Positioning method and system for rapid registration in transcranial magnetic stimulation navigation

Abstract

The invention relates to the technical field of medical instruments, and discloses a positioning method and system for rapid registration in transcranial magnetic stimulation navigation, and the method comprises the steps: S1, data collection: collecting RGBD image data of the face of a patient through a high-precision 3D camera, and obtaining skull 3D data through magnetic resonance imaging and computed tomography; s2, data processing: processing the RGBD image, and obtaining feature point information of the face of the patient and a head 3D model; s3, establishing a transcranial magnetic stimulation point model, mapping the skull 3D data to the head 3D model, and establishing a patient personalized head model and a magnetic stimulation point model.According to the method, efficient and accurate positioning in the transcranial magnetic stimulation treatment process is achieved by integrating the steps of high-precision data collection, space positioning and modeling, path planning, rapid registration, real-time tracking and the like, and the accuracy of the treatment process is improved. And the treatment effect and the patient experience are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical devices, and in particular to a positioning method and system for rapid registration in transcranial magnetic stimulation navigation. Background Art

[0002] The transcranial magnetic stimulation navigation system (TMS navigation system for short) is a technical tool that combines transcranial magnetic stimulation (TMS) and a brain navigation system to accurately locate and stimulate specific areas of the brain. This technology induces an induced electric field through a time-varying magnetic field, causing the neural tissue in the functional area of ​​the cerebral cortex to produce a circular induced current, thereby triggering a series of physiological and biochemical reactions to achieve the purpose of treatment or research. The TMS navigation system has been widely used in scientific research and clinical fields due to its advantages such as being non-invasive, painless, easy to operate, safe and reliable.

[0003] In the prior art, before transcranial magnetic stimulation, the patient will wear a positioning cap, and the target area will be manually marked directly on the positioning cap. The positioning cap is usually designed according to the EEG electrode arrangement system, and the functional areas are clearly divided. The patient only needs to wear a cap that fits the size of his head circumference and place the coil in the corresponding position for stimulation. There are positioning caps on the market that have pre-divided functional areas, or commonly used treatment sites are integrated into the cap for easy use. However, due to the different sizes of different human brains, the positioning accuracy will be affected. Summary of the invention

[0004] The object of the present invention is to provide a positioning method and system for rapid registration in transcranial magnetic stimulation navigation, so as to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a positioning method for rapid registration in transcranial magnetic stimulation navigation, comprising the following steps:

[0006] S1: Data acquisition, using a high-precision 3D camera to collect RGBD image data of the patient's face, and using magnetic resonance imaging and computer tomography to obtain skull 3D data;

[0007] S2: Data processing, processing RGBD images, obtaining the patient's facial feature point information and head 3D model;

[0008] S3: Establishment of cranial magnetic stimulation point model, mapping the skull 3D data to the head 3D model, and establishing the patient's personalized head model and magnetic stimulation point model;

[0009] S4: Path planning, planning the stimulation sequence of each magnetic stimulation point and the optimal path and movement speed of the manipulator to reach each point, and controlling the manipulator to accurately move the TMS coil to the treatment site located for the first time;

[0010] S5: Path simulation: According to the planned path, the robot is controlled to move the TMS coil to each treatment site in sequence;

[0011] S6: Path optimization, optimize the simulated path to improve the accuracy of the points;

[0012] S7: Fast matching: Use a high-precision 3D camera to take a picture of the patient's actual head posture, and match the head model with the patient's actual head posture picture to determine the spatial position of the actual stimulation site;

[0013] S8: Real-time tracking, using an infrared camera to monitor the position and posture of the patient's head in real time, and transmit the change data to the manipulator control module;

[0014] S9: Transfer learning, which transfers the infrared camera data and the high-precision 3D camera data during fast matching to the head 3D model data;

[0015] S10: Data integration, integrating the transfer learning data with the original head 3D model data to obtain the head 3D model of the new date;

[0016] S11: Data storage, saving the 3D head model of the patient's head and the path planning data of the manipulator to the patient's head model data;

[0017] S12: Data optimization. In subsequent treatments, the last saved data can be called to achieve one-click rapid repeated positioning, improve treatment efficiency, and obtain real-time 3D head model data to maintain data consistency and high accuracy.

[0018] Preferably, during data collection in S1, the patient is placed in an appropriate position to ensure that the camera can capture the entire facial area, the camera is turned on and parameters are adjusted to obtain high-quality image data, the patient remains still, and the collected data is stored on a computer for subsequent processing and analysis.

[0019] Preferably, during S2 data processing, the collected images are cleaned and preprocessed, including denoising, color space conversion, and image enhancement operations, to improve the accuracy of subsequent processing.

[0020] Preferably, in S3, the skull 3D model is spatially registered with the previously established head 3D model to ensure correct correspondence between the two in terms of spatial position and proportion, and the registered skull model is fused with the head model to construct a complete head 3D model containing skull information.

[0021] Preferably, in S4, the stimulation sequence is first determined based on medical and therapeutic needs, and when planning the movement speed, the stimulation time at each stimulation point is planned, as well as the required preparation and adjustment time.

[0022] Preferably, in S5, according to the planned path and speed, the simulation operation is started, and the manipulator gradually moves to each stimulation point along the predetermined trajectory to find out whether there is any deviation or potential collision risk in the simulation process, and make necessary path or speed adjustments until the movement of the manipulator is accurate.

[0023] Preferably, in S6, according to the simulation results, the parameters of the path planning algorithm are adjusted, such as changing the weight of the path search and optimizing the speed curve to reduce the error. In the transfer learning process of S9, various conditions of the patient are captured, including different head postures and expression changes.

[0024] Preferably, in S10, the data obtained by transfer learning is updated to the original head 3D model to reflect the latest patient head status, and version control is established for each updated head 3D model to record the integrated data source, date and relevant personnel information. S12: In data optimization, ensure that all relevant data, including head 3D model, treatment parameters and historical treatment data, are stored in a database or file in a structured manner to facilitate quick retrieval and call.

[0025] A positioning system for rapid registration in transcranial magnetic stimulation navigation comprises a six-axis manipulator control module, a data acquisition module, a data processing module, a data storage module, a head model module and a magnetic stimulation point model module.

[0026] Preferably, the data processing module includes: data integration and data optimization.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] In the transcranial magnetic stimulation navigation of the present invention, the robot automatic navigation system replaces the traditional coil holder with a mechanical arm, uses software to accurately locate the target point, and corrects the coil position in real time during the stimulation process to adapt to the dynamic changes of the patient's head. By integrating high-precision data acquisition, spatial positioning and modeling, path planning, rapid alignment and real-time tracking, etc., efficient and accurate positioning is achieved during the transcranial magnetic stimulation treatment process, and the treatment effect and patient experience are improved. In addition, transfer learning can be used to migrate the recognized user data to new appearance features, so as to quickly adapt to new facial features. In actual applications, when the user undergoes major facial surgery or significant weight changes, transfer learning can be used to quickly adjust the original facial recognition feature model to a model that adapts to the new appearance. This not only avoids the tediousness of remaking the 3D head model, but also maintains the consistency and high accuracy of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1A flowchart of a positioning method for rapid registration in transcranial magnetic stimulation navigation according to the present invention;

[0030] Figure 2 This is a connection relationship diagram of a positioning system for rapid registration in transcranial magnetic stimulation navigation according to the present invention. DETAILED DESCRIPTION

[0031] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0032] Example 1: Please refer to Figure 1 The present invention provides a technical solution: a positioning method for rapid registration in transcranial magnetic stimulation navigation, comprising the following steps:

[0033] S1: Data acquisition, using a high-precision 3D camera to collect RGBD image data of the patient's face, and using magnetic resonance imaging and computer tomography to obtain skull 3D data;

[0034] S2: Data processing, processing RGBD images, obtaining the patient's facial feature point information and head 3D model;

[0035] S3: Establishment of cranial magnetic stimulation point model, mapping the skull 3D data to the head 3D model, and establishing the patient's personalized head model and magnetic stimulation point model;

[0036] S4: Path planning, planning the stimulation sequence of each magnetic stimulation point and the optimal path and movement speed of the manipulator to reach each point, and controlling the manipulator to accurately move the TMS coil to the treatment site located for the first time;

[0037] S5: Path simulation: According to the planned path, the robot is controlled to move the TMS coil to each treatment site in sequence;

[0038] S6: Path optimization, optimize the simulated path to improve the accuracy of the points;

[0039] S7: Fast matching: Use a high-precision 3D camera to take a picture of the patient's actual head posture, and match the head model with the patient's actual head posture picture to determine the spatial position of the actual stimulation site;

[0040] S8: Real-time tracking, using an infrared camera to monitor the position and posture of the patient's head in real time, and transmit the change data to the manipulator control module;

[0041] S9: Transfer learning, which transfers the infrared camera data and the high-precision 3D camera data during fast matching to the head 3D model data;

[0042] S10: Data integration, integrating the transfer learning data with the original head 3D model data to obtain the head 3D model of the new date;

[0043] S11: Data storage, saving the 3D head model of the patient's head and the path planning data of the manipulator to the patient's head model data;

[0044] S12: Data optimization. In subsequent treatments, the last saved data can be called to achieve one-click rapid repeated positioning, improve treatment efficiency, and obtain real-time 3D head model data to maintain data consistency and high accuracy.

[0045] In S1, during data collection, the patient is placed in an appropriate position to ensure that the camera can capture the entire facial area. The camera is turned on and parameters are adjusted to obtain high-quality image data. The patient is kept still. The collected data is stored on a computer for subsequent processing and analysis. Similarly, when collecting skull 3D data, the following are also required: the patient is placed on the examination bed to ensure that the head is located in the center of the scanning area. MRI or CT equipment is selected as needed, and relevant parameters such as scanning range and resolution are adjusted. The patient is kept still and follows the instructions of the doctor or technician to avoid affecting the image quality. The scanning equipment is started for data collection, which usually takes several minutes to more than ten minutes, depending on the performance of the equipment and the amount of data required. The collected data is stored on a computer for subsequent processing and analysis.

[0046] During S2 data processing, the collected images are cleaned and preprocessed, including denoising, color space conversion, and image enhancement operations, to improve the accuracy of subsequent processing. Based on the depth information in the RGBD image, point cloud data of the patient's face is generated. In S3, the skull 3D model is spatially registered with the previously established head 3D model to ensure the correct correspondence between the two in spatial position and proportion. The registered skull model is fused with the head model to construct a complete head 3D model containing skull information. Each point contains its coordinates and color information in three-dimensional space. The point cloud segmentation and clustering algorithms are used to process the point cloud data, remove irrelevant background information, retain the facial point cloud, and reconstruct a detailed 3D model of the patient's head through the point cloud registration and fusion algorithms combined with the detected feature point information.

[0047] In S4, the order of stimulation is first determined based on medical and therapeutic needs. When planning the movement speed, the time for stimulation at each stimulation point, as well as the required preparation and adjustment time, are planned. Based on the knowledge of neuroscience and anatomy, the brain areas or neural structures that need magnetic stimulation are determined, and the accuracy of the model is verified by comparison with actual clinical data. If necessary, professional neurologists can be invited to participate in the evaluation and make necessary adjustments to the model based on feedback to ensure the accuracy and effectiveness of the magnetic stimulation points. The order of stimulation should first be determined based on medical and therapeutic needs. It may be necessary to consider the mutual influence between stimulation points, patient comfort, and maximization of therapeutic effects. On the basis of meeting medical parameters, the stimulation sequence is planned to reduce the movement distance and time of the manipulator on the head model, while avoiding safety risks that may be caused by rapid or sudden movements. In order to ensure smooth and safe movement of the manipulator, a speed curve needs to be designed, including acceleration, constant speed, and deceleration stages, to avoid equipment damage or patient discomfort due to sudden changes.

[0048] In S5, according to the planned path and speed, the simulation operation begins. The manipulator moves step by step to each stimulation point along the predetermined trajectory. If there is any deviation or potential collision risk during the simulation, the necessary path or speed adjustment is made until the manipulator moves accurately. All relevant data during the simulation are recorded, including the actual movement path and speed of the manipulator and the accuracy of reaching each stimulation point. The arrival accuracy of each stimulation point is analyzed to identify possible sources of error, such as the positioning accuracy of the manipulator and the smoothness of motion control.

[0049] In S6, according to the simulation results, adjust the parameters of the path planning algorithm, such as changing the weight of the path search and optimizing the speed curve to reduce the error. If it is found that the hardware limitation is the cause of the error, consider making necessary hardware upgrades or adjustments, such as increasing the degree of freedom of the manipulator and improving the accuracy of the sensor. Apply the optimized settings and re-run the path simulation to verify the effect of the optimization.

[0050] During the S9 transfer learning process, the shots covered various conditions of the patients, including different head postures and facial expression changes.

[0051] In S10, the data obtained by transfer learning is updated to the original head 3D model to reflect the latest patient head status, and version control is established for each updated head 3D model to record the integrated data source, date, and relevant personnel information. S12: In data optimization, ensure that all relevant data, including head 3D models, treatment parameters, and historical treatment data, are stored in a structured manner in the database or file for quick retrieval and call.

[0052] Example 2: Please refer to Figure 2The present invention provides a technical solution: a positioning system for rapid alignment in transcranial magnetic stimulation navigation, including a six-axis manipulator control module, a data acquisition module, a data processing module, a data storage module, a head model module and a magnetic stimulation point model module.

[0053] The data processing module includes: data integration and data optimization.

[0054] In the transcranial magnetic stimulation navigation of the present invention, the robot automatic navigation system replaces the traditional coil holder with a mechanical arm, uses software to accurately locate the target point, and corrects the coil position in real time during the stimulation process to adapt to the dynamic changes of the patient's head. By integrating high-precision data acquisition, spatial positioning and modeling, path planning, rapid alignment and real-time tracking, etc., efficient and accurate positioning is achieved during the transcranial magnetic stimulation treatment process, and the treatment effect and patient experience are improved. In addition, transfer learning can be used to migrate the recognized user data to new appearance features, so as to quickly adapt to new facial features. In actual applications, when the user undergoes major facial surgery or significant weight changes, transfer learning can be used to quickly adjust the original facial recognition feature model to a model that adapts to the new appearance. This not only avoids the tediousness of remaking the 3D head model, but also maintains the consistency and high accuracy of the system.

[0055] Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in the field without making any creative work shall fall within the scope of protection of the present invention.

Claims

1. A positioning method for rapid registration in transcranial magnetic stimulation navigation, characterized in that: The following steps are involved: S1: Data acquisition, using a high-precision 3D camera to collect RGBD image data of the patient's face, and using magnetic resonance imaging and computer tomography to obtain skull 3D data; S2: Data processing, processing RGBD images, obtaining the patient's facial feature point information and head 3D model; S3: Establishment of cranial magnetic stimulation point model, mapping the skull 3D data to the head 3D model, and establishing the patient's personalized head model and magnetic stimulation point model; S4: Path planning, planning the stimulation sequence of each magnetic stimulation point and the optimal path and movement speed of the manipulator to reach each point, and controlling the manipulator to accurately move the TMS coil to the treatment site located for the first time; S5: Path simulation: According to the planned path, the robot is controlled to move the TMS coil to each treatment site in sequence; S6: Path optimization, optimize the simulated path to improve the accuracy of the points; S7: Fast matching: Use a high-precision 3D camera to take a picture of the patient's actual head posture, and match the head model with the patient's actual head posture picture to determine the spatial position of the actual stimulation site; S8: Real-time tracking, using an infrared camera to monitor the position and posture of the patient's head in real time, and transmit the change data to the manipulator control module; S9: Transfer learning, which transfers the infrared camera data and the high-precision 3D camera data during fast matching to the head 3D model data; S10: Data integration, integrating the transfer learning data with the original head 3D model data to obtain the head 3D model of the new date; S11: Data storage, saving the 3D head model of the patient's head and the path planning data of the manipulator to the patient's head model data; S12: Data optimization. In subsequent treatments, the last saved data can be called to achieve one-click rapid repeated positioning, improve treatment efficiency, and obtain real-time 3D head model data to maintain data consistency and high accuracy.

2. A positioning method for rapid registration in transcranial magnetic stimulation navigation according to claim 1, characterized in that: During data collection in S1, the patient was placed in an appropriate position to ensure that the camera could capture the entire facial area, the camera was turned on and parameters were adjusted to obtain high-quality image data, the patient was asked to remain still, and the collected data was stored on a computer for subsequent processing and analysis.

3. A positioning method for rapid registration in transcranial magnetic stimulation navigation according to claim 1, characterized in that: During S2 data processing, the collected images are cleaned and preprocessed, including denoising, color space conversion, and image enhancement operations, to improve the accuracy of subsequent processing.

4. A positioning method for rapid registration in transcranial magnetic stimulation navigation according to claim 1, characterized in that: In S3, the skull 3D model is spatially registered with the previously established head 3D model to ensure the correct correspondence between the two in terms of spatial position and proportion. The registered skull model is fused with the head model to construct a complete head 3D model containing skull information.

5. A positioning method for rapid registration in transcranial magnetic stimulation navigation according to claim 1, characterized in that: In S4, the order of stimulation is first determined based on medical and therapeutic needs, and when planning the movement speed, the time for stimulation at each stimulation point is planned, as well as the required preparation and adjustment time.

6. A positioning method for rapid registration in transcranial magnetic stimulation navigation according to claim 1, characterized in that: In S5, according to the planned path and speed, the simulation operation begins, and the manipulator gradually moves to each stimulation point along the predetermined trajectory to find out whether there is any deviation or potential collision risk during the simulation process, and make necessary path or speed adjustments until the movement of the manipulator is accurate.

7. A positioning method for rapid registration in transcranial magnetic stimulation navigation according to claim 1, characterized in that: In S6, according to the simulation results, the parameters of the path planning algorithm were adjusted, such as changing the weight of the path search and optimizing the speed curve to reduce the error. In the S9 transfer learning process, the shooting covered various conditions of the patients, including different head postures and expression changes.

8. A positioning method for rapid registration in transcranial magnetic stimulation navigation according to claim 1, characterized in that: In S10, the data obtained by transfer learning is updated to the original head 3D model to reflect the latest patient head status, and version control is established for each updated head 3D model to record the integrated data source, date, and relevant personnel information. S12: In data optimization, ensure that all relevant data, including head 3D models, treatment parameters, and historical treatment data, are stored in a structured manner in the database or file to facilitate quick retrieval and call.

9. A positioning system for rapid registration in transcranial magnetic stimulation navigation, characterized in that: It includes a six-axis manipulator control module, a data acquisition module, a data processing module, a data storage module, a head model module and a magnetic stimulation point model module.

10. A positioning system for rapid registration in transcranial magnetic stimulation navigation according to claim 9, characterized in that: The data processing module includes: data integration and data optimization.

Images