Osteotomy plane intelligent planning method and system for joint replacement and electronic equipment

Abstract

The invention provides an osteotomy plane intelligent planning method and system for joint replacement and electronic equipment, and the method comprises the steps: segmenting a three-dimensional image of a target replacement joint through a deep learning algorithm, and reconstructing a skeleton three-dimensional model of the target replacement joint; based on the skeleton three-dimensional model, anatomical features of a target replacement joint are extracted, soft tissue around the target replacement joint is analyzed through a biomechanical model, and mechanical response characteristics of the soft tissue are obtained; the anatomical features and the mechanical response features serve as input features and are input into a trained artificial intelligence decision model, a target prosthesis model matched with the target replacement joint is output, osteotomy plane parameters of the target prosthesis model under a robot coordinate system are calculated, and the osteotomy plane parameters of the target prosthesis model are calculated; the technical problems that in the prior art, ligament soft tissue characteristics are not considered, intellectualization is only reflected in image segmentation, operation planning depends on doctor experience and the like are solved, the dependence of operation planning on the doctor experience is reduced, and the positioning precision of the prosthesis is improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent planning technology for osteotomy planes in joint replacement, specifically to an intelligent planning method, system, and electronic device for osteotomy planes in joint replacement. Background Technology

[0002] Total knee replacement surgery and total hip replacement surgery are effective treatments for joint diseases. The key to surgical success lies in the precise determination of the osteotomy plane during preoperative planning, which directly affects the accuracy of prosthesis implantation, the restoration of lower limb alignment, and soft tissue balance. Currently, the planning of the osteotomy plane mainly relies on the following two types of techniques: The first category is the traditional method based on mechanical locators. This method relies on rigid instruments such as intramedullary / extramedullary locators, which are aligned with the anatomical axis of the bone to determine the osteotomy angle and position. However, the installation and adjustment of the locator in this method are highly dependent on the doctor's experience, making it difficult to quantify and standardize, resulting in large fluctuations in osteotomy accuracy. Moreover, the adjustment scale of the locator is discrete, making it impossible to achieve continuous and precise personalized adjustments based on the patient's unique anatomical shape and force line. This makes it difficult to meet individual patient needs and can easily lead to improper osteotomy or inaccurate force line positioning, thereby reducing the survival rate after prosthesis surgery. The second category is the computer-aided planning method based on medical imaging. This method reconstructs a three-dimensional model of the skeleton using imaging data such as CT scans. Doctors manually mark feature points and determine the lower limb force line in the software, and plan and select the prosthesis through virtual "trial fitting". Based on the characteristics of soft tissues such as the medial and lateral collateral ligaments, the prosthesis's varus / valgus and internal / external rotation angles are fine-tuned. In addition, based on the clinical requirement of symmetrical balance between the medial and lateral femoral and tibial spaces, six osteotomy planes of the femur and tibia are determined. This method essentially transfers the doctor's manual operations from the operating room to the computer screen. The planning process still relies heavily on the doctor's manual operations (such as marking, rotating, and translating the prosthesis), which is time-consuming and labor-intensive. It fails to achieve truly automated and intelligent planning, and the planning quality is still strongly correlated with the doctor's clinical experience, making it difficult to achieve standardization and repeatability.

[0003] Most existing methods remain limited to the analysis of skeletal geometry. While some systems allow manual setting of ligament tension, they lack quantitative and dynamic simulation analysis of soft tissue biomechanical properties, making it impossible to accurately predict the balance of soft tissues after osteotomy during the planning phase. In recent years, although some studies have attempted to introduce deep learning technology, it has mainly been applied to the automatic segmentation of medical images, failing to achieve end-to-end intelligent processing from image to osteotomy parameter decision-making, and further neglecting to investigate the interpretability of deep learning algorithms in planning decisions. Furthermore, the intelligent planning of osteotomy planes for soft tissues such as ligaments is not yet adequately addressed; existing methods only plan the osteotomy plane from the perspective of the knee joint or prosthesis geometry.

[0004] Therefore, there is a lack of intelligent planning methods and systems in the existing technology that can automate the entire process, integrate bone geometry and soft tissue biomechanical properties, and have interpretability, so as to achieve high-precision, personalized and clinically reliable joint replacement osteotomy planning.

[0005] Therefore, existing technologies still need further development. Summary of the Invention

[0006] The purpose of this invention is to overcome the above-mentioned technical deficiencies and provide an intelligent planning method and system for osteotomy planes in joint replacement, so as to solve the problem that there is a lack of an intelligent planning method and system in the prior art that can be fully automated, integrate bone geometry and soft tissue biomechanical properties, and has interpretability, so as to achieve high-precision, personalized and clinically reliable osteotomy plane planning for joint replacement.

[0007] To achieve the above-mentioned technical objectives, according to a first aspect of the present invention, the present invention provides an intelligent planning method for osteotomy planes in joint replacement, comprising: S100. Obtain the patient's orthopedic medical imaging data, use deep learning algorithms to segment the three-dimensional image of the target replacement joint, and reconstruct the three-dimensional skeletal model of the target replacement joint. S200. Based on the three-dimensional model of the skeleton, extract the anatomical features of the target replacement joint, and use a biomechanical model to analyze the soft tissue around the target replacement joint to obtain the mechanical response characteristics of the soft tissue. S300: The anatomical features and the mechanical response characteristics are used as input features and input into the trained artificial intelligence decision model, and the target prosthesis model that matches the target replacement joint is output. S400. Based on the target prosthesis model, calculate the osteotomy plane parameters of the target prosthesis model in the robot coordinate system.

[0008] Preferably, the method for segmenting the three-dimensional image of the target replacement joint using a deep learning algorithm and reconstructing the three-dimensional skeletal model of the target replacement joint includes: A deep learning algorithm based on the U-Net model is used to extract the three-dimensional image of the target replacement joint from the orthopedic medical imaging data, and to reconstruct the three-dimensional model of the bone of the target replacement joint in three dimensions.

[0009] Furthermore, the method for extracting the anatomical features of the target replacement joint includes: Based on the three-dimensional skeletal model of the target replacement joint, anatomical features of the target replacement joint are extracted using morphological analysis methods.

[0010] Furthermore, the biomechanical model is a fiber-reinforced hyperelastic model, and the method for analyzing the soft tissue surrounding the target replacement joint using the biomechanical model includes: The anisotropic mechanical behavior of the soft tissue surrounding the target replacement joint was simulated using a fiber-reinforced hyperelastic model, thereby obtaining the mechanical response characteristics of the soft tissue.

[0011] Preferably, the mechanical response characteristics of the soft tissue are represented by the strain energy function of the fiber-reinforced hyperelastic model, wherein the strain energy function is composed of the superposition of the matrix contribution and the fiber contribution.

[0012] Furthermore, the artificial intelligence decision-making model is a lightweight decision tree-neural network model based on knowledge distillation, which includes a full neural network as a teacher model and a decision tree model as a student model. The teacher model analyzes the input features, outputs a first feature, and then uses the first feature as the leaf node of the student model's decision tree to output the target prosthesis model that matches the target replacement joint.

[0013] Furthermore, the method for outputting the target prosthesis model that matches the target replacement joint includes: The full neural network performs feature analysis and knowledge extraction on the input anatomical features and mechanical response characteristics, and outputs a knowledge feature vector; The decision tree model takes the knowledge feature vector as input and outputs the target prosthesis model that matches the target replacement joint based on the decision rules.

[0014] Furthermore, the method for calculating the osteotomy plane parameters of the target prosthesis model in the robot coordinate system includes: Obtain the first osteotomy plane parameters of the target prosthesis model in the prosthesis coordinate system, transform the first osteotomy plane parameters in the prosthesis coordinate system to the robot coordinate system through coordinate transformation to obtain the second osteotomy plane parameters, and use the second osteotomy plane parameters as the geometric constraint conditions of the replacement prosthesis.

[0015] According to a second aspect of the present invention, an intelligent planning system for osteotomy planes in joint replacement is provided, comprising: Image segmentation module: used to acquire orthopedic medical imaging data of patients, segment the three-dimensional image of the target replacement joint using deep learning algorithms, and reconstruct the three-dimensional skeletal model of the target replacement joint; Feature analysis module: Based on the three-dimensional model of the skeleton, it is used to extract the anatomical features of the target replacement joint and to analyze the soft tissue around the target replacement joint using a biomechanical model to obtain the mechanical response characteristics of the soft tissue; Intelligent decision-making module: used to input the anatomical features and the mechanical response characteristics as input features into the trained artificial intelligence decision-making model, and output the target prosthesis model that matches the target replacement joint; Osteotomy Calculation Module: Used to calculate the osteotomy plane parameters of the target prosthesis model in the robot coordinate system based on the target prosthesis model.

[0016] According to a third aspect of the present invention, an electronic device is provided, comprising: a memory; and a processor, wherein the memory stores computer-readable instructions that, when executed by the processor, implement the above-described intelligent planning system method for osteotomy planes in joint replacement.

[0017] Beneficial effects: This invention provides a method and system for intelligent planning of osteotomy planes in joint replacement surgery. It employs deep learning algorithms to achieve intelligent and rapid segmentation of the patient's target replacement joint, utilizes morphological analysis to extract and analyze the anatomical features of the target replacement joint, combines biomechanical models to analyze the soft tissue mechanical response characteristics, and then uses a lightweight decision tree-neural network hybrid algorithm based on knowledge distillation to achieve intelligent prosthesis matching. Finally, it calculates the osteotomy plane parameters in the robot coordinate system based on the precisely located prosthesis's inner contour plane. This reduces the reliance on surgeon experience in surgical planning, effectively improves the prosthesis's positioning accuracy, and solves the technical problems of existing intelligent planning methods for osteotomy planes, such as not considering ligament and soft tissue characteristics, intelligence only reflected in image segmentation, and the reliance on surgeon experience in current surgical planning. Attached Figure Description

[0018] Figure 1 This is a flowchart of the intelligent planning system method for osteotomy plane in joint replacement provided in a specific embodiment of the present invention; Figure 2 This is a schematic diagram of the composition of the intelligent planning system for osteotomy plane in joint replacement provided in a specific embodiment of the present invention; Figure 3 This is a schematic diagram of the knowledge distillation structure of the artificial intelligence decision-making model provided in a specific embodiment of the present invention; Figure 4 This is a flowchart of the intelligent planning system method for osteotomy plane for knee replacement provided in a specific embodiment of the present invention; Figure 5 This is a schematic diagram of the osteotomy plane of the knee joint prosthesis provided in a specific embodiment of the present invention; The reference numerals in the above figures are as follows: 1. Superior oblique surface of the femur; 2. Anterior condylar surface of the femur; 3. Distal surface of the femur; 4. Posterior condylar surface of the femur; 5. Inferior oblique surface of the femur; 6. Tibial plateau surface. Detailed Implementation

[0019] To enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Other similar embodiments obtained by those skilled in the art based on the embodiments in this application without creative effort should all fall within the scope of protection of this application. Furthermore, directional terms mentioned in the following embodiments, such as "up," "down," "left," and "right," are only for reference to the directions in the accompanying drawings; therefore, the directional terms used are for illustrative purposes and not for limiting the invention.

[0020] The present invention will be further described below with reference to the accompanying drawings and preferred embodiments.

[0021] Example 1 like Figure 1 As shown, this invention provides an intelligent planning method for osteotomy planes in joint replacement, specifically including the following steps: S100. Acquire the patient's orthopedic medical imaging data, use deep learning algorithms to segment the three-dimensional image of the target replacement joint, and reconstruct the three-dimensional skeletal model of the target replacement joint.

[0022] Specifically, a deep learning algorithm based on the U-Net model is used to extract the 3D image of the target replacement joint from orthopedic medical imaging data, and to reconstruct the 3D model of the target replacement joint's skeleton. The U-Net model is a deep learning network with an encoder-decoder structure, which is particularly suitable for medical image segmentation tasks. By preserving image detail information through skip connections, it can accurately segment the boundaries of bone structures. Through its unique structural design, this model can effectively preserve and utilize the spatial information of the image during the segmentation process, thereby improving the accuracy and reliability of segmentation.

[0023] It should be noted that the U-Net model is a commonly used medical image segmentation algorithm. Taking knee replacement as an example, the training data requires standard features such as the anatomical structures of the femur, tibia, joint space, and ankle. By using relevant image data from historical cases, key feature points such as the anatomical structures of the femur, tibia, joint space, and ankle, as well as the corresponding prosthesis feature points, are labeled. These feature points are used to train the deep learning model algorithm, thereby training the model parameters and obtaining a dedicated model for feature extraction for orthopedic surgical planning. Then, the 3D model is reconstructed using the VTK surface reconstruction filter function built into VTK.

[0024] S200: Based on the 3D model of the skeleton, the anatomical features of the target replacement joint are extracted, and the soft tissues around the target replacement joint are analyzed using a biomechanical model to obtain the mechanical response characteristics of the soft tissues.

[0025] In this step, based on the three-dimensional skeletal model of the target replacement joint, morphological analysis methods are used to extract the anatomical features of the target replacement joint. In this embodiment, the anatomical features include geometric dimensions, topological features, spatial posture features, and region segmentation features. These features should comprehensively reflect the geometric shape, spatial relationships, and biomechanical characteristics of the target replacement joint (such as the hip, knee, and shoulder). Specifically, the morphological analysis methods include bone surface curvature analysis, key anatomical landmark identification, edge detection, and joint axis localization techniques, which can accurately obtain the geometric morphological features of the joint. These techniques, through detailed analysis of the bone surface and internal structure, ensure a comprehensive understanding of the joint's anatomical features.

[0026] Furthermore, in this embodiment, a fiber-reinforced hyperelastic model is used as a biomechanical model to simulate the anisotropic mechanical behavior of the soft tissues surrounding the target replacement joint, thereby obtaining the mechanical response characteristics of the soft tissues. The fiber-reinforced hyperelastic model can accurately simulate the differences in mechanical response of soft tissues such as ligaments and tendons in different directions. The mechanical response characteristics of the soft tissues are represented by the strain energy function of the fiber-reinforced hyperelastic model. The strain energy function is composed of the superposition of the matrix contribution and the fiber contribution. The matrix contribution describes the isotropic basic characteristics of the soft tissues, while the fiber contribution describes the anisotropic characteristics caused by fiber arrangement. The combination of the two can comprehensively characterize the mechanical behavior of the soft tissues during joint movement, specifically as follows: ; ; ; in, This represents the strain energy function of soft tissue, with the matrix contributing a portion of the strain energy. and fiber contribution portion It consists of two parts, among which Energy is contributed only when the fiber is stretched. Indicates shear modulus, Indicates the first strain invariant. This indicates the fiber elongation ratio, where the fiber contribution portion... Elongation only in the fiber direction It is activated when the value is greater than 1. , For fiber stiffness parameters (human ligaments) Approximately 10~100MPa Approximately 1-10).

[0027] S300 takes anatomical features and mechanical response characteristics as input features, inputs them into a trained artificial intelligence decision-making model, and outputs the target prosthesis model that matches the target replacement joint.

[0028] like Figure 3 As shown, the artificial intelligence decision-making model is a lightweight decision tree-neural network model based on knowledge distillation. This AI decision-making model includes a full neural network as the teacher model and a decision tree model as the student model. The teacher model analyzes the input features, outputs a first feature, and then uses this first feature as the leaf node of the student model's decision tree, thereby outputting the target prosthesis model that matches the target replacement joint.

[0029] Specifically, the full neural network performs feature analysis and knowledge extraction on the input anatomical features and mechanical response characteristics, and outputs a knowledge feature vector; the decision tree model takes the knowledge feature vector as input and outputs the target prosthesis model that matches the target replacement joint based on decision rules.

[0030] Understandably, see Figure 3 The working principle of using a lightweight decision tree-neural network model based on knowledge distillation to match target replacement joint prostheses is as follows: (1) Multi-feature input and preprocessing: The collected multi-modal data, such as the anatomical features of the target replacement joint and the mechanical response characteristics of the surrounding soft tissues, are input. After feature extraction, normalization and feature selection, a set of standardized high-dimensional feature vectors are formed, and these high-dimensional feature vectors are used as the input of the subsequent model. (2) Teacher model reasoning: After performing feature analysis and knowledge extraction on the input anatomical features and mechanical response characteristics using a pre-trained teacher model, a knowledge feature vector is output; (3) Soft label generation and knowledge distillation: Construct a soft label generation function to convert the knowledge feature vector of the teacher model into a soft label with a temperature parameter. This soft label serves as a supervision signal to guide the learning process of the student model, enabling the student model to not only learn the correct classification but also to imitate the "thinking style" of the teacher model.

[0031] (4) Student model training: The student model takes soft labels as the target, combines the traditional cross-entropy loss function, optimizes its own parameters, realizes the knowledge transfer from neural network to decision tree, and takes into account both performance and interpretability; (5) Model output and interpretability enhancement: After training, the student model runs independently in the inference phase and outputs the target prosthesis model that matches the target replacement joint. This output can be used to generate a visualization report to help doctors understand the basis of the model decision and improve trust and clinical adoption rate.

[0032] S400. Based on the target prosthesis model, calculate the osteotomy plane parameters of the target prosthesis model in the robot coordinate system.

[0033] Furthermore, in this embodiment, the first osteotomy plane parameters of the target prosthesis model in the prosthesis coordinate system are obtained. These parameters are then transformed to the robot coordinate system via coordinate transformation to obtain the second osteotomy plane parameters. These second osteotomy plane parameters are used as geometric constraints for the replacement prosthesis. The coordinate transformation process includes the calculation of rotation matrices and translation vectors to ensure the precise positioning of the osteotomy plane in the robot's operating space. Through these steps, the intelligent osteotomy plane planning method of this invention can automatically select the most suitable prosthesis model based on the patient's individualized anatomical characteristics and soft tissue biomechanical properties, and accurately calculate the osteotomy plane parameters required for the robot to perform the osteotomy operation, further improving the accuracy and individualization of joint replacement surgery.

[0034] See Figure 4 The following example, focusing on total knee arthroplasty, illustrates the implementation principle of this invention: Step 1: Perform medical image segmentation of the femur and tibia of the entire knee joint.

[0035] A deep learning algorithm based on the U-Net model was used to extract images of the femur and tibia from CT images and to perform three-dimensional reconstruction of the femur and tibia models, laying the foundation for obtaining key anatomical parameters required for subsequent precise prosthesis positioning. Step 2: Analysis of key anatomical and soft tissue features.

[0036] Based on the segmentation results of step 1 above, morphological analysis methods (such as edge detection, surface fitting, etc.) are used to extract and analyze the key anatomical features of the patient's knee joint, such as the femoral / tibial condyle surface features, femoral head center, tibial plateau center, femoral-tibial space and deformity features, etc. Furthermore, the Gasser-Ogden-Holzapfel (GOH) model was used to analyze the stress-strain characteristics of soft tissues such as the medial and lateral collateral ligaments of the knee joint, simulating the mechanical analysis during the dynamic surgical process. This was used to fine-tune the prosthesis's internal and external rotation and varus / valgus angles according to the patient's individual characteristics, thereby making the prosthesis positioning more accurate. The GOH model is represented as follows: ; ; ; in, This represents the strain energy function of soft tissues such as the medial and lateral collateral ligaments of the knee joint, with the matrix contributing a portion of the strain energy. and fiber contribution portion It consists of two parts, among which Energy is contributed only when the fiber is stretched. Indicates shear modulus, Indicates the first strain invariant. This indicates the fiber elongation ratio, where the fiber contribution portion... Elongation only in the fiber direction It is activated when the value is greater than 1. , For fiber stiffness parameters (human ligaments) Approximately 10~100MPa Approximately 1-10).

[0037] Understandably, the above technical solution effectively improves the positioning accuracy of the prosthesis, preserves the patient's original bone to the greatest extent, accurately aligns the patient's lower limb force line, achieves soft tissue and interspace balance, facilitates the recovery of the patient's knee joint function, achieves the functional effect of "internal stability and external balance" after prosthesis replacement, and meets the patient's need for a high quality of life after surgery.

[0038] Step 3: Intelligent matching and decision analysis of prostheses.

[0039] Based on the properties of soft tissue, and combined with key anatomical feature points obtained from medical image segmentation results, the key anatomical features of the tibia and femur and the mechanical response characteristics of the surrounding soft tissues are used as the identification parameters of the lightweight decision tree-neural network algorithm of knowledge distillation to automatically and accurately locate the target prosthesis and output the target prosthesis model that matches the replacement knee joint. At the same time, the decision tree is used to visualize the specific decision rules and understand the contribution of each feature to the planning decision results, thereby achieving a balance between accuracy and interpretability, and thus achieving standardization of joint replacement surgery planning. The aforementioned decision rules refer to the logical combination of a series of feature judgment conditions corresponding to each path from the root node to the leaf node in the decision tree model, which ultimately determines an output category (such as prosthesis model).

[0040] While outputting the target prosthesis model, the decision tree visualizes the specific decision rules. Through the decision tree, doctors can clearly understand the contribution of each feature to the planning decision result, such as the weight of a certain anatomical feature in prosthesis selection. This approach achieves a balance between accuracy and interpretability, ensuring the accuracy of prosthesis selection while allowing doctors to understand the basis of the algorithm's decision-making. This, in turn, standardizes joint replacement surgery planning, improves the success rate of surgery, and enhances the patient's postoperative recovery.

[0041] Furthermore, the implementation principle of the lightweight decision tree-neural network hybrid model for knowledge distillation is as follows: Figure 3As shown, this utilizes the teacher model. Analyze the features, and then use the output as a decision tree for the student model. The leaf nodes then output the target prosthesis model that matches the target replacement joint; Hybrid Model and its loss function as follows: ; ; ; ; ; in, It is a teacher model logits, It is a temperature parameter used to control the smoothness of the probability distribution. The distribution is more uniform at times. When it is close to a hard label, It is the logits of the student model. ( (Number of categories), corresponding codes, These are neural network parameters. yes Regularization (reducing the number of parameters). It is the depth of the decision tree (used to control complexity). and It is the regularization coefficient. It is the output probability distribution of the teacher model. It is the output probability distribution of the student model.

[0042] Step 4: Accurate calculation of osteotomy plane parameters.

[0043] After determining the precise position of the prosthesis as described above, and combining the known osteotomy plane parameters (first osteotomy plane parameters) of each inner contour surface of the prosthesis in the prosthesis coordinate system, the osteotomy plane parameters in the image coordinate system are calculated and converted into osteotomy plane parameters (second osteotomy plane parameters) in the robot coordinate system. That is, for example... Figure 5 The five osteotomy planes of the femoral prosthesis (superior femoral oblique plane 1, anterior femoral condyle plane 2, distal femoral plane 3, posterior femoral condyle plane 4, and inferior femoral oblique plane 5) and one osteotomy plane of the tibial plateau prosthesis (tibial plateau plane 6) in the knee joint are transformed to the robot coordinate system to complete the intelligent planning of the final femoral and tibial osteotomy planes. The specific calculation method is as follows: ; in, i=1,2,3,4,5,6, which correspond to the superior oblique surface of the femur (1), the anterior condylar surface of the femur (2), the distal surface of the femur (3), the posterior condylar surface of the femur (4), the inferior oblique surface of the femur (5), and the tibial plateau surface (6), respectively). Let be the rotation matrix between the prosthetic coordinate system and the robot coordinate system. The normal vector of each osteotomy plane (superior oblique surface of the femur 1, anterior condyle surface of the femur 2, distal surface of the femur 3, posterior condyle surface of the femur 4, inferior oblique surface of the femur 5, and tibial plateau surface 6) in the prosthesis coordinate system; These are the distance parameters from the origin of the robot coordinate system to each osteotomy plane, located in the prosthesis coordinate system. Let be the translation vector from the origin of the spurious coordinate system to the origin of the robot coordinate system. Let represent the coordinate position of the point to be solved in the robot coordinate system, where . , All parameters are for the osteotomy plane.

[0044] In actual surgical procedures, the optimized osteotomy plane parameters are transmitted to the surgical robot. Based on these parameters, the surgical robot precisely controls the movement trajectory of the osteotomy tool, achieving precise osteotomy of the femur and tibia.

[0045] It should be further explained that this method can not only be used for total knee arthroplasty, but also provide surgical planning references for other procedures, such as total hip arthroplasty and unicompartmental arthroplasty. This scheme can directly obtain the osteotomy plane parameters in the robot coordinate system, which enhances the surgeon's understanding and credibility of the decision-making process. Compared with traditional methods and existing computer-aided planning methods, this scheme fully considers the impact of the patient's personalized soft tissue characteristics on prosthesis positioning, and realizes intelligent operation of the entire process from image segmentation to osteotomy plane parameter output. This reduces the surgeon's dependence on surgical planning, shortens the surgical planning time, improves the accuracy of prosthesis positioning, and helps to reduce the joint revision rate.

[0046] It should be noted that this embodiment provides an intelligent planning method for osteotomy planes in joint replacement. It employs a deep learning algorithm to achieve intelligent and rapid segmentation of the patient's target replacement joint, utilizes morphological analysis to extract and analyze the anatomical features of the target replacement joint, combines a biomechanical model to analyze the soft tissue mechanical response characteristics, and then uses a lightweight decision tree-neural network hybrid algorithm based on knowledge distillation to achieve intelligent prosthesis matching. Finally, it calculates the osteotomy plane parameters in the robot coordinate system based on the precisely located prosthesis inner contour plane. This reduces the reliance on surgeon experience in surgical planning, increases the transparency of the surgical planning decision-making process, effectively improves the positioning accuracy of the prosthesis, and solves the technical problems of existing intelligent planning methods for osteotomy planes, such as not considering ligament and soft tissue characteristics, intelligence only reflected in image segmentation, and existing surgical planning relying on surgeon experience.

[0047] Example 2 like Figure 2 As shown, the present invention provides an intelligent planning system for osteotomy planes in joint replacement surgery. The system includes an image segmentation module 100, a feature analysis module 200, an intelligent decision-making module 300, and an osteotomy calculation module 400. The modules work together to achieve precise planning of the osteotomy plane in joint replacement surgery.

[0048] Image segmentation module 100 is used to acquire orthopedic medical imaging data of patients, such as CT or MRI scans. This module utilizes deep learning algorithms to process the medical images, automatically segmenting the three-dimensional images of the target replacement joint (such as the femur and tibia in the knee joint). Through three-dimensional reconstruction technology, this module converts the segmented two-dimensional slice data into a complete three-dimensional bone model, providing a foundation for subsequent analysis. Image segmentation module 100 employs a U-Net deep neural network architecture, which has been trained on a large amount of clinical data and can accurately identify bone boundaries, further improving the accuracy of image segmentation.

[0049] The feature analysis module 200 extracts the anatomical features of the target replacement joint based on a three-dimensional skeletal model. These features include, but are not limited to, key parameters such as articular surface morphology, bone axis direction, and bone geometry. Simultaneously, this module utilizes a biomechanical model to analyze the mechanical response characteristics of the soft tissues surrounding the target replacement joint (such as ligaments, tendons, and joint capsules). The feature analysis module 200 employs a fiber-reinforced hyperelastic model to simulate the stress distribution and deformation of soft tissues under different load conditions, obtaining the mechanical response characteristics of the soft tissues. These mechanical response characteristics include parameters such as stiffness coefficient, elastic modulus, and maximum load-bearing capacity, providing important basis for prosthesis matching and osteotomy plane determination.

[0050] The intelligent decision-making module 300 takes anatomical features and mechanical response characteristics as input features and feeds them into the trained artificial intelligence decision-making model. For example... Figure 3 As shown, this AI decision-making model employs a knowledge distillation structure, comprising a full model (teacher model) and a lightweight model (student model). The teacher model is a complex deep neural network with powerful feature extraction and decision-making capabilities, while the student model is a lightweight decision tree-neural network hybrid model that learns key knowledge from the teacher model through knowledge distillation. The knowledge distillation process includes soft label generation and distillation loss function calculation, enabling the student model to maintain high accuracy while significantly reducing computational complexity. The intelligent decision module 300 ultimately outputs the target prosthesis model that best matches the target replacement joint, while also providing a matching score and interpretability analysis results to help doctors understand the basis of their decisions.

[0051] The osteotomy calculation module 400 calculates the osteotomy plane parameters of the target prosthesis in the robot coordinate system based on the prosthesis model. This module first establishes the coordinate transformation relationship between the patient's bone and the surgical robot, and then accurately calculates parameters such as the position, angle, and depth of the osteotomy plane according to the prosthesis installation requirements. The osteotomy calculation module 400 can deduce the optimal osteotomy plane from the ideal implantation position of the prosthesis, ensuring a perfect match between the bone surface and the prosthesis after osteotomy. The calculation results are output in the form of robot-recognizable coordinate parameters, directly guiding the surgical robot to perform precise osteotomy operations.

[0052] In some specific embodiments, the above-mentioned intelligent planning system for osteotomy planes for joint replacement is communicatively connected to the control system of the surgical robot, and is used to send the osteotomy plane equation parameters in the robot coordinate system to the control system of the surgical robot to guide it to perform osteotomy operations.

[0053] In practical applications, the workflow of this embodiment is as follows: Figure 1 As shown: First, the patient's CT medical image is input. After steps such as femoral and tibial image segmentation, key feature extraction and soft tissue performance analysis, intelligent prosthesis matching decision and interpretability analysis, and precise calculation of the osteotomy plane, the osteotomy plane parameters in the robot coordinate system are finally output, completing the entire intelligent planning process of the osteotomy plane.

[0054] It should be noted that this embodiment provides an intelligent planning system for osteotomy planes in joint replacement surgery, including an image segmentation module 100, a feature analysis module 200, an intelligent decision-making module 300, and an osteotomy calculation module 400. It employs deep learning algorithms to achieve intelligent and rapid segmentation of the patient's target replacement joint, utilizes morphological analysis methods to extract and analyze the anatomical features of the target replacement joint, combines biomechanical models to analyze the soft tissue mechanical response characteristics, and then uses a lightweight decision tree-neural network hybrid algorithm based on knowledge distillation to achieve intelligent prosthesis matching. Finally, it calculates the osteotomy plane parameters in the robot coordinate system based on the precisely located prosthesis inner contour plane, thereby reducing the reliance on surgeon experience in surgical planning, increasing the transparency of the surgical planning decision-making process, realizing intelligent planning of osteotomy planes in joint replacement surgery, significantly improving osteotomy accuracy and prosthesis matching degree, reducing the risk of surgical complications, and providing patients with better postoperative functional recovery.

[0055] In a preferred embodiment, this application also provides an electronic device, the electronic device comprising: The computer device includes a memory and a processor. The memory stores computer-readable instructions that, when executed by the processor, implement the intelligent planning method for osteotomy planes in joint replacement. This computer device can be broadly categorized as a server, terminal, or any other electronic device with the necessary computing and / or processing capabilities. In one embodiment, the computer device may include a processor, memory, network interface, communication interface, etc., connected via a system bus. The processor of the computer device can be used to provide the necessary computing, processing, and / or control capabilities. The memory of the computer device may include a non-volatile storage medium and internal memory. The non-volatile storage medium may store an operating system, computer programs, etc. The internal memory can provide an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface and communication interface of the computer device can be used to connect and communicate with external devices via a network. When the computer program is executed by the processor, it performs the steps of the method of the present invention.

[0056] This invention can be implemented as a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, causes the steps of the methods of embodiments of the invention to be performed. In one embodiment, the computer program is distributed across multiple network-coupled computer devices or processors, such that the computer program is stored, accessed, and executed in a distributed manner by one or more computer devices or processors. A single method step / operation, or two or more method steps / operations, may be executed by a single computer device or processor or by two or more computer devices or processors. One or more method steps / operations may be executed by one or more computer devices or processors, and one or more other method steps / operations may be executed by one or more other computer devices or processors. One or more computer devices or processors may execute a single method step / operation, or execute two or more method steps / operations.

[0057] It should be noted that this invention provides an intelligent planning method and system for osteotomy planes in joint replacement. It employs deep learning algorithms to achieve intelligent and rapid segmentation of the patient's target replacement joint, utilizes morphological analysis methods to extract and analyze the anatomical features of the target replacement joint, combines biomechanical models to analyze the soft tissue mechanical response characteristics, and then uses a lightweight decision tree-neural network hybrid algorithm based on knowledge distillation to achieve intelligent prosthesis matching. Finally, it calculates the osteotomy plane parameters in the robot coordinate system based on the precisely located prosthesis inner contour plane. This reduces the reliance on surgeon experience in surgical planning, effectively improves the positioning accuracy of the prosthesis, and solves the technical problems of existing intelligent planning methods for osteotomy planes, such as not considering ligament and soft tissue characteristics, intelligence only reflected in image segmentation, and existing surgical planning relying on surgeon experience.

[0058] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0059] The technical features described above can be combined arbitrarily. Although not all possible combinations of these technical features are described, any combination of these technical features should be considered to be covered by this specification, provided that such combination does not contain contradictions.

[0060] The specific embodiments of the present invention described above do not constitute a limitation on the scope of protection of the present invention. Any other corresponding changes and modifications made in accordance with the technical concept of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A method for intelligent planning of osteotomy planes for joint replacement, characterized in that, include: S100. Obtain the patient's orthopedic medical imaging data, use deep learning algorithms to segment the three-dimensional image of the target replacement joint, and reconstruct the three-dimensional skeletal model of the target replacement joint. S200. Based on the three-dimensional model of the skeleton, extract the anatomical features of the target replacement joint, and use a biomechanical model to analyze the soft tissue around the target replacement joint to obtain the mechanical response characteristics of the soft tissue. S300: The anatomical features and the mechanical response characteristics are used as input features and input into the trained artificial intelligence decision model, and the target prosthesis model that matches the target replacement joint is output. S400. Based on the target prosthesis model, calculate the osteotomy plane parameters of the target prosthesis model in the robot coordinate system.

2. The intelligent planning method for osteotomy planes in joint replacement according to claim 1, characterized in that, The method for segmenting the three-dimensional image of the target replacement joint using a deep learning algorithm and reconstructing the three-dimensional skeletal model of the target replacement joint includes: A deep learning algorithm based on the U-Net model is used to extract the three-dimensional image of the target replacement joint from the orthopedic medical imaging data, and to reconstruct the three-dimensional model of the bone of the target replacement joint in three dimensions.

3. The intelligent planning method for osteotomy planes in joint replacement according to claim 1, characterized in that, The method for extracting the anatomical features of the target replacement joint includes: Based on the three-dimensional skeletal model of the target replacement joint, anatomical features of the target replacement joint are extracted using morphological analysis methods.

4. The intelligent planning method for osteotomy planes in joint replacement according to claim 1, characterized in that, The biomechanical model is a fiber-reinforced hyperelastic model, and the method for analyzing the soft tissue surrounding the target replacement joint using the biomechanical model includes: The anisotropic mechanical behavior of the soft tissue surrounding the target replacement joint was simulated using a fiber-reinforced hyperelastic model, thereby obtaining the mechanical response characteristics of the soft tissue.

5. The intelligent planning method for osteotomy planes in joint replacement according to claim 4, characterized in that, The mechanical response characteristics of the soft tissue are represented by the strain energy function of the fiber-reinforced hyperelastic model, which is composed of the superposition of the matrix contribution and the fiber contribution.

6. The intelligent planning method for osteotomy planes in joint replacement according to claim 1, characterized in that, The artificial intelligence decision-making model is a lightweight decision tree-neural network model based on knowledge distillation. The artificial intelligence decision-making model includes a full neural network as the teacher model and a decision tree model as the student model. The teacher model analyzes the input features, outputs a first feature, and then uses the first feature as the leaf node of the student model's decision tree to output the target prosthesis model that matches the target replacement joint.

7. The intelligent planning method for osteotomy planes in joint replacement according to claim 6, characterized in that, The method for outputting the target prosthesis model that matches the target replacement joint includes: The full neural network performs feature analysis and knowledge extraction on the input anatomical features and mechanical response characteristics, and outputs a knowledge feature vector; The decision tree model takes the knowledge feature vector as input and outputs the target prosthesis model that matches the target replacement joint based on the decision rules.

8. The intelligent planning method for osteotomy planes in joint replacement according to claim 1, characterized in that, The method for calculating the osteotomy plane parameters of the target prosthesis model in the robot coordinate system includes: Obtain the first osteotomy plane parameters of the target prosthesis model in the prosthesis coordinate system, transform the first osteotomy plane parameters in the prosthesis coordinate system to the robot coordinate system through coordinate transformation to obtain the second osteotomy plane parameters, and use the second osteotomy plane parameters as the geometric constraint conditions of the replacement prosthesis.

9. An intelligent planning system for osteotomy planes in joint replacement, characterized in that, The system includes: Image segmentation module: used to acquire orthopedic medical imaging data of patients, segment the three-dimensional image of the target replacement joint using deep learning algorithms, and reconstruct the three-dimensional skeletal model of the target replacement joint; Feature analysis module: Based on the three-dimensional model of the skeleton, it is used to extract the anatomical features of the target replacement joint and to analyze the soft tissue around the target replacement joint using a biomechanical model to obtain the mechanical response characteristics of the soft tissue; Intelligent decision-making module: used to input the anatomical features and the mechanical response characteristics as input features into the trained artificial intelligence decision-making model, and output the target prosthesis model that matches the target replacement joint; Osteotomy Calculation Module: Used to calculate the osteotomy plane parameters of the target prosthesis model in the robot coordinate system based on the target prosthesis model.

10. An electronic device, characterized in that, include: Memory; And a processor, wherein the memory stores computer-readable instructions that, when executed by the processor, implement the intelligent planning method for osteotomy planes for joint replacement according to any one of claims 1 to 8.

Images