Arthroscopic surgery navigation and positioning system based on artificial intelligence

Abstract

The invention relates to the technical field of navigation and positioning, and discloses an arthroscopic surgery navigation and positioning system based on artificial intelligence, which comprises a preoperative image processing and lesion analysis module for analyzing preoperative images, a multi-modal data fusion module for integrating image data from different sources in the surgery, and a navigation and positioning module for performing navigation and positioning on the integrated image data. The intra-operative real-time joint tracking and posture estimation module is used for tracking the joint position and soft tissue deformation in real time, the AI dynamic operation path planning module is used for intelligently optimizing the operation path, the robot auxiliary operation and automatic adjustment module is used for automatically adjusting the position and angle of an operation instrument, and the intra-operative AI voice and visual feedback module is used for providing real-time feedback. The operation data storage and intelligent learning module is used for storing and analyzing operation data. By means of the multi-modal data fusion technology, arthroscopic images, ultrasonic images and infrared images are combined, accurate visualization of an operation area is achieved, and the effect of improving operation accuracy and safety is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of navigation and positioning, and particularly to an arthroscopic surgery navigation and positioning system based on artificial intelligence. Background Art

[0002] In modern surgical operations, especially arthroscopic surgeries, doctors often rely on preoperative images (such as CT, MRI) and intraoperative images (such as arthroscopes, ultrasounds, infrareds, etc.) to guide the surgical operations. However, a single imaging modality often cannot provide sufficient perspectives and information. Especially when dealing with complex anatomical structures and dynamically changing lesions, it is prone to cause positioning errors and insufficient surgical precision. For example, although arthroscopic images can provide direct-view images, they cannot effectively present structures such as soft tissues and blood vessels that are difficult to display; while ultrasound images can display the conditions of soft tissues and joint fluids, but due to limitations in resolution and operation, it is difficult to accurately locate deep structures; infrared images can effectively detect temperature changes and indicate potential abnormal blood flows, but cannot provide detailed anatomical structure information.

[0003] Therefore, a single-modal image has many limitations in the actual surgical process, especially in complex surgical operations, and cannot comprehensively reflect all the details and dynamic changes in the surgical area, increasing the complexity and risk of the operation. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the present invention provides an arthroscopic surgery navigation and positioning system based on artificial intelligence, which solves the problem that a single imaging modality cannot comprehensively display all important information in the surgical area, such as the morphology of soft tissues, the position of blood vessels, and the dynamic changes of lesions, resulting in doctors not being able to obtain a comprehensive view during the surgical process and increasing the risk of misoperation.

[0005] To achieve the above objectives, the present invention is realized through the following technical solutions: An arthroscopic surgery navigation and positioning system based on artificial intelligence, comprising:

[0006] A preoperative image processing and lesion analysis module for analyzing preoperative images and automatically identifying joint structures and lesion areas;

[0007] A multi-modal data fusion module for integrating intraoperative image data from different sources to provide a clear surgical view;

[0008] An intraoperative real-time joint tracking and pose estimation module for real-time tracking of joint positions and soft tissue deformations;

[0009] An AI dynamic surgical path planning module for intelligently optimizing the surgical path, preventing damage to important tissues, and dynamically adjusting the path;

[0010] The robot-assisted operation and automatic adjustment module is used to automatically adjust the position and angle of surgical instruments;

[0011] The intraoperative AI voice and visual feedback module is used to provide real-time feedback to assist the doctor in adjusting the surgical operation;

[0012] The surgical data storage and intelligent learning module is used to store and analyze surgical data.

[0013] Preferably, the preoperative image processing and lesion analysis module includes an image segmentation unit, a lesion detection unit, and a 3D reconstruction unit. The image segmentation unit is used to segment the joint structure and the lesion area by using a deep learning model, and the learning model includes U-Net. The lesion detection unit is used to automatically identify the lesion area by using an object detection network, and the object detection network includes Faster R-CNN and YOLO. The 3D reconstruction unit is used to reconstruct the 3D joint model based on the image data to provide a three-dimensional view for surgical planning.

[0014] Preferably, the multi-modal data fusion module includes an image fusion unit, an image enhancement unit, and a multi-modal tracking unit. The image fusion unit is used to fuse multi-modal image data, and the multi-modal image data fusion includes arthroscopic images, ultrasound, and infrared. The image enhancement unit is used to remove the influence by using an image enhancement algorithm, and the influence includes blood and smoke. The multi-modal tracking unit is used to combine different modal data to real-time track the surgical area to enhance visualization.

[0015] Preferably, the intraoperative real-time joint tracking and pose estimation module includes a bone tracking unit, a soft tissue deformation prediction unit, and a joint positioning and pose estimation unit. The bone tracking unit is used to real-time track the dynamic position of the bone based on a graph neural network. The soft tissue deformation prediction unit is used to predict the soft tissue deformation by using the LSTM method and correct the navigation system until the adjustment is accurate. The joint positioning and pose estimation unit is used to estimate the precise position and pose of the joint through optical tracking technology.

[0016] Preferably, the AI dynamic surgical path planning module includes a path planning unit, a dynamic adjustment unit, and an obstacle avoidance unit. The path planning unit is used to intelligently calculate the optimal surgical path based on preoperative images and real-time data. The dynamic adjustment unit is used to adjust the path planning according to the surgical progress and real-time feedback. The obstacle avoidance unit: real-time detects key structures during the operation to avoid accidental injury to the instruments, and the key structures include blood vessels and nerves.

[0017] Preferably, the robot-assisted operation and automatic adjustment module includes an automatic adjustment unit, a force sensing unit, and a feedback and cooperation unit. The automatic adjustment unit is used to automatically adjust the position, angle, and depth of the surgical instrument according to the AI navigation result. The force sensing unit is used to monitor the contact force of the instrument to prevent over-operation and tissue damage. The feedback and cooperation unit is used to provide real-time feedback and coordinate the operations of the robot and the doctor.

[0018] Preferably, the intraoperative AI voice and visual feedback module includes a voice recognition unit, a visual assistance unit, and an intelligent prompt unit. The voice recognition unit is used to enable the doctor to control the surgical system by voice through voice recognition technology. The visual assistance unit is used to provide real-time intraoperative images and navigation information through augmented reality technology. The intelligent prompt unit is used to give intraoperative suggestions in real time to remind the doctor of key operations or risks.

[0019] Preferably, the surgical data storage and intelligent learning module includes a data storage unit, a data analysis unit, and an AI learning unit. The data storage unit is used to save the images, data, and feedback information during the operation. The data analysis unit is used to analyze the surgical data to identify potential problems and improvement points. The AI learning unit is used to perform AI training using the surgical data to continuously optimize the surgical path planning and navigation model.

[0020] The present invention provides an arthroscopic surgery navigation and positioning system based on artificial intelligence, having the following beneficial effects:

[0021] 1. Through the multi-modal data fusion technology, the present invention combines arthroscopic images, ultrasound, and infrared images to achieve precise visualization of the surgical area, resulting in improved surgical accuracy and safety.

[0022] 2. Through the AI dynamic surgical path planning, the present invention combines preoperative images and real-time data to intelligently calculate the optimal path, realizing real-time adjustment of the surgical path, resulting in significantly improved surgical efficiency and safety.

[0023] 3. Through the robot-assisted operation and automatic adjustment technology, the present invention monitors and automatically adjusts the position and force of the surgical instrument in real time, resulting in reduced operation errors and lower patient risks.

[0024] 4. Through the intraoperative AI voice and visual feedback system, the present invention provides voice control and augmented reality navigation in real time, resulting in improved convenience of doctor operation and accuracy of intraoperative feedback. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 is a system architecture diagram of an arthroscopic surgery navigation and positioning system based on artificial intelligence according to the present invention;

[0026] Figure 2 This is the architecture diagram of the preoperative image processing and lesion analysis module of an arthroscopic surgery navigation and positioning system based on artificial intelligence according to the present invention;

[0027] Figure 3 This is the architecture diagram of the multimodal data fusion module of an arthroscopic surgery navigation and positioning system based on artificial intelligence according to the present invention;

[0028] Figure 4 This is the architecture diagram of the intraoperative real-time joint tracking and pose estimation module of an arthroscopic surgery navigation and positioning system based on artificial intelligence according to the present invention;

[0029] Figure 5 This is the architecture diagram of the AI dynamic surgical path planning module of an arthroscopic surgery navigation and positioning system based on artificial intelligence according to the present invention;

[0030] Figure 6 This is the architecture diagram of the robot-assisted operation and automatic adjustment module of an arthroscopic surgery navigation and positioning system based on artificial intelligence according to the present invention;

[0031] Figure 7 This is the architecture diagram of the intraoperative AI voice and visual feedback module of an arthroscopic surgery navigation and positioning system based on artificial intelligence according to the present invention;

[0032] Figure 8 This is the architecture diagram of the surgical data storage and intelligent learning module of an arthroscopic surgery navigation and positioning system based on artificial intelligence according to the present invention;

[0033] Figure 9 This is the architecture diagram of the relationship of the preoperative image processing and lesion analysis module of an arthroscopic surgery navigation and positioning system based on artificial intelligence according to the present invention. Detailed implementation manners

[0034] Next, in combination with the accompanying drawings of the present invention, the technical solutions of the present invention will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0035] Please refer to the appended Figure 1 - appended Figure 9 , the embodiment of the present invention provides an arthroscopic surgery navigation and positioning system based on artificial intelligence, including:

[0036] The preoperative image processing and lesion analysis module is used to analyze preoperative images and automatically identify joint structures and lesion areas;

[0037] The multimodal data fusion module is used to integrate intraoperative imaging data from different sources and provide a clear surgical view;

[0038] The intraoperative real-time joint tracking and pose estimation module is used to track the joint position and soft tissue deformation in real time;

[0039] The AI dynamic surgical path planning module is used to intelligently optimize the surgical path, prevent damage to important tissues, and dynamically adjust the path;

[0040] The robot-assisted operation and automatic adjustment module is used to automatically adjust the position and angle of surgical instruments;

[0041] The intraoperative AI voice and visual feedback module is used to provide real-time feedback to assist the doctor in adjusting the surgical operation;

[0042] The surgical data storage and intelligent learning module is used to store and analyze surgical data.

[0043] The preoperative image processing and lesion analysis module includes an image segmentation unit, a lesion detection unit, and a 3D reconstruction unit. The image segmentation unit is used to segment the joint structure and lesion area by using a deep learning model. The learning model includes U-Net. The lesion detection unit is used to automatically identify the lesion area by using an object detection network. The object detection network includes FasterR-CNN and YOLO. The 3D reconstruction unit is used to reconstruct the 3D joint model based on the image data to provide a three-dimensional view for surgical planning.

[0044] Specifically, first, the image segmentation unit is responsible for segmenting the preoperative image data, extracting the joint structure and lesion area, and performing pixel-level segmentation through a deep learning model (such as U-Net) to accurately segment the bones, soft tissues, and lesion area of the joint. The U-Net model can effectively process the complex background in medical images through its unique encoding-decoding structure, ensuring the segmentation accuracy and structural integrity. In addition, the image segmentation process will also combine data augmentation techniques, such as rotation, scaling, and flipping, to further improve the generalization ability of the model and adapt to the personalized data of different patients;

[0045] Secondly, the lesion detection unit uses an object detection network (such as Faster R-CNN and YOLO) to automatically identify and locate the lesion area in the image. Faster R-CNN combines the region proposal network (RPN) with the convolutional neural network (CNN) to accurately identify and locate various lesion areas, such as joint cartilage damage and meniscus tear. The YOLO network realizes the real-time recognition of the lesion area quickly and efficiently through a single-step detection method, which is suitable for application in rapid diagnosis. To further improve the lesion detection ability, the system can also integrate attention mechanisms (such as CBAM and SE-Net) to effectively improve the recognition accuracy of small lesions or areas with blurred boundaries.

[0046] Then, the 3D reconstruction unit generates a three-dimensional joint model based on preoperative imaging data to provide a stereoscopic view for surgical planning. It uses voxel reconstruction techniques (such as V-Net, VoxelMorph) or neural radiance field (NeRF) methods to convert two-dimensional imaging data into a three-dimensional joint structure, and enhances the accuracy and details of the model through 3D reconstruction algorithms. During the reconstruction process, the system uses deep learning methods to accurately restore the three-dimensional anatomical structure of the joint, and further refines the three-dimensional model through mesh optimization and super-resolution techniques to ensure high clarity and high fidelity of the three-dimensional view. It can realize the whole process from preoperative imaging data acquisition to precise surgical planning, provide high-precision preoperative data support, provide a scientific basis for the surgeon's surgical decision-making, and significantly improve the safety and accuracy of the surgery.

[0047] The multi-modal data fusion module includes an image fusion unit, an image enhancement unit, and a multi-modal tracking unit. The image fusion unit is used to fuse multi-modal imaging data, and the fusion of multi-modal imaging data includes arthroscopic images, ultrasound, and infrared. The image enhancement unit is used to remove the influences using image enhancement algorithms, and the influences include blood and smoke. The multi-modal tracking unit is used to combine different modal data to real-time track the surgical area to enhance visualization.

[0048] Specifically, first, the image fusion unit is used to fuse multi-modal imaging data to provide a more comprehensive surgical view. It supports the fusion of arthroscopic images, ultrasound images, and infrared images to make up for the limitations of single-modal images. Arthroscopic images can provide a high-resolution view under direct vision, ultrasound images can visualize soft tissue structures, and infrared images can be used for temperature detection to identify areas of inflammation or abnormal blood flow. The fusion process uses deep learning-based feature extraction methods, such as convolutional neural network (CNN) and multi-scale feature fusion (MSF) algorithms, combined with a multi-modal attention mechanism, to achieve efficient fusion of different modal information, improve the contrast and clarity of the images, and enhance the visualization effect of the surgical area.

[0049] Secondly, the image enhancement unit is used to remove intraoperative influencing factors by using image enhancement algorithms to improve the image quality. Common intraoperative interferences include blood occlusion and smoke diffusion, which can affect the surgical field of view and reduce the doctor's visualization ability. Image dehazing algorithms (such as dark channel prior dehazing algorithm, Retinex enhancement algorithm) are used to remove intraoperative smoke. At the same time, adaptive denoising technology based on generative adversarial network (GAN) is used to remove the interference of blood occlusion and other low-contrast areas. Combining dynamic contrast enhancement (Dynamic Contrast Enhancement) and adaptive histogram equalization (AHE) techniques to enhance image details, ensure the clear visibility of the surgical area, and improve the doctor's observation ability in complex surgical environments.

[0050] Then, the multi-modal tracking unit is used to combine different modal data to achieve real-time tracking of the surgical area to enhance intraoperative visualization. Deep learning tracking algorithms (such as Siamese network tracking, optical flow tracking algorithm) are used, combined with arthroscopic, ultrasonic and infrared images, to continuously monitor the changes in the surgical site. During the tracking process, the system will use temporal information for motion compensation to ensure the stability of the images, and adopt a dynamic fusion method based on image registration to accurately align the features of different modal images. In addition, this unit can combine augmented reality (AR) technology to superimpose the tracking information on the intraoperative images to achieve real-time navigation, further improving the doctor's surgical operation accuracy. The multi-modal data fusion module can effectively fuse various image data, improve the clarity and visualization level of intraoperative images, and enhance the surgical navigation ability through real-time tracking technology. The application can significantly improve the doctor's perception ability of the surgical area, improve the accuracy and safety of the surgery, reduce intraoperative uncertainties, and provide important data support for high-risk arthroscopic surgeries.

[0051] The intraoperative real-time joint tracking and pose estimation module includes a bone tracking unit, a soft tissue deformation prediction unit, and a joint positioning and pose estimation unit. The bone tracking unit is used to real-time track the dynamic position of the bone based on graph neural network. The soft tissue deformation prediction unit is used to predict the soft tissue deformation using the LSTM method and correct the navigation system until it is adjusted accurately. The joint positioning and pose estimation unit is used to estimate the precise position and pose of the joint through optical tracking technology.

[0052] Specifically, first, the bone tracking unit is used to track the dynamic position of the bone in real time based on the Graph Neural Network (GNN). By constructing a graphical representation of the bone structure, the relationships between the various joints and bones of the bone are represented as nodes and edges. On this basis, the graph neural network is used for temporal learning to capture the spatio-temporal features of bone movement. Through this method, the system can accurately track the real-time position of the bone in a dynamically changing surgical environment, provide high-precision bone positioning information, and ensure the real-time update and accurate tracking of the bone position during complex surgical procedures;

[0053] Secondly, the soft tissue deformation prediction unit uses a Long Short-Term Memory (LSTM) network to predict the deformation of the soft tissue and dynamically adjusts the surgical navigation system according to the prediction results. By modeling the temporal data of soft tissue deformation and using the LSTM model to capture the long-term dependence relationship of the soft tissue, the deformation of the soft tissue caused by operating force or other factors during the surgical procedure can be accurately predicted. Through this prediction, the navigation system can be adjusted in real time to avoid positioning errors caused by soft tissue deformation, ensure the accuracy of the surgery, and can adaptively adjust the model parameters according to the characteristics of different soft tissues to improve the accuracy of deformation prediction;

[0054] Then, the joint positioning and pose estimation unit uses optical tracking technology to estimate the precise position and pose of the joint. By installing reflective markers on the patient's joint parts and using a high-precision optical tracking system to collect the spatial coordinates of the joint markers in real time. Combining the positioning information of the optical sensor, the system can accurately calculate the position and pose of the joint and dynamically adjust the surgical path and the positioning of the tool. The high real-time performance of the optical tracking technology enables this unit to adapt to the rapid changes of the patient's joint during the surgical procedure, ensure that the precise pose of the joint is continuously monitored, and the intraoperative real-time joint tracking and pose estimation module can achieve precise tracking and prediction of the joint position and pose, provide efficient real-time feedback, and ensure high-precision navigation during the surgical procedure. This module has important clinical significance for complex joint surgeries, can significantly improve the accuracy and safety of the surgery, reduce intraoperative errors, and especially provide more reliable positioning and navigation support in cases where the operation is difficult and the field of view is limited.

[0055] The AI dynamic surgical path planning module includes a path planning unit, a dynamic adjustment unit, and an obstacle avoidance unit. The path planning unit is used to intelligently calculate the optimal surgical path based on preoperative images and real-time data. The dynamic adjustment unit is used to adjust the path planning according to the surgical progress and real-time feedback. The obstacle avoidance unit: detects key structures during the operation in real time to avoid accidental injury to instruments. The key structures include blood vessels and nerves.

[0056] Specifically, first, the path planning unit intelligently calculates the optimal surgical path based on preoperative images and real-time data. This unit uses deep learning algorithms, combines preoperative images (such as CT, MRI) and real-time data (such as arthroscopic images, ultrasound data, etc.), analyzes the surgical area, and automatically plans the best surgical path. The path planning algorithm comprehensively considers the lesion location, anatomical structure, and the movement limitations of surgical tools to ensure efficiency and safety during the operation. This unit can calculate a short path through optimization algorithms (such as A*, Dijkstra algorithms), and further optimize the path through machine learning techniques to improve the accuracy and feasibility of path selection;

[0057] Secondly, the dynamic adjustment unit adjusts the path planning according to the surgical progress and real-time feedback. During the operation, the patient's physical condition may change, and the intraoperative real-time images and sensor data will also be continuously updated. Therefore, the path planning needs to be adjusted in real time. This unit obtains the image data and operation feedback of the surgical area in real time, uses machine learning models to continuously learn the changes in the surgical environment, and automatically adjusts the path planning to cope with various situations that occur during the operation. This dynamic adjustment ensures that the surgical path always remains in the optimal state, thus avoiding the risk of path deviation or improper operation;

[0058] Then, the obstacle avoidance unit is responsible for detecting key structures during the operation in real time, such as blood vessels, nerves, etc., and avoiding accidental injury by instruments. This unit uses high-precision image processing techniques (such as real-time image segmentation and object detection) to identify the key structures in the surgical area, monitors the position and shape of these structures in real time, and ensures that the surgical instruments maintain a safe distance from the key structures. The obstacle avoidance unit combines object detection algorithms in deep learning (such as Faster R-CNN, YOLO, etc.) for accurate identification of key structures, and at the same time combines hardware devices such as force feedback sensors to further enhance the obstacle avoidance ability. In case of a risk situation, the system will issue an alarm in real time or automatically adjust the movement path of the instrument to avoid accidental injury to important structures such as blood vessels and nerves by the instrument. The AI dynamic surgical path planning module can realize the intelligent planning and real-time adjustment of the surgical path, and at the same time avoid accidental injury to key structures by surgical instruments. The application of this module will significantly improve the safety and precision of the surgical process, provide strong support for complex surgeries, reduce intraoperative risks, and improve the success rate of surgeries and the postoperative recovery quality of patients.

[0059] The robot-assisted operation and automatic adjustment module includes an automatic adjustment unit, a force sensing unit, and a feedback and cooperation unit. The automatic adjustment unit is used to automatically adjust the position, angle, and depth of the surgical instrument according to the AI navigation result. The force sensing unit is used to monitor the contact force of the instrument to prevent excessive operation and tissue damage. The feedback and cooperation unit is used to provide real-time feedback and coordinate the operations of the robot and the doctor.

[0060] Specifically, first, the automatic adjustment unit automatically adjusts the position, angle, and depth of the surgical instrument according to the AI navigation result. This unit combines the preoperative images, real-time images, and the results of AI dynamic path planning, and adjusts the movement trajectory of the robotic surgical instrument through a precise control system. The automatic adjustment unit can accurately adjust the movement parameters of the instrument according to the needs of different surgical steps, ensuring that the surgical instrument is always in the optimal position and angle. This unit also dynamically adapts to the changes in the intraoperative situation through real-time calculation, and timely adjusts the operation path, thereby improving the accuracy and safety of the surgery;

[0061] Secondly, the force sensing unit is used to monitor the contact force of the surgical instrument to prevent excessive operation and tissue damage. This unit integrates high-precision force sensors to measure the magnitude of the force when the instrument contacts the tissue in real time. The system can judge whether there is a situation of excessive operating force according to the real-time data feedback by the force sensor, and thus automatically adjust the movement of the instrument when the force is too large to avoid excessive squeezing or damage to the tissue. The force sensing unit can effectively prevent tissue damage caused by human error or excessive force during the operation and ensure the safety of the patient;

[0062] Then, the feedback and collaboration unit is used to provide real-time feedback and coordinate the operations of the robot and the doctor. This unit monitors the robot operation during the surgery in real time and transmits the operation status, progress information, and potential risks to the doctor through a feedback mechanism. Through this system, the doctor can clearly understand the current surgical status and make decisions in a timely manner. At the same time, the feedback and collaboration unit can coordinate the collaboration between the robot and the doctor to ensure that the robot operation and the doctor operation cooperate with each other, and the doctor can perform manual intervention at critical moments to enhance the flexibility and controllability of the surgery. In addition, the feedback unit can also help the doctor quickly obtain key information through voice prompts, visual displays, etc. to ensure the smooth progress of the surgery. The robot-assisted operation and the automatic adjustment module can achieve high-precision surgical operation and automatic adjustment, while providing powerful real-time feedback and collaboration capabilities. This module can significantly improve the accuracy and safety of surgical operations, reduce human errors, and improve the surgical efficiency and the quality of the patient's postoperative recovery.

[0063] The intraoperative AI voice and visual feedback module includes a voice recognition unit, a visual assistance unit, and an intelligent prompt unit. The voice recognition unit is used to enable the doctor to control the surgical system by voice through voice recognition technology. The visual assistance unit is used to provide real-time intraoperative images and navigation information through augmented reality technology. The intelligent prompt unit is used to give intraoperative suggestions in real time to remind the doctor of key operations or risks.

[0064] Specifically, first, the speech recognition unit enables doctors to control the surgical system through voice recognition technology. This unit adopts advanced speech recognition algorithms to accurately recognize doctors' instructions and convert them into operation commands, reducing the dependence on the surgical interface and enabling doctors to focus more on the surgery;

[0065] Secondly, the visual assistance unit provides real-time intraoperative images and navigation information through augmented reality (AR) technology. This unit combines intraoperative images with three-dimensional reconstruction models and overlays navigation information in real time through AR devices to help doctors intuitively understand the surgical area and improve operation accuracy;

[0066] Then, the intelligent prompt unit gives intraoperative suggestions in real time to remind doctors of key operations or risks. Based on real-time data analysis, the intelligent prompt unit identifies potential risks during the surgery and prompts doctors through voice or vision to help adjust the operation strategy, ensure surgical safety. The intraoperative AI voice and visual feedback module improves the intelligence level of the surgery, enhances doctors' real-time feedback and decision-making capabilities, reduces operation errors, and improves the efficiency and safety of the surgery.

[0067] The surgical data storage and intelligent learning module includes a data storage unit, a data analysis unit, and an AI learning unit. The data storage unit is used to save images, data, and feedback information during the surgery. The data analysis unit is used to analyze surgical data, identify potential problems and improvement points. The AI learning unit is used to perform AI training using surgical data to continuously optimize the surgical path planning and navigation model.

[0068] Specifically, first, the data storage unit is used to save images, data, and feedback information during the surgery. This unit can record all relevant data during the surgery in real time, such as intraoperative images, operation parameters, doctors' feedback, etc., and store them in a secure database to ensure the integrity and traceability of the data;

[0069] Secondly, the data analysis unit analyzes surgical data, identifies potential problems and improvement points. Through data mining and statistical analysis techniques, it automatically analyzes abnormal situations during the surgery, evaluates the surgical effect, helps doctors identify potential risks and improvement space in the operation, and provides data support for postoperative summary and improvement;

[0070] Then, the AI learning unit uses surgical data for AI training to continuously optimize the surgical path planning and navigation model. By continuously learning the data accumulated during the surgery, the AI learning unit can continuously optimize the existing path planning algorithm and navigation model, improve its accuracy and intelligence level, and provide more accurate decision-making support for future surgeries. The surgical data storage and intelligent learning module provides comprehensive data support for the surgical process, not only helping doctors analyze and optimize the surgery in real time, but also enhancing the intelligence level of the future surgical path planning and navigation system through continuous AI training.

[0071] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An arthroscopic surgery navigation and positioning system based on artificial intelligence, characterized in that, Including: The preoperative image processing and lesion analysis module is used to analyze preoperative images and automatically identify joint structures and lesion areas; The multimodal data fusion module is used to integrate intraoperative image data from different sources and provide a clear surgical view; The intraoperative real-time joint tracking and pose estimation module is used to track the joint position and soft tissue deformation in real time; The AI dynamic surgical path planning module is used to intelligently optimize the surgical path, prevent damage to important tissues, and dynamically adjust the path; The robot-assisted operation and automatic adjustment module is used to automatically adjust the position and angle of surgical instruments; The intraoperative AI voice and visual feedback module is used to provide real-time feedback to assist the doctor in adjusting the surgical operation; The surgical data storage and intelligent learning module is used to store and analyze surgical data.

2. The arthroscopic surgery navigation and positioning system based on artificial intelligence according to claim 1, wherein: The preoperative image processing and lesion analysis module includes an image segmentation unit, a lesion detection unit, and a 3D reconstruction unit. The image segmentation unit is used to segment joint structures and lesion areas by using a deep learning model, and the learning model includes U-Net. The lesion detection unit is used to automatically identify lesion areas by using an object detection network, and the object detection network includes Faster R-CNN and YOLO. The 3D reconstruction unit is used to reconstruct a 3D joint model based on image data to provide a three-dimensional view for surgical planning.

3. An arthroscopic surgery navigation and positioning system based on artificial intelligence according to claim 1, characterized in that: The multimodal data fusion module includes an image fusion unit, an image enhancement unit, and a multimodal tracking unit. The image fusion unit is used to fuse multimodal image data, and the multimodal image data fusion includes arthroscopic images, ultrasound, and infrared. The image enhancement unit is used to remove the influences by using an image enhancement algorithm, and the influences include blood and smoke. The multimodal tracking unit is used to combine different modal data to track the surgical area in real time to enhance visualization.

4. An arthroscopic surgery navigation and positioning system based on artificial intelligence according to claim 1, characterized in that: The intraoperative real-time joint tracking and pose estimation module includes a bone tracking unit, a soft tissue deformation prediction unit, and a joint positioning and pose estimation unit. The bone tracking unit is used to track the dynamic position of bones in real time based on a graph neural network. The soft tissue deformation prediction unit is used to predict soft tissue deformation by using the LSTM method and correct the navigation system until the adjustment is accurate. The joint positioning and pose estimation unit is used to estimate the precise position and pose of the joint through optical tracking technology.

5. The arthroscopic surgery navigation and positioning system based on artificial intelligence according to claim 1, characterized in that: The AI dynamic surgical path planning module includes a path planning unit, a dynamic adjustment unit, and an obstacle avoidance unit. The path planning unit is used to intelligently calculate the optimal surgical path based on preoperative images and real-time data. The dynamic adjustment unit is used to adjust the path planning according to the surgical progress and real-time feedback. The obstacle avoidance unit: detects key structures during the operation in real time to avoid accidental injury by instruments, and the key structures include blood vessels and nerves.

6. The arthroscopic surgery navigation and positioning system based on artificial intelligence according to claim 1, characterized in that: The robot-assisted operation and automatic adjustment module includes an automatic adjustment unit, a force sensing unit, and a feedback and cooperation unit. The automatic adjustment unit is used to automatically adjust the position, angle, and depth of surgical instruments according to the AI navigation results. The force sensing unit is used to monitor the contact force of the instruments to prevent excessive operation and tissue damage. The feedback and cooperation unit is used to provide real-time feedback and coordinate the operations of the robot and the doctor.

7. An arthroscopic surgery navigation and positioning system based on artificial intelligence according to claim 1, characterized in that: The intraoperative AI voice and visual feedback module includes a voice recognition unit, a visual assistance unit, and an intelligent prompt unit. The voice recognition unit is used to enable doctors to control the surgical system through voice recognition technology. The visual assistance unit is used to provide real-time intraoperative images and navigation information through augmented reality technology. The intelligent prompt unit is used to give intraoperative suggestions in real time, reminding doctors of key operations or risks.

8. An arthroscopic surgery navigation and positioning system based on artificial intelligence according to claim 1, characterized in that: The surgical data storage and intelligent learning module includes a data storage unit, a data analysis unit, and an AI learning unit. The data storage unit is used to save images, data, and feedback information during the surgical process. The data analysis unit is used to analyze surgical data to identify potential problems and improvement points. The AI learning unit is used to perform AI training using surgical data to continuously optimize the surgical path planning and navigation model.

Images