ray images cannot provide doctors with more three-dimensional information and cannot be applied to robot-assisted surgery.

[0004] Therefore, there is an urgent need for a knee replacement surgery planning method that can provide three-dimensional information, allowing doctors to more intuitively observe the prosthesis placement, while also being relatively low-cost and able to cover a wider range of people. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art described in the background section and provide a method, apparatus, device, and storage medium for knee replacement surgery planning based on X-ray images. This method utilizes statistical shape modeling to reconstruct a three-dimensional skeletal model from two-dimensional X-ray images and visualize it, allowing doctors to more intuitively observe the prosthesis placement and thus better develop surgical plans. Furthermore, the method described in this invention is low-cost and can cover a wider range of patients.

[0006] This invention is achieved through the following technical solution: Firstly, this invention provides a method for planning knee replacement surgery based on X-ray imaging, comprising the following steps:

[0007] S101, acquire two orthogonal X-ray images of the knee joint, including the anteroposterior view (AP) and the lateral view (RL).

[0008] S102, using the statistical shape model method, three-dimensional reconstruction of the femur and tibia of the knee joint is performed on the two input X-ray images, and the key point positions of each are output.

[0009] S103 recommends prosthesis type and position based on the three-dimensional model of the femur and tibia at the knee joint and the location of key points;

[0010] S104 utilizes the spatial relationships constructed from key points, allowing doctors to confirm examination results and further adjust the prosthesis position as needed, ultimately outputting a knee joint prosthesis replacement surgery plan.

[0011] Furthermore, the statistical shape model method described in S102 is used to reconstruct three-dimensional models of the femur and tibia of the knee joint from the two input X-ray images, and outputs the key point locations of each, including:

[0012] S201, pre-trained statistical shape models of the femur and tibia, and labeled the key point locations respectively. The statistical shape model includes the average shape and deformation parameters.

[0013] S202: Obtain the contours of two X-ray images, align the X-ray images according to the contour information, and obtain the initial position and scale of the femur and tibia 3D models;

[0014] S203, Load the pre-generated statistical shape model of the femur and tibia, and apply the scale and position of the initial 3D model of the femur and tibia to its average shape model as the initial statistical shape model of the femur and tibia;

[0015] S204, using AP plane contour and RL plane contour to iterate, update the parameters of the femoral-tibial statistical shape model, and generate the final femoral-tibial statistical shape model and key point positions;

[0016] S205: Obtain shape point cloud based on femoral and tibia statistical shape model, calculate the normal vector of point cloud, and use Poisson surface reconstruction method to generate surface model from point cloud to realize visualization of femoral and tibia three-dimensional model.

[0017] Furthermore, the pre-trained statistical shape models of the femur and tibia described in S201, and the annotation of key point locations, include:

[0018] S301, collect knee joint CT image samples from different patients, and generate a set of shape vector samples of the femur and tibia of the knee joint;

[0019] S302, Align the shape vectors in the shape vector sample set;

[0020] S303, process the aligned sample shape vectors to obtain parameterized statistical shape models of the femur and tibia of the knee joint;

[0021] S304: Select key points on the statistical shape model of the femur and tibia and obtain their coordinate positions.

[0022] Furthermore, S301 describes collecting knee CT image samples from different patients to generate a set of shape vector samples of the femur and tibia of the knee joint, including:

[0023] S401, collect knee joint CT image samples from different patients;

[0024] S402, after segmentation, obtain segmentation model samples of the femur and tibia at the knee joint;

[0025] S403, after the segmentation model samples are centered and scaled, a normalized shape vector is obtained;

[0026] S404, apply the same scale coefficient to the normalized shape vectors simultaneously to obtain shape vectors with uniform scale;

[0027] S405, Initial lower alignment of femoral model samples in three-dimensional space, and initial upper alignment of tibial model samples in three-dimensional space;

[0028] S406, extract the discrete coordinate points of the surface of the bone segmentation model as the initial shape vector representation to obtain the shape vector sample set of the femur and tibia of the knee joint.

[0029] Furthermore, the alignment of shape vectors in the shape vector sample set described in S302 includes:

[0030] S501, the sample with the fewest number of sample points in the shape vector sample of the femur and tibia is selected as the reference sample;

[0031] S502, determine whether the other samples in the femoral and tibia shape vector sample sets, excluding the reference sample, are aligned with the reference sample. If all samples are aligned, proceed to step S507; if there are misaligned samples, proceed to step S503 for iteration.

[0032] S503: Select one sample from the femoral and tibial shape vector sample set other than the reference sample, and use the nearest point principle to determine the one-to-one correspondence between the selected sample and the midpoint of the reference sample.

[0033] S504, use the orthogonal Pluke analysis method to calculate the rotation transformation matrix between point pairs, apply the same rotation transformation to the sample shape vector, and calculate the Pluke distance between point pairs;

[0034] S505, determine whether the Prouk distance is less than the threshold. If it is less, take the points belonging to the sample in the point pair as the shape vector of the sample, and continue to S502 to determine whether there are unaligned samples in the sample set at this time. If it exceeds the set threshold, go to S503 to re-acquire the point pair set of the transformed sample and the reference sample and continue iterating until the threshold of the selected sample meets the requirements.

[0035] S506, perform the same processing on all samples in the femoral and tibial shape vector sample sets to align them with the reference sample;

[0036] S507, obtain the shape vectors of all samples after final alignment, the aligned sample shape vectors have the same number of coordinate points.

[0037] Furthermore, the processed and aligned sample shape vectors described in S303 yield a parameterized statistical shape model of the femur and tibia of the knee joint, including:

[0038] S601, sample N points using a uniform distribution on the aligned sample shape vector;

[0039] S602, Calculate the average shape vector X_avg based on the latest sample shape vector.

[0040] S603, the sample shape vector is obtained by subtracting the average shape vector from the sample shape vector;

[0041] S604. Using the singular value decomposition method, the deviation shape vector analysis of the sample is used to obtain the corresponding eigenvector matrix P and deformation coefficient b, resulting in the parameterized statistical shape model SSM=X_avg+Pb for the femur and tibia of the knee joint.

[0042] Furthermore, S204 describes iterating using the AP plane contour and RL plane contour to update the parameters of the femoral-tibial statistical shape model, generating the final femoral-tibial statistical shape model and key point locations, including:

[0043] S701 uses the RL planar profile as a reference to iterate the statistical shape model of the femur and tibia to obtain an updated statistical shape model;

[0044] S702, Project the updated statistical shape model onto the AP plane and calculate the average Euclidean distance between the projected profile and the AP skeletal profile;

[0045] S703, determine if the average Euclidean distance is less than the threshold. If it is less, go to S707 to output the final statistical shape model of the femur and tibia; if it is greater, go to S704 to continue the iteration.

[0046] S704 uses the AP plane profile as a reference to iterate the statistical shape model of the femur and tibia to obtain an updated statistical shape model;

[0047] S705: Project the updated statistical shape model onto the RL plane and calculate the Euclidean distance between the projected profile and the RL skeletal profile.

[0048] S706: Determine if the Euclidean distance is less than the threshold. If it is less, proceed to S707 to output the final statistical shape model of the femur and tibia. If it is greater, proceed to S701 to continue iterating until the iteration requirement and exit condition are met.

[0049] S707 generates the final statistical shape model of the femur and tibia.

[0050] Furthermore, as described in S701, the statistical shape models of the femur and tibia are iterated using the RL planar profile as a reference to obtain updated statistical shape models, including:

[0051] S801, the statistical shape models of the femur and tibia are projected onto the RL plane using the Z-buffer algorithm to obtain the RL projection contours, and a one-to-one 2D point pair correspondence is established between the RL projection contours and the X-ray bone boundary contours on the RL plane.

[0052] S802, based on the position information of the points on the projected contour in three-dimensional space, assign their third-dimensional coordinates to the corresponding X-ray skeleton boundary contour points, and back-map the X-ray skeleton boundary contour to 3D space to obtain a one-to-one 3D point pair set correspondence between the statistical shape model 3D surface points and the skeleton contour points in 3D space.

[0053] S803, using the obtained 3D spatial skeleton contour points as the target, performs orthogonal Prouk analysis on the 3D point pair set to obtain the rotation matrix transformation from the initial statistical shape model to the target statistical shape model, and calculates the Prouk distance;

[0054] S804, determine if the Prouk distance is less than the threshold. If it is less, go to S807 to exit the iteration; otherwise, go to S805 to continue the iteration.

[0055] S805, apply rotation matrix transformation to the shape difference vector and the average statistical shape model X_avg respectively, and project the transformed average statistical shape model X_avg onto the two-dimensional RL plane to obtain the one-to-one 2D point pair correspondence between the projected contour and the X-ray skeleton boundary contour. Use the formula SSM=X_avg+Pb to calculate the information of the shape difference coefficient b required to update the statistical shape model.

[0056] S806: After calculating the updated statistical shape model data, proceed to S801 to continue iterating until the threshold is met and exit.

[0057] S807 outputs a statistical shape model generated by iterative RL planar contour processing.

[0058] Furthermore, as described in S704, the statistical shape models of the femur and tibia are iterated using the AP plane contour as a reference to obtain an updated statistical shape model, and the projection plane becomes the AP plane.

[0059] In a second aspect, the present invention provides a knee replacement surgery planning device based on X-ray imaging, the device comprising:

[0060] The data acquisition module 901 is used to acquire two orthogonal X-ray images of the knee joint, including an anteroposterior and lateral views;

[0061] The 3D reconstruction module 902 is used to reconstruct the 3D models of the femur and tibia of the knee joint from two input X-ray images using a statistical shape model method, and output the key point positions of each.

[0062] The prosthesis recommendation module 903 is used to recommend prosthesis models and poses based on the three-dimensional models of the femur and tibia at the knee joint and the location of key points.

[0063] Output module 904 is used to construct spatial relationships using key points, which are then confirmed by the doctor to confirm the examination results. The doctor can further adjust the prosthesis position as needed and finally output a knee joint prosthesis replacement surgery plan.

[0064] Thirdly, the present invention provides a knee replacement surgery planning device based on X-ray imaging, the device comprising: a processor, a memory, and computer program instructions stored in the memory and executable on the processor, wherein the processor is configured to execute the computer program instructions stored in the memory to implement the aforementioned knee replacement surgery planning method based on X-ray imaging.

[0065] Fourthly, the present invention also provides a knee replacement surgery planning storage medium based on X-ray images, wherein the computer storage medium stores computer program instructions, and the computer program instructions, when executed by a processor, implement the above-described knee replacement surgery planning method based on X-ray images.

[0066] Compared with the prior art, the present invention has the following technical advantages:

[0067] 1. The required bone model and corresponding key points can be reconstructed from two-dimensional X-ray images in three dimensions, thereby formulating an effective prosthesis replacement plan. Compared with three-dimensional prosthesis planning based on CT images, the method of the present invention has lower cost and can cover more people. Compared with two-dimensional prosthesis planning based on X-ray images, the method of the present invention can provide three-dimensional information, which is conducive to doctors to observe the prosthesis placement more intuitively.

[0068] 2. The method described in this invention uses a three-dimensional statistical shape model to reconstruct a three-dimensional skeleton model from a two-dimensional X-ray film. Furthermore, by selecting key points on the pre-generated statistical shape model, it is easy to obtain the key points on the registered three-dimensional shape model, thus solving the problem of simultaneously outputting the three-dimensional model and key points.

[0069] 3. Existing technical solutions generate statistical shape models of bones based on complete bones. The method described in this invention trains and generates corresponding statistical shape models for truncated femur or tibia models, solving the problem of establishing statistical shape models for bone targets with different truncated lengths and different scales. Attached Figure Description

[0070] The features, advantages and technical effects of exemplary embodiments of the present invention will now be described with reference to the accompanying drawings.

[0071] Figure 1 This is a schematic diagram of the knee replacement surgery planning method based on X-ray imaging provided in an embodiment of the present invention;

[0072] Figure 2 This is a schematic diagram of the process for reconstructing a three-dimensional model of the femur and tibia of the knee joint according to an embodiment of the present invention;

[0073] Figure 3 This is a schematic diagram of the process for pre-training and generating statistical shape models of the femur and tibia provided in an embodiment of the present invention;

[0074] Figure 4 This is a schematic diagram of the process for generating shape vector sample sets of the femur and tibia of the knee joint provided in an embodiment of the present invention;

[0075] Figure 5 This is a schematic diagram of the process for aligning shape vectors in a shape vector sample set according to an embodiment of the present invention;

[0076] Figure 6 This is a schematic flowchart of obtaining parameterized statistical shape models of the femur and tibia of the knee joint provided in an embodiment of the present invention;

[0077] Figure 7 This is a schematic diagram of the process for iteratively updating the statistical shape model parameters of the femur and tibia using the AP plane contour and RL plane contour, as provided in an embodiment of the present invention.

[0078] Figure 8 This is a schematic diagram of the process of iterating the statistical shape model of the femur and tibia using the RL planar contour as a reference, provided in an embodiment of the present invention.

[0079] Figure 9 This is a schematic diagram of a knee replacement surgery planning device based on X-ray imaging provided in an embodiment of the present invention;

[0080] Figure 10 This is a schematic diagram of the structure of a computing device provided in an embodiment of the present invention. Detailed Implementation

[0081] The features and exemplary embodiments of various aspects of this disclosure will now be described in detail. To make the objectives, technical solutions, and advantages of this disclosure clearer, the disclosure will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are intended to explain this disclosure only and not to limit it. For those skilled in the art, this disclosure can be implemented without some of these specific details. The following description of the embodiments is merely to provide a better understanding of this disclosure by illustrating examples.

[0082] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes said element.

[0083] To better understand the present invention, the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0084] Figure 1 This is a schematic diagram of the knee replacement surgery planning method based on X-ray imaging provided in an embodiment of the present invention.

[0085] like Figure 1 As shown, this invention provides a method for planning knee replacement surgery based on X-ray imaging, including the following steps:

[0086] S101, acquire two orthogonal X-ray images of the knee joint, including an anteroposterior and lateral views;

[0087] S102, using the statistical shape model method, three-dimensional reconstruction of the femur and tibia of the knee joint is performed on the two input X-ray images, and the key point positions of each are output.

[0088] S103 recommends prosthesis type and position based on the three-dimensional model of the femur and tibia at the knee joint and the location of key points;

[0089] S104 utilizes the spatial relationships constructed from key points, allowing doctors to confirm examination results and further adjust the prosthesis position as needed, ultimately outputting a knee joint prosthesis replacement surgery plan.

[0090] In S101, the knee X-ray anteroposterior view refers to the anterior-posterior lateral view (AP) of the knee, and the lateral view refers to the left and right lateral views (RL) of the knee.

[0091] like Figure 2 As shown in S102, the statistical shape model method is used to reconstruct three-dimensional models of the femur and tibia of the knee joint from two input X-ray images, and outputs the key point positions of each, including:

[0092] S201, pre-trained statistical shape models of the femur and tibia, and labeled the key point locations respectively. The statistical shape model includes the average shape and deformation parameters.

[0093] S202: Obtain the contours of two X-ray images, align the X-ray images according to the contour information, and obtain the initial position and scale of the femur and tibia 3D models;

[0094] S203, Load the pre-generated statistical shape model of the femur and tibia, and apply the scale and position of the initial 3D model of the femur and tibia to its average shape model as the initial statistical shape model of the femur and tibia;

[0095] S204, using AP plane contour and RL plane contour to iterate, update the parameters of the femoral-tibial statistical shape model, and generate the final femoral-tibial statistical shape model and key point positions;

[0096] S205: Obtain shape point cloud based on femoral and tibia statistical shape model, calculate the normal vector of point cloud, and use Poisson surface reconstruction method to generate surface model from point cloud to realize visualization of femoral and tibia three-dimensional model.

[0097] like Figure 3 As shown in S201, the pre-trained statistical shape models of the femur and tibia are generated, and the key point locations are labeled respectively, including:

[0098] S301, collect knee joint CT image samples from different patients, and generate a set of shape vector samples of the femur and tibia of the knee joint;

[0099] S302, Align the shape vectors in the shape vector sample set;

[0100] S303, process the aligned sample shape vectors to obtain parameterized statistical shape models of the femur and tibia of the knee joint;

[0101] S304: Select key points on the statistical shape model of the femur and tibia and obtain their coordinate positions.

[0102] like Figure 4 As shown in S301, collecting knee CT image samples from different patients and generating a set of shape vector samples of the femur and tibia of the knee joint includes:

[0103] S401, collect knee joint CT image samples from different patients;

[0104] S402, after segmentation, obtain segmentation model samples of the femur and tibia at the knee joint;

[0105] S403, after the segmentation model samples are centered and scaled, a normalized shape vector is obtained;

[0106] S404, apply the same scale coefficient to the normalized shape vectors simultaneously to obtain shape vectors with uniform scale;

[0107] S405, Initial lower alignment of femoral model samples in three-dimensional space, and initial upper alignment of tibial model samples in three-dimensional space;

[0108] S406, extract the discrete coordinate points of the surface of the bone segmentation model as the initial shape vector representation to obtain the shape vector sample set of the femur and tibia of the knee joint.

[0109] like Figure 5 As shown, the alignment of shape vectors in the shape vector sample set described in S302 includes:

[0110] S501, the sample with the fewest number of sample points in the shape vector sample of the femur and tibia is selected as the reference sample;

[0111] S502, determine whether the other samples in the femoral and tibia shape vector sample sets, excluding the reference sample, are aligned with the reference sample. If all samples are aligned, proceed to step S507; if there are misaligned samples, proceed to step S503 for iteration.

[0112] S503: Select one sample from the femoral and tibial shape vector sample set other than the reference sample, and use the nearest point principle to determine the one-to-one correspondence between the selected sample and the midpoint of the reference sample.

[0113] S504, use the orthogonal Pluke analysis method to calculate the rotation transformation matrix between point pairs, apply the same rotation transformation to the sample shape vector, and calculate the Pluke distance between point pairs;

[0114] S505, determine whether the Prouk distance is less than the threshold. If it is less, take the points belonging to the sample in the point pair as the shape vector of the sample, and continue to S502 to determine whether there are unaligned samples in the sample set at this time. If it exceeds the set threshold, go to S503 to re-acquire the point pair set of the transformed sample and the reference sample and continue iterating until the threshold of the selected sample meets the requirements.

[0115] S506, perform the same processing on all samples in the femoral and tibial shape vector sample sets to align them with the reference sample;

[0116] S507, obtain the shape vectors of all samples after final alignment, the aligned sample shape vectors have the same number of coordinate points.

[0117] like Figure 6 As shown, the processed and aligned sample shape vectors described in S303 yield parameterized statistical shape models of the femur and tibia of the knee joint, including:

[0118] S601, sample N points using a uniform distribution on the aligned sample shape vector;

[0119] S602, Calculate the average shape vector X_avg based on the latest sample shape vector.

[0120] S603, the sample shape vector is obtained by subtracting the average shape vector from the sample shape vector;

[0121] S604. Using the singular value decomposition method, the deviation shape vector analysis of the sample is used to obtain the corresponding eigenvector matrix P and deformation coefficient b, resulting in the parameterized statistical shape model SSM=X_avg+Pb for the femur and tibia of the knee joint.

[0122] like Figure 7 As shown in S204, the parameters of the femoral-tibial statistical shape model are updated by iterating using the AP plane contour and RL plane contour to generate the final femoral-tibial statistical shape model and key point locations, including:

[0123] S701 uses the RL planar profile as a reference to iterate the statistical shape model of the femur and tibia to obtain an updated statistical shape model;

[0124] S702, Project the updated statistical shape model onto the AP plane and calculate the Euclidean distance between the projected profile and the AP skeletal profile;

[0125] S703, determine if the Euclidean distance is less than the threshold. If it is less, go to S707 to output the final statistical shape model of the femur and tibia; if it is greater, go to S704 to continue the iteration.

[0126] S704 uses the AP plane profile as a reference to iterate the statistical shape model of the femur and tibia to obtain an updated statistical shape model;

[0127] S705: Project the updated statistical shape model onto the RL plane and calculate the Euclidean distance between the projected profile and the RL skeletal profile.

[0128] S706: Determine if the Euclidean distance is less than the threshold. If it is less, proceed to S707 to output the final statistical shape model of the femur and tibia. If it is greater, proceed to S701 to continue iterating until the iteration requirement and exit condition are met.

[0129] S707 generates the final statistical shape model of the femur and tibia.

[0130] like Figure 8 As shown in S701, the statistical shape model of the femur and tibia is iterated using the RL planar profile as a reference to obtain an updated statistical shape model, including:

[0131] S801, the statistical shape models of the femur and tibia are projected onto the RL plane using the Z-Buffer algorithm to obtain the RL projection contours, and a one-to-one 2D point pair set correspondence is established between the RL projection contours and the X-ray bone boundary contours on the RL plane.

[0132] S802, based on the position information of the points on the projected contour in three-dimensional space, assign their third-dimensional coordinates to the corresponding X-ray skeleton boundary contour points, and back-map the X-ray skeleton boundary contour to 3D space to obtain a one-to-one 3D point pair set correspondence between the statistical shape model 3D surface points and the skeleton contour points in 3D space.

[0133] S803, using the obtained 3D spatial skeleton contour points as the target, performs orthogonal Prouk analysis on the 3D point pair set to obtain the rotation matrix transformation from the initial statistical shape model to the target statistical shape model, and calculates the Prouk distance;

[0134] S804, determine if the Prouk distance is less than the threshold. If it is less, go to S807 to exit the iteration; otherwise, go to S805 to continue the iteration.

[0135] S805, apply rotation matrix transformation to the shape difference vector and the average statistical shape model X_avg respectively, and project the transformed average statistical shape model X_avg onto the two-dimensional RL plane to obtain the one-to-one 2D point pair correspondence between the projected contour and the X-ray skeleton boundary contour. Use the formula SSM=X_avg+Pb to calculate the information of the shape difference coefficient b required to update the statistical shape model.

[0136] S806: After calculating the updated statistical shape model data, proceed to S801 to continue iterating until the threshold is met and exit.

[0137] S807 outputs a statistical shape model generated by iterative RL planar contour processing.

[0138] Furthermore, as described in S704, the statistical shape models of the femur and tibia are iterated using the AP plane contour as a reference to obtain an updated statistical shape model, and the projection plane becomes the AP plane.

[0139] Figure 9 This is a schematic diagram of a knee replacement surgery planning device based on X-ray imaging provided in an embodiment of the present invention.

[0140] like Figure 9 As shown, the present invention provides a knee replacement surgery planning device based on X-ray imaging, the device comprising:

[0141] The data acquisition module 901 is used to acquire two orthogonal X-ray images of the knee joint, including an anteroposterior and lateral views;

[0142] The 3D reconstruction module 902 is used to reconstruct the 3D models of the femur and tibia of the knee joint from two input X-ray images using a statistical shape model method, and output the key point positions of each.

[0143] The prosthesis recommendation module 903 is used to recommend prosthesis models and poses based on the three-dimensional models of the femur and tibia at the knee joint and the location of key points.

[0144] Output module 904 is used to construct spatial relationships using key points, which are then confirmed by the doctor to confirm the examination results. The doctor can further adjust the prosthesis position as needed and finally output a knee joint prosthesis replacement surgery plan.

[0145] Figure 9 Each module / unit in the illustrated device has the ability to implement Figure 1 The functions of each step in the process and their corresponding technical effects are described in detail here for the sake of brevity.

[0146] like Figure 10As shown, the present invention also provides a knee replacement surgery planning device based on X-ray imaging. The device includes: a processor 1001, a memory 1002, and computer program instructions stored in the memory 1002 and executable on the processor 1001. The processor 1001 is used to execute the computer program instructions stored in the memory 1002 to implement the above-described knee replacement surgery planning method based on X-ray imaging.

[0147] Specifically, the processor 1001 may include a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the present invention.

[0148] Memory 1002 may include mass storage for data or instructions. For example, and not as a limitation, memory may include hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these devices.

[0149] In one instance, the memory 1002 may include removable or non-removable (or fixed) media, or the memory may be non-volatile solid-state memory. The memory may be internal or external to the integrated gateway disaster recovery device.

[0150] In one instance, memory 1002 may be read-only memory (ROM). In one instance, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically rewritable ROM (EAROM), or flash memory, or a combination of two or more of these.

[0151] In one example, memory 1002 may include read-only memory (ROM), random access memory (RAM), disk storage media device, optical storage media device, flash memory device, electrical, optical, or other physical / tangible memory storage device. Thus, typically, memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method according to one aspect of this disclosure.

[0152] The processor 1001 reads and executes computer program instructions stored in the memory 1002 to achieve... Figure 1 The methods / steps in the illustrated embodiments, and the corresponding technical effects they achieve, will not be elaborated upon here for the sake of brevity.

[0153] In one embodiment, the computing device may further include a communication interface 1003 and a bus 1004. For example... Figure 10 As shown, the processor 1001, memory 1002, and communication interface 1003 are connected through bus 1004 and complete communication with each other.

[0154] The communication interface 1003 is mainly used to realize communication between various modules, devices, units and / or equipment in this invention.

[0155] Bus 1004 includes hardware, software, or both, that couples components of an online data traffic metering device together. For example, and not as a limitation, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, the bus may include one or more buses. Although specific buses are described and illustrated in this invention, this disclosure contemplates any suitable bus or interconnect.

[0156] In addition, in conjunction with the knee replacement surgery planning method based on X-ray images described in the above embodiments, the present invention also provides a computer storage medium for implementation. The computer storage medium stores computer program instructions, which, when executed by a processor, implement the aforementioned knee replacement surgery planning method based on X-ray images.

[0157] The computer storage medium provided in this invention can be any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. For example, a computer-readable storage medium can be, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this invention, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0158] The functional blocks shown in the above-described structural diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this disclosure are programs or code segments used to perform the required tasks. Those skilled in the art can write computer program code for performing the operations of this invention in one or more programming languages ​​or combinations thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages—such as the "C" language or similar programming languages. Furthermore, the program or code segment can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. A machine-readable medium can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc.

[0159] The aspects of this disclosure have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by special-purpose hardware performing the specified functions or actions, or can be implemented by a combination of special-purpose hardware and computer instructions.

[0160] The above description is merely a specific embodiment of this disclosure. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of this disclosure is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this disclosure, and these modifications or substitutions should all be covered within the protection scope of this disclosure.

Claims

1. A method for knee replacement surgery planning based on X-ray images, characterized in that, The method comprises the following steps: S101, acquiring two orthogonal X-ray images of a knee joint, including an anteroposterior view and a lateral view; S102, using a statistical shape model to three-dimensionally reconstruct a femur and a tibia of the knee joint from the two input X-ray images, and outputting respective key point positions; S103, recommending a prosthesis model and a pose according to the three-dimensional models of the femur and the tibia and the key point positions at the knee joint; S104, confirming a test result by a doctor according to a spatial relationship constructed by the key points, and further adjusting the pose of the prosthesis as needed, and finally outputting a knee joint prosthesis replacement surgery planning scheme; S102, using a statistical shape model to three-dimensionally reconstruct a femur and a tibia of the knee joint from the two input X-ray images, and outputting respective key point positions, comprising: S201, pre-training a statistical shape model of the femur and the tibia to generate, and respectively labeling key point positions, wherein the statistical shape model comprises an average shape and a deformation parameter; S202, acquiring contours of the two X-ray images, aligning the X-ray images according to the contour information, and obtaining an initial position and a scale of the femur and the tibia 3D model; S203, loading a pre-generated statistical shape model of the femur and the tibia, and applying the scale and the position of the initial 3D model of the femur and the tibia to the average shape model of the statistical shape model as an initial statistical shape model of the femur and the tibia; S204, using the anteroposterior view plane contour and the lateral view plane contour to iteratively update the parameters of the statistical shape model of the femur and the tibia, and generating a final statistical shape model of the femur and the tibia and key point positions; S205, obtaining a shape point cloud according to the statistical shape model of the femur and the tibia, calculating a normal vector of the point cloud, using a Poisson surface reconstruction method to generate a surface model from the point cloud, and realizing visualization of the three-dimensional model of the femur and the tibia; S204, using the anteroposterior view plane contour and the lateral view plane contour to iteratively update the parameters of the statistical shape model of the femur and the tibia, and generating a final statistical shape model of the femur and the tibia and key point positions, comprising: S701, using the lateral view plane contour as a reference to iteratively update the statistical shape model of the femur and the tibia, and obtaining an updated statistical shape model; S702, projecting the updated statistical shape model to the anteroposterior view plane, and calculating an Euclidean distance between the projected contour and a bone contour of the anteroposterior view; S703, judging whether the Euclidean distance is less than a threshold value, if yes, proceeding to S707 to output a final statistical shape model of the femur and the tibia, and if no, proceeding to S704 to continue iteration; S704, using the anteroposterior view plane contour as a reference to iteratively update the statistical shape model of the femur and the tibia, and obtaining an updated statistical shape model; S705, projecting the updated statistical shape model to the lateral view plane, and calculating an Euclidean distance between the projected contour and a bone contour of the lateral view; S706, judging whether the Euclidean distance is less than a threshold value, if yes, proceeding to S707 to output a final statistical shape model of the femur and the tibia, and if no, proceeding to S701 to continue iteration until a required iteration exit condition is reached; S707, generating a final femur and tibia statistical shape model.

2. The method for knee replacement surgery planning based on X-ray images according to claim 1, characterized in that, The statistical shape models of the femur and the tibia are pre-trained and the positions of the key points are labeled in S201, and the statistical shape models of the femur and the tibia are pre-trained and the positions of the key points are labeled. S301, collecting knee CT image samples of different patients to generate shape vector sample sets of the femur and the tibia of the knee; S302, aligning the shape vectors in the shape vector sample sets; S303, processing the aligned sample shape vectors to obtain the parameterized statistical shape models of the femur and the tibia of the knee; S304, selecting key points on the statistical shape models of the femur and the tibia and obtaining the coordinate positions of the key points.

3. The method of planning a knee replacement surgery based on X-ray images according to claim 2, characterized in that, The knee CT image samples of different patients are collected to generate shape vector sample sets of the femur and the tibia of the knee in S301, and the knee CT image samples of different patients are collected. S401, collecting knee CT image samples of different patients; S402, obtaining segmented model samples of the femur and the tibia at the knee after segmentation; S403, obtaining normalized shape vectors after the segmented model samples are decentered and descaled; S404, applying the same scale proportion coefficient to the normalized shape vectors to obtain shape vectors with uniform scales; S405, aligning the femur model samples at the initial lower part in the three-dimensional space and aligning the tibia model samples at the initial upper part in the three-dimensional space; S406, extracting surface discrete coordinate points of the segmented bone model as initial shape vectors to obtain shape vector sample sets of the femur and the tibia of the knee.

4. The method of planning a knee replacement surgery based on X-ray images according to claim 2, wherein, The shape vectors in the shape vector sample sets are aligned in S302, and the shape vectors in the shape vector sample sets are aligned. S501, selecting the sample with the least number of points in the femur shape vector sample set and the tibia shape vector sample set as the reference sample; S502, determining whether the samples other than the reference sample in the femur shape vector sample set and the tibia shape vector sample set are aligned with the reference sample, if all the samples are aligned, proceeding to step S507, if there are unaligned samples, proceeding to step S503 for iteration; S503, selecting a sample other than the reference sample in the femur shape vector sample set and the tibia shape vector sample set, and determining the one-to-one correspondence between the selected sample and the reference sample using the iterative closest point algorithm; S504, calculating the rotation transformation matrix between the point pairs using the orthogonal Procrustes analysis method, applying the same rotation transformation to the sample shape vectors, and calculating the Procrustes distance between the point pairs; S505, determining whether the Procrustes distance is less than a threshold value, if it is less than the threshold value, taking the points belonging to the sample in the point pair as the shape vector of the sample, and continuing to determine whether there are unaligned samples in the sample set in S502, if it exceeds the set threshold value, reacquiring the point pair set of the sample after the transformation and the reference sample, and proceeding to S504 for iteration until the threshold value of the selected sample meets the requirements; S506, processing all the samples in the femur shape vector sample set and the tibia shape vector sample set to align them with the reference sample; S507, obtaining the shape vectors of all the samples after alignment, and the aligned sample shape vectors have the same number of coordinate points.

5. The method of planning a knee replacement surgery based on X-ray images according to claim 2, wherein, The aligned sample shape vectors are processed to obtain the parameterized statistical shape models of the femur and the tibia of the knee in S303, and the aligned sample shape vectors are processed to obtain the parameterized statistical shape models of the femur and the tibia of the knee. S601, using uniform distribution sampling N points on the aligned sample shape vector; S602, calculating the average statistical shape model X_avg according to the latest sample shape vector, S603, obtaining the shape difference vector of the sample by subtracting the average statistical shape model from the sample shape vector; S604, using the singular value decomposition method to analyze the shape difference vector of the sample to obtain the corresponding shape difference vector eigenvalue P and shape difference vector eigenvector b, and obtaining the parameterized statistical shape model SSM of the knee femur and tibia through the formula SSM=X_avg+Pb.

6. The method of planning a knee replacement surgery based on X-ray images according to claim 1, wherein, The iteration of the statistical shape model of the femur and tibia is performed using the contour of the left and right view planes of the knee as the reference in S701, and an updated statistical shape model is obtained, comprising: S801, respectively projecting the statistical shape model of the femur and tibia onto the left and right view planes of the knee through the Z-Buffer algorithm to obtain the left and right view projection contours of the knee, and establishing a one-to-one 2D point pair set correspondence between the left and right view projection contours of the knee and the X-ray skeletal boundary contours on the left and right view planes of the knee; S802, according to the position information of the points on the projection contour in the three-dimensional space, assigning the third-dimensional coordinate to the corresponding X-ray skeletal boundary contour point, and mapping the X-ray skeletal boundary contour to the 3D space to obtain a one-to-one 3D point pair set correspondence between the 3D surface points of the statistical shape model and the skeletal contour points in the 3D space; S803, taking the obtained 3D skeletal contour points as the target, performing orthogonal Procrustes analysis on the 3D point pair set to obtain the rotation matrix transformation from the initial statistical shape model to the target statistical shape model, and calculating the Procrustes distance; S804, judging whether the Procrustes distance is less than a threshold value, if yes, turning to S807 to exit the iteration; otherwise, turning to S805 to continue the iteration; S805, applying the rotation matrix transformation to the shape difference vector and the average statistical shape model X_avg respectively, projecting the transformed average statistical shape model X_avg onto the two-dimensional left and right view planes of the knee again to obtain the one-to-one 2D point pair set correspondence between the projection contour and the X-ray skeletal boundary contour, and calculating the shape difference scale information required for updating the statistical shape model, i.e. the information of the shape difference vector eigenvalue P, by using the formula SSM=X_avg+Pb; S806, obtaining the updated statistical shape model data by calculation, and turning to S801 to continue the iteration until the threshold value meets the requirement to exit; S807, outputting the statistical shape model finally generated by iteration using the contour of the left and right view planes of the knee.

7. The method of planning a knee replacement surgery based on X-ray images according to claim 1, wherein, The iteration of the statistical shape model of the femur and tibia is performed using the contour of the left and right view planes of the knee as the reference in S701, and an updated statistical shape model is obtained, comprising:

8. An X-ray image-based knee replacement surgery planning apparatus, characterized by, A data acquisition module is configured to acquire two orthogonal X-ray images of a knee joint, including an anteroposterior image and a lateral image; A three-dimensional reconstruction module is configured to use a statistical shape model to three-dimensionally reconstruct a three-dimensional model of a femur and a tibia of a knee joint from the input two X-ray images, and output respective key point positions. ​ The prosthesis recommendation module is configured to recommend a prosthesis model and a pose according to the three-dimensional model of the femur and the tibia at the knee joint and the key point positions; The output module is configured to confirm the inspection result by the doctor by using the spatial relationship constructed by the key points, and further adjust the prosthesis pose as needed, and finally output a knee joint prosthesis replacement surgery planning scheme; The three-dimensional reconstruction module is configured to use a statistical shape model method to three-dimensionally reconstruct the input two X-ray images into three-dimensional models of the femur and the tibia of the knee joint, and output respective key point positions, including: S201, a statistical shape model of the femur and the tibia is pre-trained and generated, and key point positions are labeled respectively, wherein the statistical shape model includes an average shape and deformation parameters; S202, contours of two X-ray images are obtained, the X-ray images are aligned according to the contour information, and initial positions and scales of the femur and the tibia 3D models are obtained; S203, the pre-generated statistical shape model of the femur and the tibia is loaded, and the scale and the position of the initial 3D model of the femur and the tibia are applied to the average shape model as an initial statistical shape model of the femur and the tibia; S204, the statistical shape model of the femur and the tibia is updated by iteration using the knee front and rear view plane contours and the knee left and right view plane contours, and a final statistical shape model of the femur and the tibia and key point positions are generated; S205, shape point clouds are obtained according to the statistical shape model of the femur and the tibia, normal vectors of the point clouds are calculated, and a surface model is generated from the point clouds by using a Poisson surface reconstruction method, so as to realize visualization of the three-dimensional model of the femur and the tibia; S204, the statistical shape model of the femur and the tibia is updated by iteration using the knee front and rear view plane contours and the knee left and right view plane contours, and a final statistical shape model of the femur and the tibia and key point positions are generated, including: S701, the statistical shape model of the femur and the tibia is iterated using the knee left and right view plane contours as a reference to obtain an updated statistical shape model; S702, the updated statistical shape model is projected to the knee front and rear view plane, and the Euclidean distance between the projected contour and the knee front and rear view bone contour is calculated; S703, whether the Euclidean distance is less than a threshold value is judged, if yes, the final statistical shape model of the femur and the tibia is output in S707, if no, the iteration is continued in S704; S704, the statistical shape model of the femur and the tibia is iterated using the knee front and rear view plane contours as a reference to obtain an updated statistical shape model; S705, the updated statistical shape model is projected to the knee left and right view plane, and the Euclidean distance between the projected contour and the knee left and right view bone contour is calculated; S706, whether the Euclidean distance is less than a threshold value is judged, if yes, the final statistical shape model of the femur and the tibia is output in S707, if no, the iteration is continued in S701 until the iteration requirement exit condition is reached; S707, a final statistical shape model of the femur and the tibia is generated.

9. A knee joint replacement surgery planning device based on X-ray images, characterized in that, The device comprises a processor, a memory, and computer program instructions stored in the memory and executable on the processor, wherein the processor is configured to execute the computer program instructions stored in the memory to implement the method for planning knee replacement surgery based on X-ray images according to any one of claims 1 to 7.

10. A computer storage medium, comprising computer program instructions. The computer storage medium stores computer program instructions, and the computer program instructions are executed by a processor to implement the method for planning knee replacement surgery based on X-ray images according to any one of claims 1 to 7.

Images