A robot safe tool path trajectory planning method and system

Abstract

The present invention discloses a method and system for planning safe toolpath trajectories for a robot. The method comprises: obtaining a medical image set containing a lesion to be operated on and a three-dimensional anatomical virtual model of a target patient; determining a point on the surface of the lesion as a target surgical point set, searching a medical database for a historical set of actual incision points of the same type to be operated on; determining the incision points to be operated on on the surface of the target patient based on the historical set of incision points; generating a predicted toolpath trajectory for the surgical robot in the three-dimensional anatomical virtual model based on the incision points to be operated on and the target set of incision points using a pre-trained artificial intelligence model; and planning a final toolpath trajectory to be operated on based on the predicted toolpath trajectory and the historical set of actual toolpath trajectories of the same type to be operated on. Using the embodiments of the present invention, intelligent toolpath planning can be performed based on the patient's anatomical characteristics in a personalized manner, and previous surgical experience can be effectively absorbed and utilized to ensure the safety and effectiveness of the surgery.

Description

Technical Field

[0001] The present invention belongs to the field of artificial intelligence technology, and in particular to a method and system for planning a safe tool path trajectory of a robot. Background Art

[0002] In modern medicine, robotic surgery has become an increasingly important surgical tool due to its high precision, minimal trauma, and rapid recovery. The introduction of robotic surgical systems allows surgeons to perform more precise procedures, effectively reducing the risks associated with human factors, especially during complex procedures. However, despite the continuous advancement of robotic surgical technology, surgical path planning remains a critical factor in ensuring surgical safety and success.

[0003] Traditional surgical path planning often relies on the surgeon's experience and intuition, lacking a systematic scientific basis. This approach is not only limited by the surgeon's individual skills but also susceptible to factors such as the surgical environment, individual patient differences, and anatomical complexity, leading to increased surgical risks and inconsistent outcomes. Summary of the Invention

[0004] The purpose of the present invention is to provide a robot safe tool path trajectory planning method and system to address the deficiencies in the existing technology. It can perform intelligent tool path planning based on the patient's anatomical characteristics in a personalized manner, and can effectively absorb and utilize previous surgical experience to ensure the safety and effectiveness of the operation.

[0005] An embodiment of the present application provides a method for planning a safe tool path trajectory of a robot, the method comprising:

[0006] Obtaining a medical image set and a three-dimensional anatomical virtual model of a target patient containing the lesion to be operated on;

[0007] Determine the points on the lesion surface as the target surgical point set, and search the medical database for the historical actual entry point set for the same type of surgery.

[0008] Determining a target patient's surface surgical entry point based on the historical entry point set;

[0009] Based on the entry point to be operated on and the target surgical point set, a predicted tool path trajectory of the surgical robot is generated in the three-dimensional anatomical virtual model using a pre-trained artificial intelligence model, wherein the predicted tool path trajectory connects the entry point to be operated on and the target surgical point set;

[0010] The final tool path trajectory to be operated is planned according to the predicted tool path trajectory and a historical actual tool path trajectory set of the same type of operation to be operated.

[0011] Optionally, determining the target patient's surface entry point to be operated on based on the historical entry point set includes:

[0012] From a set of historical lesion locations of the same type to be operated on in a large medical database, find a historical lesion location whose distance from the center of the lesion to be operated on is less than a preset distance;

[0013] Determine the historical incision points corresponding to the positions of the historical lesions, and draw a circle for each of the determined historical incision points so that the circle includes the most historical incision points;

[0014] The center of the circle is calculated as the target patient's surgical entry point.

[0015] Optionally, generating a predicted tool path trajectory of the surgical robot in the three-dimensional anatomical virtual model using a pre-trained artificial intelligence model based on the surgical entry point to be operated on and the target surgical point set includes:

[0016] In the three-dimensional anatomical virtual model, a partial anatomical virtual model data set related to the lesion to be operated on is obtained;

[0017] The entry point to be operated on, the target surgical point set and the partial anatomical virtual model data set are input into a pre-trained artificial intelligence model to generate a predicted tool path trajectory of the surgical robot in the partial anatomical virtual model, wherein the artificial intelligence model is trained based on historical surgical entry points, historical target surgical point sets, historical actual tool path trajectories and corresponding historical partial anatomical virtual model data sets, and the historical actual tool path trajectories are artificial actual tool path trajectories, and the historical partial anatomical virtual model data set is the minimum partial anatomical virtual model that contains the historical actual tool path trajectories.

[0018] Optionally, planning a final tool path trajectory to be operated on based on the predicted tool path trajectory and a set of historical actual tool path trajectories for the same type of operation to be operated on includes:

[0019] From a set of historical actual tool path trajectories of the same type of surgery to be performed, the historical actual tool path trajectories corresponding to the historical lesion positions whose distance from the center of the lesion position to be operated on is less than a preset distance are screened;

[0020] The screened historical actual tool path trajectories are segmented according to the anatomical tissue type, and each segment of the historical actual tool path trajectories is correspondingly superimposed;

[0021] For each segment of the superimposed historical actual tool path trajectory, determine the sub-segment trajectory with an overlap degree higher than a preset threshold;

[0022] Deleting the first part of the predicted tool path trajectory corresponding to each partial sub-segment trajectory in the predicted tool path trajectory to obtain a first predicted tool path trajectory;

[0023] Determine whether there is a non-cutting tissue structure in the predicted tool path trajectory after deletion. If so, delete the second part of the predicted tool path trajectory corresponding to the non-cutting tissue structure to obtain a second predicted tool path trajectory. Otherwise, use the first predicted tool path trajectory directly as the second predicted tool path trajectory.

[0024] The second predicted tool path trajectory is spliced ​​with the determined partial sub-segment trajectories to obtain the final tool path trajectory to be operated on.

[0025] Another embodiment of the present application provides a robot safe tool path trajectory planning system, the system comprising:

[0026] An acquisition module, configured to acquire a medical image set and a three-dimensional anatomical virtual model of a target patient including a lesion to be operated on;

[0027] The search module is used to determine the points on the lesion surface as the target surgical point set and search the medical database for the historical actual entry point set of the same type of surgery to be performed;

[0028] a determination module, configured to determine a target patient's surface entry point for surgery based on the historical entry point set;

[0029] a prediction module, configured to generate a predicted tool path trajectory of the surgical robot in the three-dimensional anatomical virtual model based on the entry point to be operated on and the target surgical point set, using a pre-trained artificial intelligence model, wherein the predicted tool path trajectory connects the entry point to be operated on and the target surgical point set;

[0030] A planning module is used to plan a final tool path for surgery based on the predicted tool path and a set of historical actual tool path trajectories for the same type of surgery to be performed.

[0031] Yet another embodiment of the present application provides a storage medium, wherein the storage medium stores a computer program, wherein the computer program is configured to execute any of the above methods when run.

[0032] Yet another embodiment of the present application provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute any of the above methods.

[0033] Compared with the existing technology, the present invention provides a robot safe tool path trajectory planning method, which obtains a medical image set containing the lesion to be operated on and a three-dimensional anatomical virtual model of the target patient; determines the points on the surface of the lesion as the target surgical point set, and searches for a historical actual entry point set of the same type of surgery to be performed from a medical big data database; determines the entry point to be operated on the surface of the target patient based on the historical entry point set; generates a predicted tool path trajectory of the surgical robot in the three-dimensional anatomical virtual model based on the entry point to be operated and the target surgical point set using a pre-trained artificial intelligence model; plans the final tool path trajectory to be operated based on the predicted tool path trajectory and the historical actual tool path trajectory set of the same type of surgery to be performed, thereby enabling personalized intelligent tool path planning based on the patient's anatomical characteristics, and effectively absorbing and utilizing previous surgical experience to ensure the safety and effectiveness of the operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 A hardware structure block diagram of a computer terminal for a robot safety tool path trajectory planning method provided by an embodiment of the present invention;

[0035] Figure 2 A schematic diagram of a flow chart of a robot safe tool path trajectory planning method provided by an embodiment of the present invention;

[0036] Figure 3 A schematic structural diagram of a robot safe tool path trajectory planning system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0037] The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and are not to be construed as limiting the present invention.

[0038] The embodiment of the present invention first provides a robot safe tool path trajectory planning method, which can be applied to electronic devices such as computer terminals, specifically ordinary computers, etc.

[0039] The following describes it in detail by taking running on a computer terminal as an example. Figure 1 The hardware structure block diagram of a computer terminal for a robot safety tool path trajectory planning method provided by an embodiment of the present invention. Figure 1 As shown, the computer terminal may include one or more ( Figure 1 Only one is shown) a processor 102 (the processor 102 may include but is not limited to a microprocessor MCU or a programmable logic device FPGA and other processing devices) and a memory 104 for storing data. Optionally, the computer terminal may also include a transmission device 106 for communication functions and an input and output device 108. It will be understood by those skilled in the art that Figure 1The structure shown is only for illustration and does not limit the structure of the above-mentioned computer terminal. For example, the computer terminal may also include Figure 1 More or fewer components than shown, or with Figure 1 Different configurations shown.

[0040] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / modules corresponding to the robot safety tool path trajectory planning method in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, implementing the above-mentioned method. The memory 104 may include a high-speed random access memory and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely located relative to the processor 102, and these remote memories may be connected to the computer terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0041] The transmission device 106 is used to receive or send data via a network. Specific examples of the aforementioned network may include a wireless network provided by a communications provider of a computer terminal. In one embodiment, the transmission device 106 includes a network interface controller (NIC), which can be connected to other network devices via a base station to enable communication with the Internet. In another embodiment, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0042] See also Figure 2 , an embodiment of the present invention provides a robot safe tool path trajectory planning method, which may include the following steps:

[0043] S201, obtaining a medical image set and a three-dimensional anatomical virtual model of a target patient including a lesion to be operated on;

[0044] This step first requires obtaining the target patient's relevant medical imaging data and constructing a 3D anatomical virtual model based on this imaging data. Key to this process is an accurate understanding of the patient's internal anatomy, particularly the specific location of the surgical lesion and the identification of surrounding anatomical structures. Medical imaging datasets typically include various types of images, such as CT (computed tomography) and MRI (magnetic resonance imaging), which provide crucial visual and spatial information for subsequent surgical planning.

[0045] The primary purpose of acquiring medical image datasets and 3D anatomical virtual models is to provide a precise anatomical context for surgical path planning. Clear image data enables surgeons to accurately identify the lesion to be operated on and clarify the relationships between surrounding tissues, enabling safer and more effective surgical strategies. This process not only enhances surgical visualization but also provides foundational data support for subsequent AI model training, further enhancing the accuracy of path planning.

[0046] Specifically, image data can be acquired from multimodal imaging devices (CT, MRI, etc.) of the target patient and image registration can be performed using specialized image processing software. This image processing technology ensures that images from different sources are precisely aligned in space, facilitating subsequent information fusion.

[0047] Deep learning segmentation algorithms (such as U-Net) can be applied to image data to identify the surgical lesion and its surrounding anatomical structures. The algorithm classifies the images at the pixel level, ultimately obtaining the outlines of the lesion area and its adjacent important anatomical structures, ensuring segmentation accuracy and consistency.

[0048] The lesion and anatomical structure information obtained in the above steps is volumetrically reconstructed, and a three-dimensional anatomical virtual model is generated using stereo modeling techniques (such as the Marching Cubes algorithm). This model should accurately reflect the patient's anatomical characteristics, including the shape, size, and location of the lesion, and incorporate the spatial relationships of adjacent tissues and organs to provide visualization support for surgery.

[0049] Data optimization of the 3D model is performed, using mesh simplification and refinement techniques to improve computational efficiency and visual quality. Excessive noise points and unnecessary details are removed to ensure the accuracy and operability of the final model. Verification is performed on the generated 3D anatomical virtual model, incorporating the surgeon's professional experience to ensure the model's accuracy and authenticity. Virtual reality technology can be used to allow the surgeon to interact with the model, further confirming the relationship between the lesion location and surrounding tissue, ensuring the reliability of subsequent tool path planning.

[0050] The integrated medical image collection and 3D anatomical virtual models are stored in a cloud database for easy access and sharing. This process also provides the necessary data foundation for subsequent artificial intelligence model training, tool path planning, and surgical decision-making.

[0051] Through the above methods, a medical image set and a 3D anatomical virtual model of the target patient, including the surgical lesion, can be successfully acquired. This process not only ensures the accuracy and completeness of the data but also lays a solid foundation for safe surgical path planning. By combining advanced image processing technology with deep learning algorithms, the intelligence and personalization of robotic surgery are significantly improved, ultimately promoting the precision and safety of surgical procedures.

[0052] S202, determining points on the lesion surface as a target surgical point set, and searching a large medical database for a historical set of actual surgical entry points for the same type of surgery to be performed;

[0053] In this step, the patient's 3D anatomical virtual model is first analyzed to identify key points on the lesion's surface. These points will be considered the target surgical point set. The target surgical point is typically the ideal location on the lesion's surface for surgical intervention, ensuring the surgeon can effectively access and treat the lesion. Simultaneously, a set of historical entry points from past surgeries of the same type must be searched in a large medical database. These points—the points selected by surgeons for entering the patient's body during similar procedures—help form a reference for surgical planning.

[0054] The process of identifying the target surgical site and tracing historical entry points is crucial to the success of surgical procedures. Determining the surface location of the lesion provides direct operational guidance, while analyzing historical entry points accumulates valuable experience for finding the optimal approach. This approach not only improves surgical safety and minimizes damage to surrounding tissue, but also allows surgeons to optimize surgical plans using historical data to ensure efficient and successful procedures.

[0055] Specifically, 3D reconstruction technology can be used to construct a clear 3D anatomical model from acquired medical imaging data. Image processing algorithms, such as deep learning models based on graph convolutional networks (GCNs), are then used to segment the lesion area and accurately identify feature points on the lesion surface. These feature points serve as preliminary target surgical points.

[0056] The extracted feature points on the lesion surface (i.e., the initial target surgical point) are converted into point cloud data. Clustering algorithms (such as DBSCAN) can be used to process the point cloud data, extracting meaningful feature points and filtering out noise and irrelevant points. Normal calculation and curvature analysis can also be used to assess the geometric importance of each point, ensuring that the final target surgical point has optimal contact conditions to achieve the surgical goal.

[0057] It can access a large medical database and, through a content-based retrieval system, extract the actual entry point related to the current lesion type and location from historical surgical records. By dynamically annotating historical data, it ensures that it can accurately match the current surgical type.

[0058] S203, determining a target surgical entry point on the surface of the target patient based on the historical entry point set;

[0059] Specifically, a historical lesion location set of the same type of surgery to be performed can be searched from a large medical database, and a historical lesion location whose distance from the center of the surgery lesion location is less than a preset distance;

[0060] This process aims to identify lesion locations similar to those of the target patient's surgical lesion by analyzing historical case histories. Medical databases contain extensive historical case data, including the spatial locations of lesions and corresponding entry points. By calculating distances, we can screen for historical lesion locations with similar characteristics to the current lesion, ensuring the reference value of the selected entry point. This data leverages existing case data to provide empirical support from medical practice. By comparing similar lesions, we can better predict the potential entry point for surgery at the current lesion location, reducing surgical risks and improving surgical success rates.

[0061] Determine the historical incision points corresponding to the positions of the historical lesions, and draw a circle for each of the determined historical incision points so that the circle includes the most historical incision points;

[0062] Once the locations of historical lesions similar to the target lesion have been identified, the next step is to identify the historical entry points for these lesions. Using a "circle drawing" method, these entry points are clustered together to form a circular area encompassing as many entry points as possible. The goal is to find a reasonable area that can cover multiple entry points. This method, by clustering similar entry points, increases flexibility and rationality in entry point selection. This ensures that the selected entry points are statistically reliable, reflect historical surgical successes, and reduce errors and risks.

[0063] The center of the circle is calculated as the target patient's surgical entry point.

[0064] The final step is to calculate the center of the circle, which becomes the selected surgical entry point. This spatially significant point serves as the starting point for surgical entry, allowing access to the lesion. By calculating the center of the circle, the entry point, derived from historical data analysis, is rationalized into a specific location, providing surgeons with clear guidance when planning the surgery. This data-driven approach can effectively improve surgical precision, reduce unnecessary complexity, and provide patients with a safer surgical experience.

[0065] The above steps systematically determine the target surgical entry point for the target patient by combining historical case data with geometric analysis. This process not only relies on the statistical properties of historical data but also incorporates scientific computational methods, providing a solid foundation for subsequent surgical planning. This approach can more effectively ensure the safety and effectiveness of the surgery.

[0066] S204, generating a predicted tool path trajectory of the surgical robot in the three-dimensional anatomical virtual model based on the pending surgical entry point and the target surgical point set using a pre-trained artificial intelligence model, wherein the predicted tool path trajectory connects the pending surgical entry point and the target surgical point set;

[0067] This step aims to improve the surgical robot's precision while minimizing damage to normal tissue through the application of machine learning and intelligent algorithms. The resulting predicted toolpath provides surgeons with intuitive surgical path planning, enabling them to better control the robot during surgery and optimize surgical outcomes. This approach can increase surgical safety, improve postoperative recovery rates, and reduce the risk of postoperative complications for patients.

[0068] Specifically, a partial anatomical virtual model data set related to the lesion to be operated on can be obtained in the three-dimensional anatomical virtual model;

[0069] This process involves extracting a subset of datasets relevant to the lesion being operated on from the overall 3D anatomical virtual model. This dataset includes information on key anatomical structures surrounding the lesion, such as blood vessels, nerves, and other vital tissues. Obtaining relevant partial anatomical model data helps the AI ​​model more accurately understand the anatomical environment surrounding the lesion, reducing potential surgical risks. This ensures that the surgical path does not conflict with critical anatomical structures, thereby improving the safety and effectiveness of the procedure.

[0070] A graphics-based region growing algorithm can be used to automatically extract data related to the lesion to be operated on from the overall 3D anatomical model. The algorithm first places a seed point at the location of the lesion to be operated on and then expands to adjacent voxels based on similarity thresholds (such as color, density, texture, etc.). Through continuous growth, a region is eventually formed that encompasses the lesion and its surrounding anatomical structures, ensuring that the extracted dataset contains all key anatomical features that may affect the surgical path.

[0071] The entry point to be operated on, the target surgical point set and the partial anatomical virtual model data set are input into a pre-trained artificial intelligence model to generate a predicted tool path trajectory of the surgical robot in the partial anatomical virtual model, wherein the artificial intelligence model is trained based on historical surgical entry points, historical target surgical point sets, historical actual tool path trajectories and corresponding historical partial anatomical virtual model data sets, and the historical actual tool path trajectories are artificial actual tool path trajectories, and the historical partial anatomical virtual model data set is the minimum partial anatomical virtual model that contains the historical actual tool path trajectories.

[0072] In this step, the planned surgical entry point, the target surgical point set, and a partial anatomical virtual model dataset are input into a pre-trained AI model. This model learns from historical data to predict the tool path. By analyzing the input data, the model calculates an optimal tool path. This data-driven approach provides the optimal tool path for robotic surgery. The model is trained based on historical surgical experience, effectively reducing the risk of unexpected procedures and significantly improving surgical accuracy and repeatability. This AI-based trajectory prediction helps surgeons achieve higher surgical standards and optimize overall surgical outcomes.

[0073] A graph neural network (GNN) can be used to process the extracted anatomical model dataset. By constructing a graph structure, the entry point, target surgical point, and anatomical feature nodes are connected. A graph convolutional layer is used to extract feature information from the graph nodes, capturing their spatial relationships and interactions. Using a head-level attention mechanism, the model can better focus on toolpath patterns that have demonstrated high success rates in historical data, generating corresponding predicted toolpath trajectories.

[0074] S205 , planning a final tool path trajectory to be operated on according to the predicted tool path trajectory and a set of historical actual tool path trajectories of the same type of operation to be operated on.

[0075] The purpose of this step is to compare and integrate the generated predicted surgical path with historical data to ensure that the final surgical path is both consistent with historical success cases and adaptable to the specific circumstances of the current patient. By analyzing the historical actual surgical path set, it is possible to identify which paths have been shown to be safer and more effective in similar cases. This process typically involves screening, segmenting, and overlaying historical data, as well as optimizing the path based on anatomical features to find the optimal surgical implementation plan.

[0076] By comparing and integrating the predicted tool path with historically accurate tool paths, this step effectively reduces risks during surgery and improves the robustness and reliability of the system. This ensures that the final tool path not only takes into account the patient's specific lesion condition but also draws on the experience of successful previous surgeries, thereby improving surgical safety and success rates. This comprehensive consideration helps surgeons make more informed decisions during surgery and optimizes the patient's surgical experience.

[0077] Specifically, the historical actual tool path trajectory corresponding to the historical lesion position whose distance from the center of the lesion position to be operated on is less than a preset distance may be screened from the historical actual tool path trajectory set of the same type of surgery to be operated on;

[0078] This step aims to identify actual surgical paths corresponding to historical lesions located near the intended lesion through distance screening. By setting a distance threshold, surgical paths performed under similar anatomical circumstances can be effectively selected for reference. This process makes the selected historical paths more relevant, helping to improve the scientific nature of surgical path selection while avoiding the risks associated with selecting irrelevant historical data.

[0079] The screened historical actual tool path trajectories are segmented according to the anatomical tissue type, and each segment of the historical actual tool path trajectories is correspondingly superimposed;

[0080] In this step, the historical toolpath is segmented to better analyze and address the specific anatomical features of each segment. Overlay processing highlights the overlap of different trajectories in the same anatomical region, facilitating subsequent analysis. This allows identification of critical tissue structures and potential risk areas, enabling better optimization of surgical paths.

[0081] For example, 10 historical actual tool path trajectories are screened out, and each historical actual tool path trajectory is divided into 5 segments. Then, the first segment of the 10 historical actual tool path trajectories is superimposed, and the second segment of the 10 historical actual tool path trajectories is superimposed, until the fifth segment of the 10 historical actual tool path trajectories is superimposed.

[0082] For each segment of the superimposed historical actual tool path trajectory, determine the sub-segment trajectory with an overlap degree higher than a preset threshold;

[0083] In this step, by setting an overlap threshold, we can identify tool paths that show high consistency in historical surgeries. This process helps determine which tool paths are more reliable choices. By focusing on areas with high overlap, we can effectively eliminate unreliable tool paths, thereby increasing the certainty of the optimal tool path.

[0084] Deleting the first part of the predicted tool path trajectory corresponding to each partial sub-segment trajectory in the predicted tool path trajectory to obtain a first predicted tool path trajectory;

[0085] The corresponding first part of the predicted trajectory is deleted in order to replace the sub-trajectories with higher overlap in the future, thereby optimizing the safety of the tool path and ensuring that the surgical path remains consistent with the known successful cases in the highly overlapping parts.

[0086] Determine whether there is a non-cutting tissue structure in the predicted tool path trajectory after deletion. If so, delete the second part of the predicted tool path trajectory corresponding to the non-cutting tissue structure to obtain a second predicted tool path trajectory. Otherwise, use the first predicted tool path trajectory directly as the second predicted tool path trajectory.

[0087] This step ensures that the final tool path matches the target anatomical structure and avoids non-cutting tissue (critical tissue structures that must be avoided) to improve the safety of the surgical process. Through this judgment, it can effectively prevent cutting the wrong area during surgery and protect the patient's normal tissue structure.

[0088] The second predicted tool path trajectory is spliced ​​with the determined partial sub-segment trajectories to obtain the final tool path trajectory to be operated on.

[0089] The final step is to combine the optimized tool path with each sub-segment trajectory to form a complete surgical tool path. This ensures that the tool path is not only safe and effective, but also highly consistent with historical surgical experience and actual needs, ensuring that the final tool path can achieve the surgical purpose while maintaining safety and improving surgical operability.

[0090] For example, a path smoothing algorithm, such as B-spline interpolation, can be used to consider potential smoothness and continuity during the splicing process to avoid unnecessary turns in the tool path. This approach ensures the smoothness and naturalness of the final tool path by controlling the position of the splicing points, thereby improving the operational precision and stability of the robotic scalpel.

[0091] It can be seen that a medical image set containing the lesion to be operated on and a three-dimensional anatomical virtual model of the target patient are obtained; points on the surface of the lesion are determined as the target surgical point set, and a historical actual entry point set of the same type of surgery to be operated on is searched from the medical big data database; based on the historical entry point set, the entry point to be operated on the surface of the target patient is determined; based on the entry point to be operated on and the target surgical point set, a pre-trained artificial intelligence model is used to generate a predicted tool path trajectory of the surgical robot in the three-dimensional anatomical virtual model; based on the predicted tool path trajectory and the historical actual tool path trajectory set of the same type of surgery to be operated on, the final tool path trajectory to be operated is planned, so that personalized intelligent tool path planning can be performed according to the patient's anatomical characteristics, and past surgical experience can be effectively absorbed and utilized to ensure the safety and effectiveness of the operation.

[0092] Another embodiment of the present invention provides a robot safety tool path trajectory planning system, see Figure 3 , the system may include:

[0093] An acquisition module 301 is used to acquire a medical image set and a three-dimensional anatomical virtual model of a target patient including a lesion to be operated on;

[0094] Search module 302 is used to determine the points on the surface of the lesion as the target surgical point set, and search the medical database for the historical actual entry point set of the same type of surgery to be performed;

[0095] A determination module 303 is configured to determine a target surgical entry point on the surface of the target patient based on the historical entry point set;

[0096] A prediction module 304 is configured to generate a predicted tool path trajectory of the surgical robot in the three-dimensional anatomical virtual model based on the entry point to be operated on and the target surgical point set using a pre-trained artificial intelligence model, wherein the predicted tool path trajectory connects the entry point to be operated on and the target surgical point set;

[0097] The planning module 305 is configured to plan a final tool path for the surgery to be performed based on the predicted tool path and a set of historical actual tool path trajectories for the same type of surgery to be performed.

[0098] It can be seen that a medical image set containing the lesion to be operated on and a three-dimensional anatomical virtual model of the target patient are obtained; points on the surface of the lesion are determined as the target surgical point set, and a historical actual entry point set of the same type of surgery to be operated on is searched from the medical big data database; based on the historical entry point set, the entry point to be operated on the surface of the target patient is determined; based on the entry point to be operated on and the target surgical point set, a pre-trained artificial intelligence model is used to generate a predicted tool path trajectory of the surgical robot in the three-dimensional anatomical virtual model; based on the predicted tool path trajectory and the historical actual tool path trajectory set of the same type of surgery to be operated on, the final tool path trajectory to be operated is planned, so that personalized intelligent tool path planning can be performed according to the patient's anatomical characteristics, and past surgical experience can be effectively absorbed and utilized to ensure the safety and effectiveness of the operation.

[0099] An embodiment of the present invention further provides a storage medium storing a computer program, wherein the computer program is configured to execute the steps of any one of the above method embodiments when running.

[0100] Specifically, in this embodiment, the above-mentioned storage medium may be configured to store a computer program for performing the following steps:

[0101] S201, obtaining a medical image set and a three-dimensional anatomical virtual model of a target patient including a lesion to be operated on;

[0102] S202, determining points on the lesion surface as a target surgical point set, and searching a large medical database for a historical set of actual surgical entry points for the same type of surgery to be performed;

[0103] S203, determining a target surgical entry point on the surface of the target patient based on the historical entry point set;

[0104] S204, generating a predicted tool path trajectory of the surgical robot in the three-dimensional anatomical virtual model based on the pending surgical entry point and the target surgical point set using a pre-trained artificial intelligence model, wherein the predicted tool path trajectory connects the pending surgical entry point and the target surgical point set;

[0105] S205 , planning a final tool path trajectory to be operated on according to the predicted tool path trajectory and a set of historical actual tool path trajectories of the same type of operation to be operated on.

[0106] It can be seen that a medical image set containing the lesion to be operated on and a three-dimensional anatomical virtual model of the target patient are obtained; points on the surface of the lesion are determined as the target surgical point set, and a historical actual entry point set of the same type of surgery to be operated on is searched from the medical big data database; based on the historical entry point set, the entry point to be operated on the surface of the target patient is determined; based on the entry point to be operated on and the target surgical point set, a pre-trained artificial intelligence model is used to generate a predicted tool path trajectory of the surgical robot in the three-dimensional anatomical virtual model; based on the predicted tool path trajectory and the historical actual tool path trajectory set of the same type of surgery to be operated on, the final tool path trajectory to be operated is planned, so that personalized intelligent tool path planning can be performed according to the patient's anatomical characteristics, and past surgical experience can be effectively absorbed and utilized to ensure the safety and effectiveness of the operation.

[0107] An embodiment of the present invention further provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to perform the steps in any one of the above method embodiments.

[0108] Specifically, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0109] Specifically, in this embodiment, the processor may be configured to execute the following steps through a computer program:

[0110] S201, obtaining a medical image set and a three-dimensional anatomical virtual model of a target patient including a lesion to be operated on;

[0111] S202, determining points on the lesion surface as a target surgical point set, and searching a large medical database for a historical set of actual surgical entry points for the same type of surgery to be performed;

[0112] S203, determining a target surgical entry point on the surface of the target patient based on the historical entry point set;

[0113] S204, generating a predicted tool path trajectory of the surgical robot in the three-dimensional anatomical virtual model based on the pending surgical entry point and the target surgical point set using a pre-trained artificial intelligence model, wherein the predicted tool path trajectory connects the pending surgical entry point and the target surgical point set;

[0114] S205 , planning a final tool path trajectory to be operated on according to the predicted tool path trajectory and a set of historical actual tool path trajectories of the same type of operation to be operated on.

[0115] It can be seen that a medical image set containing the lesion to be operated on and a three-dimensional anatomical virtual model of the target patient are obtained; points on the surface of the lesion are determined as the target surgical point set, and a historical actual entry point set of the same type of surgery to be operated on is searched from the medical big data database; based on the historical entry point set, the entry point to be operated on the surface of the target patient is determined; based on the entry point to be operated on and the target surgical point set, a pre-trained artificial intelligence model is used to generate a predicted tool path trajectory of the surgical robot in the three-dimensional anatomical virtual model; based on the predicted tool path trajectory and the historical actual tool path trajectory set of the same type of surgery to be operated on, the final tool path trajectory to be operated is planned, so that personalized intelligent tool path planning can be performed according to the patient's anatomical characteristics, and past surgical experience can be effectively absorbed and utilized to ensure the safety and effectiveness of the operation.

[0116] The above describes in detail the structure, features and effects of the present invention based on the embodiments shown in the drawings. The above is only a preferred embodiment of the present invention, but the scope of implementation of the present invention is not limited to what is shown in the drawings. Any changes made in accordance with the concept of the present invention, or modifications to equivalent embodiments with equivalent changes, which do not exceed the spirit covered by the description and drawings, should be within the scope of protection of the present invention.

Claims

1. A robot safe tool path trajectory planning method, characterized in that: The method comprises: Obtaining a medical image set and a three-dimensional anatomical virtual model of a target patient containing the lesion to be operated on; Determine the points on the lesion surface as the target surgical point set, and search the medical database for the historical actual entry point set for the same type of surgery. Determining a target patient's surface surgical entry point based on the historical actual entry point set; Based on the entry point to be operated on and the target surgical point set, a predicted tool path trajectory of the surgical robot is generated in the three-dimensional anatomical virtual model using a pre-trained artificial intelligence model, wherein the predicted tool path trajectory connects the entry point to be operated on and the target surgical point set; Planning a final tool path for surgery based on the predicted tool path and a set of historical actual tool path trajectories for the same type of surgery to be performed, wherein, from the set of historical actual tool path trajectories for the same type of surgery to be performed, screening the historical actual tool path trajectories corresponding to the historical lesion location whose distance from the center of the lesion location to be operated on is less than a preset distance; segmenting each of the screened historical actual tool path trajectories according to anatomical tissue type, and performing corresponding superposition processing on each segment of each historical actual tool path trajectory; For each superimposed historical actual tool path trajectory, determining a partial sub-segment trajectory having an overlap degree higher than a preset threshold; deleting the first partial predicted trajectory corresponding to each partial sub-segment trajectory in the predicted tool path trajectory to obtain a first predicted tool path trajectory; Determine whether there is a non-cutting tissue structure in the predicted tool path trajectory after deletion. If so, delete the second part of the predicted trajectory corresponding to the non-cutting tissue structure to obtain the second predicted tool path trajectory. Otherwise, the first predicted tool path trajectory is directly used as the second predicted tool path trajectory; the second predicted tool path trajectory is spliced ​​with the determined partial sub-segment trajectories to obtain the final tool path trajectory to be operated on.

2. The method according to claim 1, characterized in that Determining the target patient's surface surgical entry point based on the historical actual surgical entry point set includes: From a set of historical lesion locations of the same type to be operated on in a large medical database, find a historical lesion location whose distance from the center of the lesion to be operated on is less than a preset distance; Determine the historical incision points corresponding to the positions of the historical lesions, and draw a circle for each of the determined historical incision points so that the circle includes the most historical incision points; The center of the circle is calculated as the target patient's surgical entry point.

3. The method according to claim 2, characterized in that The method of generating a predicted tool path of the surgical robot in the three-dimensional anatomical virtual model based on the surgical entry point and the target surgical point set using a pre-trained artificial intelligence model includes: In the three-dimensional anatomical virtual model, a partial anatomical virtual model data set related to the lesion to be operated on is obtained; The entry point to be operated on, the target surgical point set and the partial anatomical virtual model data set are input into a pre-trained artificial intelligence model to generate a predicted tool path trajectory of the surgical robot in the partial anatomical virtual model, wherein the artificial intelligence model is trained based on historical surgical entry points, historical target surgical point sets, historical actual tool path trajectories and corresponding historical partial anatomical virtual model data sets, and the historical actual tool path trajectories are artificial actual tool path trajectories, and the historical partial anatomical virtual model data set is the minimum partial anatomical virtual model that contains the historical actual tool path trajectories.

4. A robot safe tool path trajectory planning system, characterized in that: The system comprises: An acquisition module, configured to acquire a medical image set and a three-dimensional anatomical virtual model of a target patient including a lesion to be operated on; The search module is used to determine the points on the lesion surface as the target surgical point set and search the medical database for the historical actual entry point set of the same type of surgery to be performed; a determination module, configured to determine a target patient's surface entry point for surgery based on the historical actual entry point set; a prediction module, configured to generate a predicted tool path trajectory of the surgical robot in the three-dimensional anatomical virtual model based on the entry point to be operated on and the target surgical point set, using a pre-trained artificial intelligence model, wherein the predicted tool path trajectory connects the entry point to be operated on and the target surgical point set; a planning module for planning a final tool path trajectory to be operated on based on the predicted tool path trajectory and a set of historical actual tool path trajectories for the same type of operation to be operated on, wherein, from the set of historical actual tool path trajectories for the same type of operation to be operated on, historical actual tool path trajectories corresponding to historical lesion locations whose distance from the center of the lesion location to be operated on is less than a preset distance are screened; each screened historical actual tool path trajectories is segmented according to anatomical tissue type, and each segment of each historical actual tool path trajectory is correspondingly superimposed; For each superimposed historical actual tool path trajectory, determining a partial sub-segment trajectory having an overlap degree higher than a preset threshold; deleting the first partial predicted trajectory corresponding to each partial sub-segment trajectory in the predicted tool path trajectory to obtain a first predicted tool path trajectory; Determine whether there is a non-cutting tissue structure in the predicted tool path trajectory after deletion. If so, delete the second part of the predicted trajectory corresponding to the non-cutting tissue structure to obtain the second predicted tool path trajectory. Otherwise, the first predicted tool path trajectory is directly used as the second predicted tool path trajectory; the second predicted tool path trajectory is spliced ​​with the determined partial sub-segment trajectories to obtain the final tool path trajectory to be operated on.

5. The system according to claim 4, characterized in that The determining module is specifically configured to: From a set of historical lesion locations of the same type to be operated on in a large medical database, find a historical lesion location whose distance from the center of the lesion to be operated on is less than a preset distance; Determine the historical incision points corresponding to the positions of the historical lesions, and draw a circle for each of the determined historical incision points so that the circle includes the most historical incision points; The center of the circle is calculated as the target patient's surgical entry point.

6. The system according to claim 5, characterized in that The prediction module is specifically used to: In the three-dimensional anatomical virtual model, a partial anatomical virtual model data set related to the lesion to be operated on is obtained; The entry point to be operated on, the target surgical point set and the partial anatomical virtual model data set are input into a pre-trained artificial intelligence model to generate a predicted tool path trajectory of the surgical robot in the partial anatomical virtual model, wherein the artificial intelligence model is trained based on historical surgical entry points, historical target surgical point sets, historical actual tool path trajectories and corresponding historical partial anatomical virtual model data sets, and the historical actual tool path trajectories are artificial actual tool path trajectories, and the historical partial anatomical virtual model data set is the minimum partial anatomical virtual model that contains the historical actual tool path trajectories.

7. A storage medium, characterized in that: The storage medium stores a computer program, wherein the computer program is configured to execute the method according to any one of claims 1 to 3 when run.

8. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to run the computer program to perform the method according to any one of claims 1 to 3.

Images