Method for rapid positioning of brain stimulation target based on facial recognition and ai neural navigation

Abstract

The present application relates to a kind of brain stimulation target point fast positioning method based on facial recognition and AI neural navigation, by obtaining the MRI data of user's head, construct the three-dimensional digital brain model including brain function partition information;Acquisition facial three-dimensional point cloud, with the head surface model reconstructed by MRI registration, establish the space mapping of actual head and brain model;Based on this mapping, determine stimulation target point on model.In treatment, real-time tracking head pose using optical system, and in navigation interface fusion display brain model, target point and the real-time position and angle of magnetic stimulation coil.Determine the best target stimulation angle of coil according to the gyri and sulci orientation of target point, guide operator to adjust coil pose by comparing the difference between real-time angle and target angle.The present application realizes the fast, accurate, visual positioning and dynamic maintenance of target point in transcranial magnetic stimulation treatment, and upgrades the traditional experience-dependent operation to individualized, real-time feedback precision navigation treatment.

Description

Technical Field

[0001] This invention relates to the fields of medical instruments and neuronavigation technology, and in particular to a method for rapid localization of brain stimulation targets based on facial recognition and AI neuronavigation. Background Technology

[0002] Transcranial magnetic stimulation (TMS), as a non-invasive brain modulation technique, is increasingly widely used in the treatment and research of neuropsychiatric diseases. Its therapeutic principle involves inducing an electric field in the brain through a time-varying magnetic field, thereby modulating the excitability of specific neural circuits. However, traditional TMS treatment procedures face significant challenges in achieving individualization and precision: before treatment, the determination of motor thresholds and the localization of stimulation targets are highly dependent on the operator's experience, making the process cumbersome and lacking repeatability; during treatment, unavoidable micro-movements of the patient's head can easily lead to off-target stimulation, affecting the stability of the therapeutic effect; furthermore, stimulation protocols are often based on group brain maps, making it difficult to match with the fine differences in individual brain anatomy and function, and it is impossible to intuitively assess the actual distribution of the magnetic field within the brain. Although existing neuronavigation systems incorporate medical imaging, they still have limitations such as complex registration processes, inability to guide the optimal stimulation angle in real time, and lack of dynamic compensation during treatment, which restrict the full realization and development of the therapeutic effect of TMS technology. Summary of the Invention

[0003] The main objective of this invention is to provide a method for rapid localization of brain stimulation targets based on facial recognition and AI neuronavigation. By integrating facial recognition, individualized three-dimensional brain model construction, real-time optical navigation, and intelligent angle guidance based on cortical anatomy, it aims to achieve rapid, accurate, and visual localization and dynamic maintenance of transcranial magnetic stimulation treatment targets, thereby upgrading traditional experience-dependent operations to personalized, real-time feedback-based precision navigation therapy.

[0004] To achieve the above objectives, this invention provides a method for rapid localization of brain stimulation targets based on facial recognition and AI neural navigation, comprising the following steps:

[0005] The system acquires the user's head MRI image data and uploads it to an image workstation, which then constructs a three-dimensional digital brain model containing information on the sulci and gyri of the cerebral cortex and brain functional regions.

[0006] Collect three-dimensional point cloud data of the user's face, register the three-dimensional point cloud data with the three-dimensional surface model of the head reconstructed from the head MRI image data, and establish a mapping relationship between the user's actual head space and the three-dimensional digital brain model space.

[0007] Stimulation targets are determined based on brain functional partitioning information on the three-dimensional digital brain model.

[0008] The real-time spatial pose of the positioning headgear, which is equipped with optical positioning markers and is worn by the user, is tracked by an optical positioning system.

[0009] In the navigation interface of the image workstation, the real-time spatial position and angle of the three-dimensional digital brain model, the stimulation target point, and the magnetic stimulation coil are displayed based on the real-time spatial pose fusion.

[0010] Based on the direction of the cortical sulci at the stimulation target point in the three-dimensional digital brain model, the target stimulation angle of the magnetic stimulation coil relative to the stimulation target point is determined. Based on the difference between the real-time angle of the magnetic stimulation coil displayed in the navigation interface and the target stimulation angle, the spatial pose of the magnetic stimulation coil is guided and adjusted.

[0011] Further, the steps of acquiring the user's head MRI image data, uploading the head MRI image data to an image workstation, and having the image workstation construct a three-dimensional digital brain model including information on the sulci and gyri of the cerebral cortex and brain functional regions include:

[0012] The head MRI image data is automatically segmented to obtain the brain tissue structure;

[0013] The brain tissue structure is nonlinearly registered with a standard brain atlas template, which includes at least one of the Brodmann partition template, the AAL partition template, and a brain template based on individual brain network connectivity omics.

[0014] Based on the registration results, the brain functional partition information in the standard brain atlas template is mapped to the brain tissue structure to construct a three-dimensional digital brain model containing brain region functional labels.

[0015] Further, the step of acquiring three-dimensional point cloud data of the user's face, registering the three-dimensional point cloud data with a three-dimensional head surface model reconstructed from the head MRI image data, and establishing a mapping relationship between the user's actual head space and the three-dimensional digital brain model space includes:

[0016] A three-dimensional surface model of the user's head is reconstructed from the head MRI image data;

[0017] Collect 3D point cloud data of key facial features of the user;

[0018] Extract the location of feature points corresponding to the key facial features of the user from the three-dimensional surface model of the head;

[0019] Using the correspondence between the key facial features of the user and the feature point positions extracted from the three-dimensional surface model of the head, a point cloud registration algorithm is executed to calculate the coordinate transformation relationship from the user's actual head space to the three-dimensional digital brain model space, thus establishing the mapping relationship.

[0020] Furthermore, the step of determining stimulation targets based on brain functional partitioning information on the three-dimensional digital brain model includes:

[0021] According to the treatment needs, a pre-stored treatment plan is invoked, which includes information on at least one recommended brain region to be stimulated based on a standard brain atlas;

[0022] On the navigation interface of the three-dimensional digital brain model, initial target markers are generated and displayed in the corresponding brain functional areas based on the recommended stimulation brain region information.

[0023] Receive confirmation or adjustment instructions for the initial target mark, and set the determined three-dimensional coordinate position as the stimulation target.

[0024] Furthermore, the step of tracking the real-time spatial pose of the positioning headgear, which is equipped with optical positioning markers and worn by the user, using an optical positioning system includes:

[0025] The infrared camera array of the optical positioning system captures the reflected signals of the optical positioning markers on the positioning head;

[0026] Based on the reflected signals, the three-dimensional spatial coordinates of each optical positioning marker are calculated in real time.

[0027] Based on the three-dimensional spatial coordinates of multiple optical positioning markers and fixed geometric relationships, the real-time spatial pose of the positioning headgear is calculated and output.

[0028] Furthermore, in the navigation interface of the image workstation, the step of displaying the real-time spatial position and angle of the three-dimensional digital brain model, the stimulation target point, and the magnetic stimulation coil based on the real-time spatial pose fusion includes:

[0029] Based on the real-time spatial pose of the positioning headgear, coordinate transformation is calculated and applied to correct the display orientation of the three-dimensional digital brain model and the stimulation target point in the navigation interface to be consistent with the actual orientation of the user's head.

[0030] The optical positioning system tracks the optical markers set on the magnetic stimulation coil to obtain the real-time spatial pose of the magnetic stimulation coil.

[0031] In the same three-dimensional scene of the navigation interface, the model of the magnetic stimulation coil is rendered synchronously based on the real-time spatial pose of the magnetic stimulation coil, thereby completing the fusion display of the three-dimensional digital brain model, the stimulation target point and the relative spatial relationship of the magnetic stimulation coil.

[0032] Further, based on the direction of the cortical sulci at the stimulation target point in the three-dimensional digital brain model, the target stimulation angle of the magnetic stimulation coil relative to the stimulation target point is determined. The step of guiding the adjustment of the spatial pose of the magnetic stimulation coil based on the difference between the real-time angle of the magnetic stimulation coil displayed in the navigation interface and the target stimulation angle includes:

[0033] In the three-dimensional digital brain model, the local tangential direction at the stimulation target point is determined based on the orientation of the main brain sulci.

[0034] The orientation of the magnetic stimulation coil, with its central axis parallel to the local tangent and the direction of maximum stimulation field strength pointing deep into the brain, is set as the target stimulation angle.

[0035] In the navigation interface, guidance for adjusting the spatial pose of the magnetic stimulation coil is generated.

[0036] This invention also provides a rapid brain stimulation target localization device based on facial recognition and AI neuronavigation, comprising:

[0037] The model building module is used to acquire the user's head MRI image data, upload the head MRI image data to the image workstation, and the image workstation constructs a three-dimensional digital brain model containing information on the sulci and gyri of the cerebral cortex and brain functional areas.

[0038] The registration and mapping module is used to acquire three-dimensional point cloud data of the user's face, register the three-dimensional point cloud data with the three-dimensional surface model of the head reconstructed from the head MRI image data, and establish a mapping relationship between the user's actual head space and the three-dimensional digital brain model space.

[0039] The target determination module is used to determine stimulation targets based on brain functional partitioning information on the three-dimensional digital brain model.

[0040] The pose tracking module is used to track the real-time spatial pose of the positioning headgear, which is equipped with optical positioning markers, worn by the user, using an optical positioning system.

[0041] The navigation display module displays the real-time spatial position and angle of the three-dimensional digital brain model, the stimulation target point, and the magnetic stimulation coil in the navigation interface of the image workstation based on the real-time spatial pose fusion.

[0042] The pose guidance module is used to determine the target stimulation angle of the magnetic stimulation coil relative to the stimulation target point based on the direction of the cortical sulci at the stimulation target point in the three-dimensional digital brain model, and guide the adjustment of the spatial pose of the magnetic stimulation coil based on the difference between the real-time angle of the magnetic stimulation coil displayed in the navigation interface and the target stimulation angle.

[0043] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described method for rapid localization of brain stimulation targets based on facial recognition and AI neural navigation.

[0044] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described method for rapid localization of brain stimulation targets based on facial recognition and AI neural navigation.

[0045] The rapid target localization method for brain stimulation based on facial recognition and AI neuronavigation provided by this invention has the following beneficial effects: This invention rapidly and markerlessly registers a 3D facial point cloud with a head surface model reconstructed from MRI, replacing the traditional time-consuming and complex process of marker pasting and registration, significantly shortening preoperative preparation time and improving spatial mapping accuracy. Based on individual MRI data, it automatically constructs a 3D digital brain model integrating multimodal standard brain atlases, achieving precise mapping from group-based coordinates to individualized brain functional areas, making the selection of stimulation targets more scientific and intuitive. Furthermore, by using an optical positioning system to track the head and coil pose in real time and performing fusion display and dynamic compensation in the navigation interface, it fundamentally solves the problem of stimulation off-target caused by patient head movement, ensuring continuous accuracy of stimulation position during treatment. In addition, this invention intelligently calculates and visualizes the optimal placement angle of the magnetic stimulation coil based on the individualized anatomical orientation of the cortical sulci at the target point, ensuring that the stimulation electric field can act on the target brain region along the optimal path, thereby improving the physiological effectiveness and treatment efficiency of the stimulation. It also provides operators with an objective tool for predicting and verifying therapeutic effects by calculating and visualizing the predicted distribution of the stimulation electric field in individual brain models in real time. Attached Figure Description

[0046] Figure 1 This is a flowchart illustrating a method for rapid localization of brain stimulation targets based on facial recognition and AI neural navigation in one embodiment of the present invention.

[0047] Figure 2 This is a structural block diagram of a brain stimulation target rapid localization device based on facial recognition and AI neuronavigation in one embodiment of the present invention;

[0048] Figure 3This is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention.

[0049] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0051] Reference Figure 1 The diagram below illustrates a method for rapid localization of brain stimulation targets based on facial recognition and AI neural navigation, as proposed in this invention. The method includes the following steps:

[0052] S1. Acquire the user's head MRI image data and upload the head MRI image data to the image workstation, and the image workstation constructs a three-dimensional digital brain model containing information on the sulci and gyri of the cerebral cortex and brain functional areas.

[0053] S2, collect the three-dimensional point cloud data of the user's face, register the three-dimensional point cloud data with the three-dimensional surface model of the head reconstructed from the head MRI image data, and establish a mapping relationship between the user's actual head space and the three-dimensional digital brain model space.

[0054] S3, on the three-dimensional digital brain model, determine the stimulation target based on brain functional partition information;

[0055] S4, using an optical positioning system to track the real-time spatial pose of a positioning headgear equipped with optical positioning markers worn by the user.

[0056] S5, in the navigation interface of the image workstation, the real-time spatial position and angle of the three-dimensional digital brain model, the stimulation target point and the magnetic stimulation coil are displayed based on the real-time spatial pose fusion.

[0057] S6. Based on the direction of the cortical sulci at the stimulation target point in the three-dimensional digital brain model, determine the target stimulation angle of the magnetic stimulation coil relative to the stimulation target point. Based on the difference between the real-time angle of the magnetic stimulation coil displayed in the navigation interface and the target stimulation angle, guide the adjustment of the spatial pose of the magnetic stimulation coil.

[0058] In one embodiment, for step S1,

[0059] The steps of acquiring the user's head MRI image data, uploading the head MRI image data to an image workstation, and constructing a three-dimensional digital brain model including the sulci and gyri of the cerebral cortex and information on brain functional regions by the image workstation include:

[0060] The head MRI image data is automatically segmented to obtain the brain tissue structure;

[0061] The brain tissue structure is nonlinearly registered with a standard brain atlas template, which includes at least one of the Brodmann partition template, the AAL partition template, and a brain template based on individual brain network connectivity omics.

[0062] Based on the registration results, the brain functional partition information in the standard brain atlas template is mapped to the brain tissue structure to construct a three-dimensional digital brain model containing brain region functional labels.

[0063] In practice, high-resolution structural magnetic resonance imaging (MRI) images of the user's whole brain are acquired and uploaded to a dedicated image workstation in digital format. The image workstation initiates an automated processing flow: the raw head MRI images are automatically segmented. This process uses image recognition algorithms to intelligently distinguish and extract key brain tissue structures such as complete cerebral cortex gray matter, white matter, and cerebrospinal fluid, while removing non-brain tissue interference such as skull bones, obtaining pure brain tissue structure data suitable for advanced analysis. To achieve precise brain function localization, the segmented user brain tissue structure data is nonlinearly registered with one or more high-precision standard brain atlas templates. The standard brain atlas template is a selectable set of templates, including at least a Brodmann partition template based on cellular architecture features, an AAL (Automated Anatomical Labeling) partition template based on anatomical landmarks, and a brain template constructed based on individual brain network connectomics. One or more templates can be used for registration. Nonlinear registration technology can finely correct anatomical differences in shape, size, and sulcus morphology between different individuals' brains, elastically deforming the coordinate system of the standard template to perfectly match the user's brain. Based on this registration result, the scientifically defined brain functional partition information (e.g., the boundaries of brain regions responsible for specific functions such as movement, language, and vision) from the selected standard brain atlas template is accurately mapped and assigned to the user's own brain structure. By integrating the segmented brain anatomy and the mapped functional partition information, a personalized 3D digital brain model unique to the user is constructed, containing a detailed 3D geometric shape of cortical sulci and gyri, and embedding multi-level brain functional partition labels.

[0064] In one embodiment, for step S2,

[0065] The steps of acquiring three-dimensional point cloud data of the user's face, registering the three-dimensional point cloud data with a three-dimensional head surface model reconstructed from the head MRI image data, and establishing a mapping relationship between the user's actual head space and the three-dimensional digital brain model space include:

[0066] A three-dimensional surface model of the user's head is reconstructed from the head MRI image data;

[0067] Collect 3D point cloud data of key facial features of the user;

[0068] Extract the location of feature points corresponding to the key facial features of the user from the three-dimensional surface model of the head;

[0069] Using the correspondence between the key facial features of the user and the feature point positions extracted from the three-dimensional surface model of the head, a point cloud registration algorithm is executed to calculate the coordinate transformation relationship from the user's actual head space to the three-dimensional digital brain model space, thus establishing the mapping relationship.

[0070] In practice, a three-dimensional surface model of the user's head is reconstructed from head MRI image data. This model specifically extracts and reconstructs the skin surface contour of the user's head (especially the face, which contains rich geometric features) from the MRI data. It is a triangular mesh model that only expresses the external morphology, providing a digital twin target for subsequent registration with real-world sensor data. Simultaneously, during the treatment preparation phase, three-dimensional point cloud data of the user's key facial features is acquired using three-dimensional sensing devices such as depth cameras or structured light scanners. This point cloud data is a collection of high-density spatial coordinate points of the user's actual face in the coordinate system of the treatment room, realistically reflecting its immediate three-dimensional facial morphology. In parallel, feature point locations corresponding to the user's key facial features are extracted from the head three-dimensional surface model. These feature points are typically selected from areas with stable anatomical significance and easy identification in the point cloud, such as the tip of the nose, the corners of the eyes, and the corners of the mouth. Accordingly, in the acquired facial three-dimensional point cloud, feature detection algorithms automatically or assistedly identify these identical anatomical points. A point cloud registration algorithm is executed, establishing a correspondence between the user's key facial features (from point cloud data) and the feature point positions extracted from the 3D head surface model. In practice, the Iterative Closest Point (ICP) algorithm or its variants are commonly used: two sets of feature points are used as initial matching pairs, and an optimal spatial rigid body transformation (typically including three translation parameters and three rotation parameters) is iteratively calculated to minimize the overall spatial distance error between the transformed feature points in the point cloud data and their corresponding feature points on the model surface. This is used to calculate the coordinate transformation relationship from the user's actual head space (i.e., the treatment room coordinate system where the point cloud is located) to the 3D digital brain model space (i.e., the MRI image coordinate system). This transformation relationship is a precise 4x4 homogeneous transformation matrix that defines how points and directions are transformed between two completely independent coordinate systems. Once this transformation relationship is determined and verified, the mapping relationship is considered established. Subsequently, the pose of any object tracked by the optical system in the real head space (such as a positioning headgear or magnetic stimulation coil) can be accurately mapped to the virtual three-dimensional digital brain model space through this transformation relationship, thereby achieving millimeter-level precision fusion of virtual navigation information and the real physical world, laying a spatial foundation for subsequent real-time visualization and guidance.

[0071] In one embodiment, for step S3,

[0072] The steps for determining stimulation targets based on brain functional partitioning information on the three-dimensional digital brain model include:

[0073] According to the treatment needs, a pre-stored treatment plan is invoked, which includes information on at least one recommended brain region to be stimulated based on a standard brain atlas;

[0074] On the navigation interface of the three-dimensional digital brain model, initial target markers are generated and displayed in the corresponding brain functional areas based on the recommended stimulation brain region information.

[0075] Receive confirmation or adjustment instructions for the initial target mark, and set the determined three-dimensional coordinate position as the stimulation target.

[0076] In practice, pre-stored treatment plans are invoked according to treatment needs. The treatment plan library is a digital database integrating a large amount of clinical medical knowledge and practical experience, pre-stored with standardized treatment strategies supported by evidence-based medicine for different neuropsychiatric indications (such as depression, motor dysfunction, etc.). Each treatment plan contains at least one recommended brain region stimulation information based on a standard brain atlas. This information is usually in the form of three-dimensional coordinates in a standard brain atlas space (such as MNI space) or specific functional brain region names, clearly identifying the core brain network nodes that should be intervened for the current condition. Treatment plans can also be manually entered. After invoking a treatment plan, the interactive target planning stage begins: on the navigation interface of the three-dimensional digital brain model, initial target markers are generated and displayed in the corresponding brain functional areas based on the recommended brain region stimulation information. Specifically, using the brain atlas registration results completed in step S1, the recommended coordinates or brain region names based on standard templates in the treatment plan are automatically converted and mapped onto the current user's personalized three-dimensional digital brain model. The navigation interface automatically generates an initial target marker at the corresponding location after conversion, using prominent visual elements (such as a 3D cursor, a highlighted sphere, or a crosshair). This marker is then overlaid and rendered onto the 3D brain model, visually displaying the anatomical location of the recommended stimulus and its functional zoning environment. The system receives confirmation or adjustment instructions for the initial target marker. Based on their clinical experience, interpretation of individual images (e.g., avoiding obvious blood vessels or sulci), or other physiological considerations, physicians can freely move and finely adjust the initial target marker in 3D space using mouse dragging, touch, or a dedicated 3D interactive device. After confirmation or adjustment, the physician sets this final determined 3D coordinate position as the stimulation target and records its precise value in the 3D digital brain model coordinate system.

[0077] In one embodiment, for step S4,

[0078] The steps of tracking the real-time spatial pose of a positioning headgear equipped with optical positioning markers worn by the user using an optical positioning system include:

[0079] The infrared camera array of the optical positioning system captures the reflected signals of the optical positioning markers on the positioning head;

[0080] Based on the reflected signals, the three-dimensional spatial coordinates of each optical positioning marker are calculated in real time.

[0081] Based on the three-dimensional spatial coordinates of multiple optical positioning markers and fixed geometric relationships, the real-time spatial pose of the positioning headgear is calculated and output.

[0082] In practical implementation, this step is a crucial physical perception step for achieving dynamic, real-time navigation. Its core is to continuously and accurately track the spatial movement of the user's head during treatment, thus providing a data foundation for real-time registration of virtual navigation information with the real physical space. This relies on a positioning headgear pre-worn by the user, on which multiple optical positioning markers (e.g., spherical infrared reflective markers) are rigidly fixed. These markers work in conjunction with an optical positioning system. This system comprises an array of two or more infrared cameras, geometrically distributed in space to cover the entire field of view. The infrared camera array of the optical positioning system captures the reflected signals from the optical positioning markers on the headgear. Each camera in the array synchronously emits infrared light and receives the light signal reflected back from the marker surface, i.e., the reflected signal. Since multiple cameras observe the same set of markers from different perspectives, multi-view two-dimensional image information of the markers is obtained. Based on the reflected signals, the three-dimensional spatial coordinates of each optical positioning marker are calculated in real time using the principle of stereoscopic triangulation. By matching the imaging position of each marker in different camera images and combining the precisely calibrated internal parameters (such as focal length and distortion) and external parameters (relative position and attitude) between the cameras, the precise three-dimensional coordinates (X, Y, Z) of each marker in the coordinate system of the optical positioning system can be obtained through geometric calculations. Based on the three-dimensional spatial coordinates of multiple optical positioning markers and their fixed geometric relationships, the real-time spatial pose of the positioning headgear is calculated and output. The real-time spatial pose is a six-degree-of-freedom concept, including three translational degrees of freedom (position in the X, Y, and Z axes) and three rotational degrees of freedom (rotation angles around the X, Y, and Z axes, usually represented by Euler angles or quaternions). Specifically, the calculation involves the multiple markers on the headgear forming a rigid geometric structure. By identifying this rigid structure and comparing it with the known theoretical geometric layout of the markers on the headgear, mathematical algorithms (such as singular value decomposition (SVD)) are used to solve for the unique spatial transformation that best matches the observed set of marker coordinates to its theoretical layout. This transformation represents the real-time position and orientation of the positioning headgear (i.e., the user's head) relative to the coordinate system of the optical positioning system. Step S4 uses sophisticated hardware systems and algorithms to convert the physical movement of the user's head into a continuous, high-frequency digital pose data stream. This data stream is a prerequisite for subsequent steps to achieve real-time, dynamic fusion display of the three-dimensional brain model and stimulation coils, as well as for compensating for head movements during treatment, ensuring spatial consistency and treatment accuracy throughout the entire navigation guidance process.

[0083] In one embodiment, for step S5,

[0084] In the navigation interface of the image workstation, the steps of displaying the real-time spatial position and angle of the three-dimensional digital brain model, the stimulation target point, and the magnetic stimulation coil based on the real-time spatial pose fusion include:

[0085] Based on the real-time spatial pose of the positioning headgear, coordinate transformation is calculated and applied to correct the display orientation of the three-dimensional digital brain model and the stimulation target point in the navigation interface to be consistent with the actual orientation of the user's head.

[0086] The optical positioning system tracks the optical markers set on the magnetic stimulation coil to obtain the real-time spatial pose of the magnetic stimulation coil.

[0087] In the same three-dimensional scene of the navigation interface, the model of the magnetic stimulation coil is rendered synchronously based on the real-time spatial pose of the magnetic stimulation coil, thereby completing the fusion display of the three-dimensional digital brain model, the stimulation target point and the relative spatial relationship of the magnetic stimulation coil.

[0088] In practical implementation, the synchronous display of the head and the virtual model is handled as follows: Based on the real-time spatial pose of the positioning headgear, coordinate transformation is calculated and applied to correct the display orientation of the three-dimensional digital brain model and the stimulation target points in the navigation interface to be consistent with the orientation of the user's actual head. Specifically, using the "mapping relationship between the user's actual head space and the three-dimensional digital brain model space" established in step S2 and the "pose of the positioning headgear in actual space" obtained in real time in step S4, through continuous coordinate transformation (i.e., applying the inverse transformation of the above mapping relationship and combining it with the real-time pose of the headgear), the dynamic calculation is made as to how the three-dimensional digital brain model should be rendered on the screen at the current moment, simulating the visual effect of viewing the user's real head and its internal brain structure from the operator's perspective. When the user's head rotates or moves, the perspective of the brain model displayed on the screen also changes naturally and accurately, ensuring the consistency between the virtual scene and physical reality in spatial perception.

[0089] Simultaneously, the system synchronously tracks the treatment device: the optical positioning system tracks optical markers set on the magnetic stimulation coil to obtain the real-time spatial pose of the magnetic stimulation coil. This is similar to the tracking and positioning headgear, where optical markers are also fixedly installed on the magnetic stimulation coil to form a rigid body. The optical positioning system calculates the precise position (X, Y, Z) and three-dimensional angular orientation of the coil in the same world coordinate system at a high frame rate (e.g., tens of hertz). Information fusion and presentation are then completed: in the same three-dimensional scene of the navigation interface, based on the real-time spatial pose of the magnetic stimulation coil, a model of the magnetic stimulation coil is synchronously rendered, completing the fusion display of the three-dimensional digital brain model, the stimulation target point, and the relative spatial relationship of the magnetic stimulation coil. Specifically, a unified three-dimensional rendering scene is created through a graphics workstation. In this scene, based on the aforementioned calculations, a three-dimensional digital brain model with corrected orientation and stimulation target points marked on it are drawn in real time. Simultaneously, based on the real-time acquired magnetic stimulation coil pose data, a three-dimensional model of the magnetic stimulation coil, consistent with the actual geometry, is drawn at the corresponding spatial position. All elements are rendered based on the same world coordinate system, allowing the operator to directly and in real-time observe on the screen whether the virtual magnetic stimulation coil model is aligned with the target point inside the brain model, the distance between them, and the current angle of the coil. This fusion display, which maps the real spatial state of the device in real time and integrates it into the individualized brain anatomy scene, greatly reduces the operator's cognitive load, transforming the originally abstract spatial alignment problem into an intuitive visual alignment task. It is an indispensable human-computer interaction interface for achieving accurate and efficient positioning.

[0090] In one embodiment, for step S6,

[0091] Based on the direction of the cortical sulci at the stimulation target point in the three-dimensional digital brain model, the target stimulation angle of the magnetic stimulation coil relative to the stimulation target point is determined. Based on the difference between the real-time angle of the magnetic stimulation coil displayed in the navigation interface and the target stimulation angle, the spatial pose of the magnetic stimulation coil is guided and adjusted, including:

[0092] In the three-dimensional digital brain model, the local tangential direction at the stimulation target point is determined based on the orientation of the main brain sulci.

[0093] The orientation of the magnetic stimulation coil, with its central axis parallel to the local tangent and the direction of maximum stimulation field strength pointing deep into the brain, is set as the target stimulation angle.

[0094] In the navigation interface, guidance for adjusting the spatial pose of the magnetic stimulation coil is generated.

[0095] In practical implementation, key geometric information is extracted from individualized anatomical data: In the three-dimensional digital brain model, the local tangent direction at the stimulation target point is determined based on the orientation of the major sulci. The cortical surface geometry at the target point is automatically analyzed, and by calculating the principal curvature direction or fitting the local surface, the tangent vector consistent with the extension direction of the major sulci (such as the central sulcus, superior frontal sulcus, etc.) on the cortical surface is identified. Anatomically, this direction is usually closely related to the orientation of the subcortical white matter fiber bundles and is the optimal path for determining how the magnetic field effectively couples into the brain and induces neural activity. This anatomical information is then converted into precise device control parameters: the orientation of the magnetic stimulation coil, with its central axis parallel to the local tangent direction and the direction of maximum stimulation field strength pointing deep into the brain, is set as the target stimulation angle. Here, the "target stimulation angle" is a complete spatial orientation containing six degrees of freedom. The core criteria for determining this posture are twofold: firstly, the central axis of the coil must be parallel to the calculated local tangent direction of the cortex, ensuring that the magnetic field generated by the coil can penetrate the cortical surface most effectively; secondly, the coil's asymmetrical figure-eight or biconical design ensures that the peak value of the induced electric field (i.e., the direction of the maximum stimulation field strength) is perpendicular to the central axis and precisely points to the deep white matter region of the brain, rather than spreading along the tangential direction. This calculation comprehensively considers the principles of electromagnetic physics and individual brain anatomy, aiming to maximize the efficiency of stimulation energy transfer to the target neural circuit. Finally, a seamless transition from static planning to dynamic operation is achieved: the navigation interface generates guidance for adjusting the spatial pose of the magnetic stimulation coil. The actual spatial pose of the coil, currently obtained through optical positioning, is compared in real time with the calculated "target stimulation angle." Based on this difference, the navigation interface proactively generates intuitive and clear guidance instructions. For example, in a three-dimensional scene, this might be indicated in real time, in the form of dynamic arrows, colored highlights, or numerical deviation bars, showing the operator which spatial direction to rotate the coil in and by how many degrees to achieve the optimal angle. Operators no longer need to rely on abstract experience-based judgments; they can simply follow the visual cues on the interface, much like following a navigation system, to gradually adjust the coil to the theoretically optimal stimulation posture. This step transforms the geometric features of an individual's cortical anatomy into the optimal spatial posture of the stimulation device and constructs a real-time, visualized closed-loop guidance system. This solves the problems of traditional methods where coil angle settings rely on experience and are difficult to standardize and optimize. Furthermore, it elevates the precision of transcranial magnetic stimulation from "target localization" to a new dimension of "vector field modulation."

[0096] Reference Figure 2 The diagram shows a structural block diagram of a brain stimulation target rapid localization device based on facial recognition and AI neural navigation in one embodiment of the present invention, comprising:

[0097] The model building module is used to acquire the user's head MRI image data, upload the head MRI image data to the image workstation, and the image workstation constructs a three-dimensional digital brain model containing information on the sulci and gyri of the cerebral cortex and brain functional areas.

[0098] The registration and mapping module is used to acquire three-dimensional point cloud data of the user's face, register the three-dimensional point cloud data with the three-dimensional surface model of the head reconstructed from the head MRI image data, and establish a mapping relationship between the user's actual head space and the three-dimensional digital brain model space.

[0099] The target determination module is used to determine stimulation targets based on brain functional partitioning information on the three-dimensional digital brain model.

[0100] The pose tracking module is used to track the real-time spatial pose of the positioning headgear, which is equipped with optical positioning markers, worn by the user, using an optical positioning system.

[0101] The navigation display module displays the real-time spatial position and angle of the three-dimensional digital brain model, the stimulation target point, and the magnetic stimulation coil in the navigation interface of the image workstation based on the real-time spatial pose fusion.

[0102] The pose guidance module is used to determine the target stimulation angle of the magnetic stimulation coil relative to the stimulation target point based on the direction of the cortical sulci at the stimulation target point in the three-dimensional digital brain model, and guide the adjustment of the spatial pose of the magnetic stimulation coil based on the difference between the real-time angle of the magnetic stimulation coil displayed in the navigation interface and the target stimulation angle.

[0103] For the specific implementation of each module in the above device example, please refer to the above method embodiments, which will not be repeated here.

[0104] Reference Figure 3 This invention also provides a computer device, which can be a server, and its internal structure can be as follows: Figure 3 As shown, the computer device includes a processor, memory, display screen, input device, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores the data corresponding to this embodiment. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements the above-described method.

[0105] Those skilled in the art will understand that Figure 3 The structures shown are merely block diagrams of some structures related to the present invention and do not constitute a limitation on the computer devices on which the present invention is applied.

[0106] An embodiment of the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method. It is understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0107] In summary, this invention acquires MRI images of a user's head and uploads them to an image workstation, where the workstation constructs a three-dimensional digital brain model including the sulci and gyri of the cerebral cortex and information on brain functional regions. It also acquires three-dimensional point cloud data of the user's face and registers this data with a three-dimensional head surface model reconstructed from the MRI images, establishing a mapping between the user's actual head space and the space of the three-dimensional digital brain model. Stimulation targets are determined on the three-dimensional digital brain model based on brain functional region information. Finally, an optical positioning system tracks a positioning headgear equipped with optical positioning markers worn by the user. The real-time spatial pose of the image workstation is displayed in the navigation interface based on the real-time spatial pose. The real-time spatial position and angle of the three-dimensional digital brain model, the stimulation target point, and the magnetic stimulation coil are fused and displayed. According to the direction of the cortical sulci at the stimulation target point in the three-dimensional digital brain model, the target stimulation angle of the magnetic stimulation coil relative to the stimulation target point is determined. Based on the difference between the real-time angle of the magnetic stimulation coil displayed in the navigation interface and the target stimulation angle, the spatial pose of the magnetic stimulation coil is guided and adjusted to achieve rapid, accurate, visual positioning and dynamic maintenance of the transcranial magnetic stimulation treatment target point. This upgrades the traditional experience-dependent operation to individualized, real-time feedback precision navigation treatment.

[0108] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the present invention and embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual-rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM, etc.

[0109] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0110] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for rapid localization of brain stimulation targets based on facial recognition and AI neural navigation, characterized in that, Includes the following steps: The system acquires the user's head MRI image data and uploads it to an image workstation, which then constructs a three-dimensional digital brain model containing information on the sulci and gyri of the cerebral cortex and brain functional regions. Collect three-dimensional point cloud data of the user's face, register the three-dimensional point cloud data with the three-dimensional surface model of the head reconstructed from the head MRI image data, and establish a mapping relationship between the user's actual head space and the three-dimensional digital brain model space. Stimulation targets are determined based on brain functional partitioning information on the three-dimensional digital brain model. The real-time spatial pose of the positioning headgear, which is equipped with optical positioning markers and is worn by the user, is tracked by an optical positioning system. In the navigation interface of the image workstation, the real-time spatial position and angle of the three-dimensional digital brain model, the stimulation target point, and the magnetic stimulation coil are displayed based on the real-time spatial pose fusion. This includes: calculating and applying coordinate transformation based on the real-time spatial pose of the positioning headgear to correct the display orientation of the three-dimensional digital brain model and the stimulation target point in the navigation interface to be consistent with the actual orientation of the user's head; tracking optical markers set on the magnetic stimulation coil through the optical positioning system to obtain the real-time spatial pose of the magnetic stimulation coil; and synchronously rendering the model of the magnetic stimulation coil in the same three-dimensional scene of the navigation interface based on the real-time spatial pose of the magnetic stimulation coil to complete the fusion display of the relative spatial relationship between the three-dimensional digital brain model, the stimulation target point, and the magnetic stimulation coil. Based on the direction of the cortical sulci at the stimulation target point in the three-dimensional digital brain model, the target stimulation angle of the magnetic stimulation coil relative to the stimulation target point is determined. Based on the difference between the real-time angle of the magnetic stimulation coil displayed in the navigation interface and the target stimulation angle, the spatial pose of the magnetic stimulation coil is guided and adjusted.

2. The method for rapid localization of brain stimulation targets based on facial recognition and AI neural navigation according to claim 1, characterized in that, The steps of acquiring the user's head MRI image data, uploading the head MRI image data to an image workstation, and having the image workstation construct a three-dimensional digital brain model including information on the sulci and gyri of the cerebral cortex and brain functional areas include: The head MRI image data is automatically segmented to obtain the brain tissue structure; The brain tissue structure is nonlinearly registered with a standard brain atlas template, which includes at least one of the Brodmann partition template, the AAL partition template, and a brain template based on individual brain network connectivity omics. Based on the registration results, the brain functional partition information in the standard brain atlas template is mapped to the brain tissue structure to construct a three-dimensional digital brain model containing brain region functional labels.

3. The method for rapid localization of brain stimulation targets based on facial recognition and AI neural navigation according to claim 1, characterized in that, The steps of acquiring three-dimensional point cloud data of the user's face, registering the three-dimensional point cloud data with a three-dimensional head surface model reconstructed from the head MRI image data, and establishing a mapping relationship between the user's actual head space and the three-dimensional digital brain model space include: A three-dimensional surface model of the user's head is reconstructed from the head MRI image data; Collect 3D point cloud data of key facial features of the user; Extract the location of feature points corresponding to the key facial features of the user from the three-dimensional surface model of the head; Using the correspondence between the key facial features of the user and the feature point positions extracted from the three-dimensional surface model of the head, a point cloud registration algorithm is executed to calculate the coordinate transformation relationship from the user's actual head space to the three-dimensional digital brain model space, thus establishing the mapping relationship.

4. The method for rapid localization of brain stimulation targets based on facial recognition and AI neural navigation according to claim 1, characterized in that, The step of determining stimulation targets based on brain functional partitioning information on the three-dimensional digital brain model includes: According to the treatment needs, a pre-stored treatment plan is invoked, which includes information on at least one recommended brain region to be stimulated based on a standard brain atlas; On the navigation interface of the three-dimensional digital brain model, initial target markers are generated and displayed in the corresponding brain functional areas based on the recommended stimulation brain region information. Receive confirmation or adjustment instructions for the initial target mark, and set the determined three-dimensional coordinate position as the stimulation target.

5. The method for rapid localization of brain stimulation targets based on facial recognition and AI neural navigation according to claim 1, characterized in that, The step of tracking the real-time spatial pose of a positioning headgear equipped with optical positioning markers worn by the user using an optical positioning system includes: The infrared camera array of the optical positioning system captures the reflected signals of the optical positioning markers on the positioning head; Based on the reflected signals, the three-dimensional spatial coordinates of each optical positioning marker are calculated in real time. Based on the three-dimensional spatial coordinates of multiple optical positioning markers and fixed geometric relationships, the real-time spatial pose of the positioning headgear is calculated and output.

6. The method for rapid localization of brain stimulation targets based on facial recognition and AI neural navigation according to claim 1, characterized in that, The step of determining the target stimulation angle of the magnetic stimulation coil relative to the target stimulation point based on the direction of the cortical sulci at the stimulation target point in the three-dimensional digital brain model, and guiding the adjustment of the spatial pose of the magnetic stimulation coil based on the difference between the real-time angle of the magnetic stimulation coil displayed in the navigation interface and the target stimulation angle, includes: In the three-dimensional digital brain model, the local tangential direction at the stimulation target point is determined based on the orientation of the main brain sulci. The orientation of the magnetic stimulation coil, with its central axis parallel to the local tangent and the direction of maximum stimulation field strength pointing deep into the brain, is set as the target stimulation angle. In the navigation interface, guidance for adjusting the spatial pose of the magnetic stimulation coil is generated.

7. A rapid brain stimulation target localization device based on facial recognition and AI neural navigation, characterized in that, include: The model building module is used to acquire the user's head MRI image data, upload the head MRI image data to the image workstation, and the image workstation constructs a three-dimensional digital brain model containing information on the sulci and gyri of the cerebral cortex and brain functional areas. The registration and mapping module is used to acquire three-dimensional point cloud data of the user's face, register the three-dimensional point cloud data with the three-dimensional surface model of the head reconstructed from the head MRI image data, and establish a mapping relationship between the user's actual head space and the three-dimensional digital brain model space. The target determination module is used to determine stimulation targets based on brain functional partitioning information on the three-dimensional digital brain model. The pose tracking module is used to track the real-time spatial pose of the positioning headgear, which is equipped with optical positioning markers, worn by the user, using an optical positioning system. The navigation display module displays the real-time spatial position and angle of the three-dimensional digital brain model, the stimulation target point, and the magnetic stimulation coil in the navigation interface of the image workstation based on the real-time spatial pose fusion. The pose guidance module is used to determine the target stimulation angle of the magnetic stimulation coil relative to the stimulation target point based on the direction of the cortical sulci at the stimulation target point in the three-dimensional digital brain model, and guide the adjustment of the spatial pose of the magnetic stimulation coil based on the difference between the real-time angle of the magnetic stimulation coil displayed in the navigation interface and the target stimulation angle.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the rapid localization method for brain stimulation targets based on facial recognition and AI neural navigation as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for rapid localization of brain stimulation targets based on facial recognition and AI neuronavigation as described in any one of claims 1 to 6.

Images