Two-dimensional total hip replacement prosthesis type selection method and device based on reinforcement learning, medium, program product and terminal

Abstract

According to the reinforcement learning-based two-dimensional total hip replacement prosthesis type selection method and device, the medium, the program product and the terminal provided by the invention, the prosthesis type selection model is constructed based on the Unity platform and the reinforcement learning algorithm, so that a proper prosthesis is automatically selected according to the medullary cavity condition of a patient and preoperative planning is carried out, the operation efficiency and accuracy are remarkably improved, and the operation cost is reduced. And errors caused by manual operation are reduced. Meanwhile, the model supports personalized setting, a labeling scheme can be customized according to patient requirements, personalized prosthesis type selection is achieved, and individual requirements of different patients are met. Besides, the model has expansibility, can be used for preoperative planning to a certain extent, can be applied to automatic planning in other surgical fields in the future, and has important clinical application value.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of intelligent medical treatment, and in particular to a two-dimensional total hip arthroplasty prosthesis selection method based on reinforcement learning. BACKGROUND

[0002] Total hip arthroplasty (THA) is a common repair surgery that replaces all or part of the hip joint with an artificial hip joint to restore and improve joint movement function. For femoral head necrosis, there are various non-surgical options in clinical practice, such as using non-steroidal anti-inflammatory drugs to reduce inflammation, using physiotherapy to restore joint and muscle function, etc. However, these methods can only alleviate the worsening of the disease and cannot completely cure the inconvenience caused by femoral head necrosis. Therefore, in order to completely cure femoral head necrosis or in the case where other treatment options are ineffective, THA surgery is often used to remove the necrotic femoral head and replace it with a prosthetic device to replace its function.

[0003] Traditional surgical options rely on manual analysis of medical imaging data to select prostheses, which is time-consuming and has limited accuracy, and lacks quantitative evaluation. Existing prosthetic materials and designs have defects and cannot meet the individual needs of patients. Incorrect prosthesis selection can lead to serious complications, causing great pain and economic burden to patients. Therefore, there is an urgent need for a precise and efficient prosthesis selection method to improve surgical outcomes and patient quality of life. SUMMARY

[0004] In view of the above-mentioned shortcomings of the prior art, the purpose of the present application is to provide a two-dimensional total hip arthroplasty prosthesis selection method, device, medium, program product and terminal based on reinforcement learning, which is used to solve the problems of long time consumption, low accuracy, inability to meet the individual needs of patients, high risk of incorrect selection, and great pain and economic burden to patients caused by the traditional artificial total hip arthroplasty prosthesis selection method in the prior art.

[0005] To achieve the above-mentioned purposes and other related purposes, the first aspect of the present application provides a two-dimensional total hip arthroplasty prosthesis selection method based on reinforcement learning, comprising: acquiring two-dimensional section data of a selected medullary cavity, and acquiring two-dimensional information of a plurality of types of prostheses; performing anchor point labeling operation in the two-dimensional section data to generate a plurality of anchor points; importing the two-dimensional information of the plurality of types of prostheses and the plurality of anchor points into a Unity platform; in the Unity platform, constructing a reinforcement learning model based on the two-dimensional information of the plurality of types of prostheses and the plurality of anchor points and performing training optimization, and selecting a prosthesis type with the highest matching degree from the plurality of types of prostheses.

[0006] In some embodiments of the first aspect of the present application, the process of obtaining the two-dimensional section data of the selected type of medullary cavity comprises: obtaining three-dimensional image data of the selected type of medullary cavity; slicing the three-dimensional image data based on a preset direction to generate a two-dimensional section; extracting the intensity value of each voxel on the two-dimensional section to generate two-dimensional section data.

[0007] In some embodiments of the first aspect of the present application, the process of constructing a reinforcement learning model based on multiple types of two-dimensional information of prostheses and multiple anchor points and optimizing the training comprises: constructing a Unity environment for prosthesis selection based on multiple types of two-dimensional information of prostheses and multiple anchor points; importing an ML-Agents toolkit into the Unity environment for prosthesis selection; defining a state space, an action space and a reward function of a prosthesis selection model based on a preset reinforcement learning algorithm; training and evaluating the prosthesis selection model to generate an optimized prosthesis selection model.

[0008] In some embodiments of the first aspect of the present application, the process of selecting a prosthesis type with the highest matching degree with the anchor points from multiple types of prostheses comprises: inputting the multiple types of two-dimensional information of prostheses into the trained prosthesis selection model to generate a total hip arthroplasty simulation result corresponding to each type of prosthesis; calculating the matching degree of multiple simulation results and multiple anchor points to generate a prosthesis type with the highest matching degree with the anchor points.

[0009] In some embodiments of the first aspect of the present application, the preset reinforcement learning algorithm comprises any one of a Q-learning algorithm, a DDPG algorithm and a proximal policy optimization algorithm.

[0010] In some embodiments of the first aspect of the present application, the two-dimensional information of the prosthesis comprises one or more of a prosthesis number, a prosthesis size, a prosthesis shape and a prosthesis material.

[0011] To achieve the above object and other related objects, the second aspect of the present application provides a two-dimensional total hip arthroplasty prosthesis selection device based on reinforcement learning, comprising: a data acquisition module for obtaining two-dimensional section data of a selected type of medullary cavity and obtaining two-dimensional information of multiple types of prostheses; an anchor point labeling module for performing anchor point labeling operation in the two-dimensional section data to generate multiple anchor points; a prosthesis selection module for importing multiple types of two-dimensional information of prostheses and multiple anchor points into a Unity platform; in the Unity platform, constructing a reinforcement learning model based on multiple types of two-dimensional information of prostheses and multiple anchor points and optimizing the training, and selecting a prosthesis type with the highest matching degree with the anchor points from multiple types of prostheses.

[0012] To achieve the above object and other related objects, the third aspect of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the two-dimensional total hip arthroplasty prosthesis selection method based on reinforcement learning.

[0013] To achieve the above object and other related objects, the fourth aspect of the present application provides a computer program product, which includes computer program codes, and when the computer program codes are run on a computer, the computer program codes make the computer implement the two-dimensional total hip arthroplasty prosthesis selection method based on reinforcement learning.

[0014] To achieve the above object and other related objects, the fifth aspect of the present application provides an electronic terminal, which includes a memory, a processor and a computer program stored in the memory; the processor executes the computer program to implement the two-dimensional total hip arthroplasty prosthesis selection method based on reinforcement learning.

[0015] As described above, the present application has the following beneficial effects: the present application can automatically select a suitable prosthesis according to the patient's medullary cavity condition and perform preoperative planning, significantly improving the efficiency and accuracy of the operation and reducing the errors caused by manual operation. At the same time, the model supports personalized settings, can customize the labeling scheme according to the patient's needs, realizes personalized prosthesis selection, and meets the individual needs of different patients. In addition, the model also has extensibility and can perform a certain degree of preoperative planning, which can be applied to automatic planning in other surgical fields in the future, and has important clinical application value. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 A flowchart of an embodiment of the two-dimensional total hip arthroplasty prosthesis selection method based on reinforcement learning of the present application is shown.

[0017] Figure 2 A schematic diagram of the relative position of the medullary cavity and the prosthesis in an embodiment of the two-dimensional total hip arthroplasty prosthesis selection method based on reinforcement learning of the present application is shown.

[0018] Figure 3 A schematic diagram of the medullary cavity after anchor point labeling in an embodiment of the two-dimensional total hip arthroplasty prosthesis selection method based on reinforcement learning of the present application is shown.

[0019] Figure 4 A schematic diagram of the comparison results of different prosthesis selection in an embodiment of the two-dimensional total hip arthroplasty prosthesis selection method based on reinforcement learning of the present application is shown.

[0020] Figure 5 A structural schematic diagram of an embodiment of the two-dimensional total hip arthroplasty prosthesis selection device based on reinforcement learning of the present application is shown.

[0021] Figure 6 A structural schematic diagram of an embodiment of a two-dimensional total hip arthroplasty prosthesis selection terminal based on reinforcement learning according to the present application is shown. DETAILED DESCRIPTION

[0022] The embodiments of the present application will be described in detail with specific examples below, and other advantages and effects of the present application can be easily understood by those skilled in the art from the disclosure. The present application can also be implemented or applied by other different specific embodiments, and various modifications or changes can be made to the details in the specification based on different views and applications without departing from the spirit of the present application. It should be noted that the following examples and features in the examples can be combined with each other without conflict.

[0023] Before the present application is further described, the terms and terminology used in the embodiments of the present application are explained, and the terms and terminology used in the embodiments of the present application are applicable to the following explanations:

[0024] <1> Unity platform: Unity is a cross-platform game development engine that provides powerful tools and resources for creating various types of 2D and 3D games.

[0025] <2> NII.GZ format: NII.GZ is a file format for medical image data, commonly used to store medical image data such as magnetic resonance imaging (MRI) and computed tomography (CT).

[0026] <3> Harris corner detection algorithm: Harris corner detection algorithm is a commonly used algorithm in computer vision field, which can be used to detect corner points (local maximum feature points) in images, with stability and robustness.

[0027] <4> SIFT algorithm: SIFT (Scale-Invariant Feature Transform) algorithm is an algorithm for extracting key points in images and describing local features, with rotation, scale and illumination invariance.

[0028] <5> Proximal policy optimization algorithm: Proximal policy optimization algorithm is an algorithm for solving proximal optimization problems, which optimizes the objective function by finding the proximal solution of the gradient, and is widely used in machine learning, data mining and optimization.

[0029] To facilitate understanding of the embodiments of the present application, first, the structure and working principle of the Unity platform will be described in detail. Figure 1 Detailed description. Figure 1 A flowchart of a two-dimensional total hip arthroplasty prosthesis selection method based on reinforcement learning in an embodiment of the present application is shown. The two-dimensional total hip arthroplasty prosthesis selection method based on reinforcement learning in the present embodiment mainly includes the following steps:

[0030] Step S11: Obtain two-dimensional section data of the selected type of medullary cavity, and obtain two-dimensional information of multiple types of prostheses.

[0031] In an embodiment of the present application, the process of obtaining two-dimensional section data of the selected type of medullary cavity includes: obtaining three-dimensional image data of the selected type of medullary cavity; slicing the three-dimensional image data based on a preset direction to generate a two-dimensional section; and extracting the intensity value of each voxel on the two-dimensional section to generate two-dimensional section data. Figure 2 The relative position between the medullary cavity and the prosthesis is shown in the embodiment.

[0032] Further, the three-dimensional image data of the selected type of medullary cavity is in a medical image format, including but not limited to DICOM format or NIfTI format. The preset slicing direction includes a transverse section, a sagittal section, and a coronal section.

[0033] In an embodiment of the present application, the two-dimensional section data extraction process includes: voxel intensity value extraction, data format conversion, and data storage. Each pixel point on each two-dimensional section represents a voxel, and each voxel has an intensity value for describing the gray value or other physical properties of the voxel. First, extract the intensity value of each voxel on each section, store the extracted voxel intensity value as a matrix or other data structure, and exemplarily store it using a two-dimensional array. Finally, store the two-dimensional section data using an NII.GZ format file. NII.GZ is a compression format for storing neuroimaging data, which has high storage and transmission efficiency.

[0034] Step S12: Perform anchor point labeling operation in the two-dimensional section data to generate multiple anchor points.

[0035] In an embodiment of the present application, when performing the anchor point labeling operation on the two-dimensional section data, a feature point detection algorithm in computer vision technology is used, exemplarily, a Harris corner detection algorithm or a SIFT (Scale-Invariant Feature Transform) algorithm is used to identify key feature points in the image. Subsequently, a clustering algorithm (such as k-means clustering) is used to generate multiple anchor points according to the positions of these feature points. In addition, a deep learning model such as a convolutional neural network (CNN) is used to train an end-to-end anchor point detector to improve the accuracy and efficiency of anchor point generation. By combining these technical means, the operation of automatically generating multiple anchor points is realized.

[0036] In an embodiment of the present application, key anchor points such as joint heads on the medullary cavity and the prosthesis can be automatically or manually labeled according to clinical actual needs.

[0037] It is worth mentioning that the calibrated anchor points can serve as evaluation benchmarks in reinforcement learning training, helping to speed up the subsequent training process. In addition, these key points are crucial for clinical applications, ensuring that the automatically selected prosthesis model meets the actual needs and facilitates the integration of future complex automation projects. For example, when annotating the position of the joint head corresponding to the upper end of the prosthesis, the annotated model achieves significant improvements in training speed and final results compared to the unannotated model. Figure 3 The figure shows the intramedullary canal after the anchor point annotation in an embodiment of the present application.

[0038] Step S13: Importing multiple prosthesis two-dimensional information of the type and multiple anchor points into the Unity platform; in the Unity platform, constructing a reinforcement learning model based on multiple prosthesis two-dimensional information of the type and multiple anchor points and performing training optimization to select the prosthesis type with the highest matching degree from multiple prostheses of the type.

[0039] In an embodiment of the present application, the process of constructing a reinforcement learning model based on multiple prosthesis two-dimensional information of the type and multiple anchor points and performing training optimization includes: based on multiple prosthesis two-dimensional information of the type and multiple anchor points, constructing a prosthesis selection Unity environment; importing an ML-Agents toolkit into the prosthesis selection Unity environment; based on a preset reinforcement learning algorithm, defining the state space, action space, and reward function of the prosthesis selection model; training and evaluating the prosthesis selection model to generate an optimized prosthesis selection model.

[0040] In an embodiment of the present application, before training, the clinician uses a deep reinforcement learning algorithm to optimize the model, which includes but is not limited to adjusting the complexity and realism of the simulation environment, optimizing the design of the reward function, and optimizing the network structure and learning rate. In addition, by manually performing and recording key information during the simulation implantation process, it helps to identify problems and bottlenecks in model training, further fine-tune model parameters, and improve the effectiveness and reliability of the model in clinical practice. These measures not only effectively improve the performance of the model, but also ensure the fit and safety of the model with actual clinical operations, providing more accurate and reliable decision-making tools for clinicians.

[0041] In an embodiment of the present application, the process of selecting the prosthesis type with the highest matching degree from multiple prostheses of the type includes: inputting the multiple prosthesis two-dimensional information of the type into the trained prosthesis selection model to generate the total hip arthroplasty simulation result corresponding to each type of prosthesis; calculating the matching degree of multiple simulation results and multiple anchor points to generate the prosthesis type with the highest matching degree with the anchor points.

[0042] Further, the process of calculating the matching degree of the plurality of simulation results and the plurality of anchor points includes: calculating the similarity or distance of the simulation results and the anchor points to determine the matching degree. Wherein, the calculation method of similarity or relative distance includes but is not limited to: Euclidean distance, cosine similarity, correlation coefficient, Mahalanobis distance and Hamming distance algorithm. After multiple matching degree calculations are performed on the plurality of simulation results and the anchor points, the prosthesis type with the highest matching degree is found, that is, the most suitable simulation result.

[0043] In an embodiment of the present application, the present application builds an automatic prosthesis selection model for the two-dimensional structure of the prosthesis and the medullary cavity based on Unity and its ML-agents reinforcement learning toolkit. The model mainly consists of two core algorithms, namely collision detection algorithm and reinforcement learning algorithm. The collision detection algorithm is based on the existing Ray Perception Sensor component of Unity, which realizes the detection of the collision position and collision degree that may occur in the THA operation area such as the inner wall of the medullary cavity. In addition, since this project focuses on two-dimensional problems, a Camera Sensor is also used to process information under two-dimensional conditions to provide more comprehensive information support and assist the performance of the Ray Perception Sensor in the training process. The present application optimizes the calculation process on the basis of the algorithm, so that the model can efficiently calculate the collision between the prosthesis and the medullary cavity with less consumption of computing resources, providing a basis for the model to quickly give automatic prosthesis selection.

[0044] Further, the Ray Perception Sensor and the Camera Sensor are commonly used sensor components in Unity, which are used to simulate the perception of light by objects and the function of camera shooting. The Ray Perception Sensor detects the position and normal direction of the collision point by emitting virtual light, helping the model to detect the position and degree of possible collision between the prosthesis and the inner wall of the medullary cavity. The Camera Sensor simulates the shooting and recognition function of the camera to the object, obtains image information to provide visual information to assist the model to understand the environment. Through the cooperation of the Ray Perception Sensor and the Camera Sensor, the model can more accurately perceive environmental information, make collision detection and selection of prosthesis decision, and improve the accuracy and efficiency in the process of automatic prosthesis selection.

[0045] In an embodiment of the present application, the preset reinforcement learning algorithm includes any one of Q-learning algorithm, DDPG algorithm and proximal policy optimization algorithm.

[0046] Further, the process of training and optimization using the proximal policy optimization algorithm includes: first, by adjusting the parameters in the collision detection algorithm, such as the number of light rays, the range and the angle range of the light ray perception detector, to improve the perception accuracy and range of the model to collision information. At the same time, the resolution and angle of the camera detector are adjusted to improve the model's ability to identify and understand environmental information. Subsequently, by optimizing the hyperparameters of the reinforcement learning algorithm, such as learning rate, discount factor and exploration rate, to improve the training efficiency and performance of the model. At the same time, the experience replay mechanism and the design of the reward function are introduced to speed up the convergence speed of the model and improve the stability of the training.

[0047] In an embodiment of the present application, the prosthesis two-dimensional information includes one or more of the following: prosthesis number, prosthesis size, prosthesis shape, prosthesis material.

[0048] Further, Figure 4 In this embodiment, the comparison results of different prosthesis selection are shown. In THA surgery, the size of the prosthesis is usually selected according to the size and anatomical structure of the patient's hip joint to ensure that the prosthesis can be properly installed and restore the function of the hip joint. Common prosthesis size ranges from small, medium, to large. The shape of the prosthesis includes straight handle, tapered handle, S-shaped handle and other forms. Different shapes of prosthesis are suitable for different hip joint diseases and patient characteristics. The material of the prosthesis usually includes metal alloy, ceramic and plastic. Metal alloy (such as titanium alloy, stainless steel) prosthesis has good durability and bearing capacity, ceramic prosthesis has good anti-wear performance, and plastic prosthesis can reduce friction noise.

[0049] It should be noted that in the embodiments of the present application, the words "exemplary" or "for example" indicate an example, illustration or description. Any embodiment or design scheme described as "exemplary" or "for example" in the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. Rather, the use of "exemplary" or "for example" is intended to present the relevant concept in a specific manner.

[0050] In the embodiments of the present application, "at least one" means one or more, and "multiple" means two or more. The association relationship between the associated objects is described, which means that there can be three kinds of relationships, for example, A and / or B, which can represent the following cases: A exists alone, A and B exist together, and B exists alone, where A and B can be singular or plural. The character " / " generally represents an "or" relationship between the associated objects before and after it. "At least one of the following" or similar expressions means any combination of these items, including any combination of single item or multiple items. For example, at least one of a, b or c can represent a, b, c, a-b, a-c, b-c or a-b-c, where a, b and c can be single or multiple.

[0051] Figure 5 is a schematic block diagram of a two-dimensional total hip arthroplasty prosthesis selection device based on reinforcement learning provided by an embodiment of the present application. As shown in the figure, Figure 5 the device comprises a data acquisition module 501, an anchor point labeling module 502, and a prosthesis selection module 503.

[0052] The data acquisition module 501 is configured to acquire two-dimensional section data of a medullary cavity to be selected, and acquire two-dimensional information of multiple types of prostheses.

[0053] The anchor point labeling module 502 is configured to perform anchor point labeling operations in the two-dimensional section data to generate multiple anchor points.

[0054] The prosthesis selection module 503 is configured to import the two-dimensional information of multiple types of prostheses and the multiple anchor points into a Unity platform. In the Unity platform, a reinforcement learning model is constructed based on the two-dimensional information of multiple types of prostheses and the multiple anchor points, and is trained and optimized, and a type of prosthesis with the highest matching degree with the anchor points is selected from the multiple types of prostheses.

[0055] It should be understood that the specific processes of each module performing the above corresponding steps have been described in detail in the above method embodiments, and for the sake of brevity, will not be described here.

[0056] It should also be understood that the division of the modules in the embodiments of the present application is illustrative, and is only a logical functional division. In actual implementation, there can be another division mode. In addition, each functional module in each embodiment of the present application can be integrated in one processor, or can be physically separated, or two or more modules can be integrated in one module. The integrated module can be realized in the form of hardware or in the form of a software functional module.

[0057] Figure 6 is a schematic block diagram of an electronic terminal provided by an embodiment of the present application. As shown in the figure, Figure 6 the electronic terminal comprises at least one processor 601, a memory 602, at least one network interface 603, and a user interface 605. Each component in the device is coupled together through a bus system 604. It can be understood that the bus system 604 is used to realize the connection and communication between the components. In addition to the data bus, the bus system 604 also includes a power bus, a control bus, and a status signal bus. However, for the sake of clarity, all kinds of buses are marked as a bus system in Figure 6 .

[0058] The user interface 605 can include a display, a keyboard, a mouse, a trackball, a click gun, a key, a button, a touchpad, or a touch screen, etc.

[0059] It is to be understood that the memory 602 can be volatile or nonvolatile memory, or both. Non-volatile memory can be read only memory (ROM), programmable ROM (PROM), which is used as an external cache. By way of example, but not limitation, many forms of RAM can be used, such as static random access memory (SRAM), synchronous static random access memory (SSRAM). The memory described in embodiments of the present application is intended to include, but not be limited to, these and any other suitable category of memory.

[0060] The memory 602 in embodiments of the present application is used to store various categories of data to support the operation of the electronic terminal 600. Examples of these data include: any executable programs used to operate on the electronic terminal 600, such as an operating system 6021 and application programs 6022; the operating system 6021 contains various system programs, such as a framework layer, a core library layer, a driver layer, etc., for implementing various basic services and processing hardware-based tasks. The application programs 6022 can contain various application programs, such as a media player (Media Player), a browser (Browser), etc., for implementing various application services. The implementation of the 66 method provided by embodiments of the present application can be included in the application programs 6022.

[0061] The method disclosed in the embodiments of the present application can be applied to the processor 601 or implemented by the processor 601. The processor 601 can be an integrated circuit chip having a signal processing capability. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in the processor 601 or an instruction in the form of software. The processor 601 described above can be a general processor, a digital signal processor (DSP), or other programmable logic device, discrete gate or transistor logic device, discrete hardware component, etc. The processor 601 can implement or execute the disclosed methods, steps and logic block diagrams in the embodiments of the present application. The general processor 601 can be a microprocessor or any conventional processor, etc. In combination with the steps of the accessory optimization method provided in the embodiments of the present application, the steps can be directly embodied as hardware decoding processor execution, or executed by a combination of hardware and software modules in the decoding processor. The software module can be located in a storage medium, which is located in a memory, and the processor reads the information in the memory and combines the hardware to complete the steps of the foregoing method.

[0062] In the exemplary embodiments, the electronic terminal 600 can be one or more application specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), or the like for executing the foregoing method.

[0063] According to the method provided in the embodiments of the present application, the present application further provides a computer program product, which comprises computer program code, when the computer program code runs on a computer, so that the computer executes Figures 1 to 4 The method for selecting a two-dimensional total hip arthroplasty prosthesis based on reinforcement learning according to any one of the embodiments shown.

[0064] According to the method provided in the embodiments of the present application, the present application further provides a computer readable storage medium, which stores program code, when the program code runs on a computer, so that the computer executes Figures 1 to 4 The method for selecting a two-dimensional total hip arthroplasty prosthesis based on reinforcement learning according to any one of the embodiments shown.

[0065] As used in this description, the terms "component," "module," "system," and the like are intended to refer to a computer-related entity, either hardware, firmware, a combination of hardware and software, software, or software in execution. For example, a component can be, but is not limited to being, a process running on a processor, a processor, an object, an executable, a thread of execution, a program, and / or a computer. By way of illustration, both an application running on a computing device and the computing device can be a component. One or more components can reside within a process and / or thread of execution and a component can be localized, partially localized, and / or distributed across two or more computers. Also, these components can execute from various computer readable media having various data structures stored thereon. The components can communicate by way of local and / or remote processes such as in accordance with a signal having one or more data packets (e.g., data from one component interacting with another component in a local system, distributed system, and / or across a network such as the Internet with other systems via the signal).

[0066] Those of skill in the art would understand that the various illustrative logical blocks, modules, and steps described in connection with the embodiments disclosed herein can be implemented as electronic hardware, or a combination of computer software and electronic hardware. The choice of hardware and software would depend on the particular application and general purpose nature of the technical solution. Skilled artisans can implement the described functionality in varying ways for each particular application, but such implementation would not render the application or its various embodiments of the application invalid.

[0067] Those of skill in the art would understand that, for the described convenience and brevity, the specific working process of the system, device and unit described above can refer to the corresponding process in the foregoing method embodiment, which will not be described here.

[0068] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the device embodiments described above are merely illustrative, for example, the division of units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed each other can be through some interface, indirect coupling or communication connection between the devices or units, which can be electrical, mechanical or other forms.

[0069] The units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, i.e., may be located in one place, or may be distributed to multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0070] In addition, each functional unit in each embodiment of the present application can be integrated into one processing unit, or each unit can exist physically, or two or more units can be integrated into one unit.

[0071] In the above embodiments, the functions of each functional unit can be implemented by software, hardware, firmware, or any combination thereof, in whole or in part. When implemented by software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions (programs). When the computer program instructions (programs) are loaded and executed on a computer, the flow or function according to the embodiments of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. Computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another, for example, computer instructions can be transferred from one website, computer, server or data center to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (Digital Subscriber Line, DSL)) or wireless (such as infrared, wireless, microwave, etc.) mode. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. integrated with one or more available media. The available media can be magnetic media (for example, floppy disk, hard disk, magnetic tape), optical media (for example, high-density digital video disc (Digital Video Disc, DVD), or semiconductor media (for example, solid state disk (Solid State Disk, SSD) and the like.

[0072] If the functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application can be embodied in the form of a software product, which is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0073] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0074] In summary, the present application provides a two-dimensional total hip arthroplasty prosthesis selection method and device based on reinforcement learning, a medium, a program product and a terminal. The present application provides a method for improving the efficiency of total hip arthroplasty prosthesis selection. A prosthesis selection model is constructed based on the Unity platform and reinforcement learning algorithm to automatically select appropriate prostheses according to the patient's medullary cavity condition and perform preoperative planning, significantly improving the efficiency and accuracy of the operation and reducing the errors caused by manual operation. At the same time, the model supports personalized settings, can customize the labeling scheme according to patient needs, and realizes personalized prosthesis selection to meet the individual needs of different patients. In addition, the model also has scalability and can perform a certain degree of preoperative planning. It can be applied to automatic planning in other surgical fields in the future and has important clinical application value. Therefore, the present application effectively overcomes the various shortcomings in the prior art and has high industrial utilization value.

[0075] The above embodiments only exemplarily illustrate the principles and effects of the present application, and are not intended to limit the present application. Any person skilled in the art can modify or change the above embodiments without departing from the spirit and scope of the present application. Therefore, all equivalent modifications or changes made by those skilled in the art without departing from the spirit and technical idea disclosed in the present application should be covered by the claims of the present application.

Claims

1. A method for prosthesis selection in two-dimensional total hip arthroplasty based on reinforcement learning, characterized in that, include: Obtain two-dimensional cross-sectional data of the medullary cavity of the candidate type, and obtain two-dimensional information of multiple types of prostheses; Anchor point annotation is performed on the two-dimensional section data to generate multiple anchor points; Import the two-dimensional information of multiple prostheses of the aforementioned types and the multiple anchor points into the Unity platform; In the Unity platform, a reinforcement learning model is constructed and trained and optimized based on the two-dimensional information of multiple types of prostheses and multiple anchor points, and the prosthesis type with the highest matching degree with the anchor point is selected from the multiple types of prostheses.

2. The method for selecting prostheses for two-dimensional total hip arthroplasty based on reinforcement learning according to claim 1, characterized in that, The process of obtaining two-dimensional section data of the candidate medullary canal includes: Obtain three-dimensional image data of the medullary cavities to be selected; The three-dimensional image data is sliced ​​according to a preset direction to generate a two-dimensional cross-section; The intensity value of each voxel on the two-dimensional cross-section is extracted to generate two-dimensional cross-section data.

3. The method for selecting prostheses for two-dimensional total hip arthroplasty based on reinforcement learning according to claim 1, characterized in that, The process of constructing and training a reinforcement learning model based on two-dimensional information of multiple types of prostheses and multiple anchor points includes: Based on the two-dimensional information of multiple types of prostheses and the multiple anchor points, a prosthesis selection Unity environment is constructed. Import the ML-Agents toolkit into the prosthesis selection Unity environment; Based on a pre-defined reinforcement learning algorithm, the state space, action space, and reward function of the prosthesis selection model are defined. The prosthesis selection model is trained and evaluated to generate an optimized prosthesis selection model.

4. The method for selecting prostheses for two-dimensional total hip arthroplasty based on reinforcement learning according to claim 3, characterized in that, The process of selecting the prosthesis type with the highest matching degree to the anchor point from a plurality of prostheses of the aforementioned types includes: The two-dimensional information of the multiple types of prostheses is input into the trained prosthesis selection model to generate the simulation results of total hip arthroplasty corresponding to each type of prosthesis; Matching degree calculations are performed on multiple simulation results and multiple anchor points to generate the prosthesis type with the highest matching degree to the anchor points.

5. The method for selecting prostheses for two-dimensional total hip arthroplasty based on reinforcement learning according to claim 3, characterized in that, The preset reinforcement learning algorithm includes any one of the following: Q-learning algorithm, DDPG algorithm, and proximal policy optimization algorithm.

6. The method for selecting prostheses for two-dimensional total hip arthroplasty based on reinforcement learning according to claim 1, characterized in that, The two-dimensional information of the prosthesis includes one or more of the following: prosthesis number, prosthesis size, prosthesis shape, and prosthesis material.

7. A two-dimensional total hip arthroplasty prosthesis selection device based on reinforcement learning, characterized in that, include: Data acquisition module: used to acquire two-dimensional cross-sectional data of the medullary canal of the candidate and to acquire two-dimensional information of multiple types of prostheses; Anchor point annotation module: used to perform anchor point annotation operations on the two-dimensional section data to generate multiple anchor points; Prosthesis selection module: used to import two-dimensional information of multiple prostheses of the aforementioned types and multiple anchor points into the Unity platform; In the Unity platform, a reinforcement learning model is constructed and trained and optimized based on the two-dimensional information of multiple types of prostheses and multiple anchor points, and the prosthesis type with the highest matching degree with the anchor point is selected from the multiple types of prostheses.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the reinforcement learning-based prosthesis selection method for two-dimensional total hip arthroplasty as described in any one of claims 1 to 6.

9. A computer program product, characterized in that, The computer program product includes computer program code, which, when run on a computer, enables the computer to implement the reinforcement learning-based two-dimensional total hip arthroplasty prosthesis selection method as described in any one of claims 1 to 6.

10. An electronic terminal, comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the reinforcement learning-based prosthesis selection method for two-dimensional total hip arthroplasty as described in any one of claims 1 to 6.

Images