Orthodontic micro-implant ai-mr navigation and marker clip registration system

Abstract

The application relates to the technical field of oral medicine, and discloses an orthodontic micro-implant AI-MR navigation and mark clamp registration system, which comprises the following modules: a data fusion module, which collects CBCT data and intraoral scanning data, and uses an image fusion algorithm to construct a three-dimensional digital dental model; an intelligent path planning module, which analyzes tooth root and alveolar bone morphology by using an AI algorithm, and plans an optimal implantation path; a registration and virtual-real registration module, which realizes virtual-real registration by clamping a registration mark clamp or other registration anchor point in the incisal edge area of anterior teeth, constructs an MR space registration post-model, and maps the optimal implantation path to a real coordinate system to obtain a virtual path; and an intraoperative navigation module, which tracks real-time position changes of an MR head-mounted device and updates a virtual path superposition position, and performs navigation guidance in the whole implantation process. The application scheme can improve positioning accuracy, operation visibility and system adaptability in the implantation process of an orthodontic micro-implant.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of digital dentistry, computer vision and mixed reality, and in particular to an orthodontic micro-implant AI-MR navigation and marker clip registration system. BACKGROUND

[0002] Orthodontic micro-implants (bracket pins) have become an important auxiliary device for modern orthodontic treatment, but due to the close proximity of the implant path to the tooth root and the fine angle control, the implant operation has a high technical sensitivity. Doctors need to manually complete the position and angle judgment based on experience, and thus there is a high risk of error, such as loose micro-implant pins due to poor implant position, and touching the tooth root position during the micro-implant pin implantation process. The traditional method relies on two-dimensional or three-dimensional images, or the production of a guide plate, which is complex to design, has many limitations, and is difficult to adapt to complex soft tissue or limited mouth opening conditions.

[0003] In recent years, mixed reality (MR) technology has been explored for application in oral navigation, but existing systems have two key problems:

[0004] 1) In the process of applying mixed reality to oral navigation, the system must first achieve spatial registration between the real oral cavity and the three-dimensional virtual model through markers. Existing methods often rely on bite plates with recognition maps or orthodontic bracket markers, but such solutions are not available in bracketless invisible aligners, or have problems such as large foreign object volume, complex production, and large individual differences, limiting their wide adaptability and stability.

[0005] Especially in orthodontic micro-implant implantation surgery, the anchorage pin can be implanted on the buccal side (closed mouth state) or the palatal side (open mouth state), with high requirements for the position, stability and compatibility of the registration device, and existing registration markers are difficult to meet all scenarios.

[0006] 2) The navigation path needs to be manually developed, and there is a lack of optimal path intelligent recommendation and real-time feedback mechanism based on anatomical structure. Especially in patients with narrow tooth root spacing, complex alveolar bone morphology or asymmetric oral structure, path development relies on the experience of doctors, and there are problems such as insufficient risk structure avoidance and poor pin stability. SUMMARY

[0007] The present application provides an orthodontic micro-implant AI-MR navigation and marker clip registration system for improving the positioning accuracy, operation visibility and system adaptability in the process of orthodontic micro-implant implantation, and is suitable for preoperative design, intraoperative navigation and postoperative review in complex oral structure scenarios.

[0008] To this end, the present application provides the following technical solutions:

[0009] An orthodontic micro-implant AI-MR navigation and landmark clip registration system, the system comprises:

[0010] A data fusion module collects cone beam computed tomography data and intraoral scanning data, and uses an image fusion algorithm to construct a three-dimensional digital dental model containing crowns, roots, and alveolar bone;

[0011] An intelligent path planning module analyzes the morphology of the roots and alveolar bone based on the three-dimensional digital dental model, and generates path data containing heat map scores and optimal implantation paths;

[0012] A registration and virtual-real registration module uses at least one registration anchor to complete virtual-real coordinate alignment, the registration anchor is preferably a incisal edge registration clip clamped in the incisal edge area of the maxillary central incisor, and a mixed reality MR display terminal including a head-mounted / handheld device identifies a standard image recognition pattern provided on the top of the incisal edge registration clip, estimates the 6 degrees of freedom DoF pose of the standard image recognition pattern, obtains a registration matrix and a spatial conversion matrix, and maps the path data and the three-dimensional digital dental model to a real coordinate system;

[0013] An intraoperative navigation module superimposes the path data and reference guide information in real time in the real scene based on the MR display terminal, and compares with the implant tool pose to obtain correction prompts for output direction, angle, and depth;

[0014] A data closed loop module is used to record intraoperative deviations and implant tool trajectories, and to perform postoperative precision analysis and visual report generation.

[0015] Optionally, the cone beam computed tomography data collected by the data fusion module includes crown, root, and alveolar bone morphology, and the intraoral scanning data collected by the intraoral scanner includes tooth surface, i.e., crown, and gingival morphology; the data fusion module performs medical image processing on the cone beam computed tomography data, performs oral scanning modeling on the intraoral scanning data, uses an image fusion algorithm to reconstruct and fuse a three-dimensional model, and obtains the three-dimensional digital dental model including the uniform scale and coordinate system of the roots, alveolar bone, and crowns.

[0016] Optionally, the intelligent path planning module uses a neural network or a heuristic algorithm based on anatomical rules to plan an optimal implantation path including position and direction and visualize it as a heat map score distribution based on alveolar bone thickness, interradicular distance measurement, and surgical accessibility, and micro-implant nail implantation angle requirements.

[0017] Optionally, the intelligent path planning module supports interactive fine-tuning of entry point, incident angle, and target depth, and is constrained by a safety threshold. When it exceeds the threshold, it gives a prompt and refuses to update.

[0018] Optionally, the registration and virtual-real registration module uses at least one of AprilTag, ArUco, checkerboard or two-dimensional code recognition algorithm to recognize the standard image recognition pattern, and uses PnP algorithm + ICP algorithm to complete 6DoF pose solution of the standard image recognition pattern in space, complete three-dimensional pose recognition, obtain a registration matrix and a space conversion matrix, and a mapped real-time navigation view of virtual-real superposition, construct a registered MR space model, and complete alignment of the entity model to the virtual model coordinate system.

[0019] Optionally, the intraoperative navigation module superimposes a virtual path and a space reference guide line in real time according to the registered MR space model, and assists the doctor in controlling the direction, depth and angle, and supports comparison and correction of the implant tool posture; the real-time position change of the user's head is collected through an MR device camera or an inertial measurement unit (IMU).

[0020] Optionally, the system further comprises a user interaction and tracking module, which uses a LeapMotion gesture recognition device or voice control to switch models, adjust paths, track head position or the shift of the resection registration marker clip in real time, maintain registration consistency, and cooperate with the implant tool to realize direction and depth correction prompts by comparing the implant tool posture with the virtual path; specifically including: dynamically adjusting the navigation interface in real time using a space tracking and posture updating algorithm; comparing the direction of the implant tool with the angle of the preset path, and prompting the deviation through an angle difference feedback algorithm; displaying depth scale, direction guide line and hot map area marking information; intraoperative navigation errors are recorded and used for postoperative path trajectory comparison and precision analysis.

[0021] A resection registration marker clip for oral digital navigation, which is applied to the orthodontic micro-implant AI-MR navigation and marker clip registration system, the clip body is a U-shaped clamping structure, the buccal and lingual arms of which are used to cover the upper and lower edges of the tooth resection and are provided with anti-slip pads, which are made of silicone or foam and can be replaced or reused after disinfection. A pattern recognition platform is provided at the top of the clip body to install or integrally form the coded planar marker; the clip body material is elastic to adapt to different crown shapes and the presence of brackets / attachments; the pattern recognition platform is elevated relative to the resection and has a forward inclination angle to improve the visibility and pose analysis stability of the machine-readable pattern.

[0022] Optionally, the pattern recognition platform is elevated by 2-5 mm relative to the resection and inclined forward by 5-15°, and the effective recognition area of the pattern recognition platform is ≥8 mm in length.

[0023] A path planning method for orthodontic micro-implants, the method comprising: constructing a three-dimensional digital dental model; calculating a root distance map and a bone thickness map and generating a feasible region by combining prohibited structures and surgical accessibility constraints; outputting a candidate path set and a scoring heatmap using a neural network or an anatomical rule-based heuristic algorithm; and interactively fine-tuning the entry point, incident angle and target depth under threshold constraints and outputting the optimal implantation path.

[0024] The orthodontic micro-implant AI-MR navigation and marker clip registration system provided by this invention introduces an AI path planning module. Utilizing anatomical data, such as alveolar bone thickness, root position and spacing, and surgical approach angle reconstructed by cone-beam computed tomography (CBCT), it automatically generates implantation path plans. Combined with a heatmap scoring module, it indicates safe implantation zones, significantly improving the scientific rigor, individual adaptability, and clinical safety of path selection. This invention integrates CBCT and intraoral scanning to create a 3D oral model, uses AI to generate the optimal implantation path, and completes spatial registration using 3D-printed marker clips that hold the incisors. A pattern platform equipped with AprilTag recognition patterns allows MR equipment to track and identify the patterns. The system supports intraoperative guidance, postoperative analysis, path heatmap display, navigation error feedback, and interactive control, achieving a closed loop of preoperative modeling, intelligent path planning, registration and alignment, intraoperative navigation, and interactive feedback. It features stability, high precision, and wide adaptability, making it suitable for various scenarios such as orthodontic navigation, implant assistance, and educational demonstrations. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly described below. Obviously, the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0026] Figure 1 This is a structural diagram of the AI-MR navigation and marker clip registration system for orthodontic micro-implants in a specific embodiment of the present invention;

[0027] Figure 2 This is a schematic diagram of the structure of the cut-end registration mark clip used in a specific embodiment of the present invention;

[0028] Figure 3 This is a schematic diagram of the clamping and holding state of the cut-end registration mark clamp used in a specific embodiment of the present invention;

[0029] Figure 4 This is a schematic diagram illustrating the use of an MR head-mounted display device for tag recognition and model registration in a specific embodiment of the present invention;

[0030] Figure 5 Structure diagram of MR head-mounted device used in a specific embodiment of the present application;

[0031] Figure 6 Optimal implantation path, heat map and deviation prompt map of orthodontic micro-implant in AI-MR navigation in a specific embodiment of the present application. DETAILED DESCRIPTION

[0032] The specific embodiments of the present application will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only used to illustrate and explain the present application, and are not used to limit the present application.

[0033] In order to enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should fall within the scope of protection of the present application.

[0034] Figure 1 Structure diagram of orthodontic micro-implant AI-MR navigation and marker clip registration system in a specific embodiment of the present application, which uses a marker clip clamped at the end of the central incisor as a registration anchor, combines posture recognition and path prompt, and realizes navigation guidance in the whole implantation process. The system comprises:

[0035] The data fusion module 101 collects CBCT data and intraoral scanning data, respectively performs medical image processing and intraoral scanning modeling, generates a three-dimensional digital dental model containing dental crowns / roots / alveolar bone by using a rigid registration + surface registration image fusion algorithm. The intraoral scanner is used to obtain the intraoral scanning data, which is output as a triangular mesh / point cloud (such as STL / OBJ / PLY) of a unified scale and coordinate system, and the model metadata (coordinate system convention, unit, density threshold) is recorded for downstream path and registration. The image fusion algorithm is used for reconstruction and fusion of the dental model to obtain a three-dimensional digital dental model. The CBCT data includes the shapes of the dental crown, dental root and alveolar bone; the intraoral scanning data is the surface of the teeth, mainly including the shapes of the dental crown and gingiva. The CBCT data and the intraoral scanning data are respectively imported into Mimics (medical image processing software) and 3Shape (intraoral scanning modeling software); the three-dimensional model reconstruction algorithm (such as rigid registration + surface registration) is used for three-dimensional model reconstruction and fusion to obtain a high-precision three-dimensional digital dental model containing dental roots, alveolar bones and dental crowns.

[0036] The intelligent path planning module 102 analyzes the tooth root and alveolar bone morphology based on the three-dimensional digital dental model using an AI algorithm, plans the optimal implant path, and generates a heat map score. As shown in Figure 6 According to the surgical supplementary information, i.e., the alveolar bone thickness and the inter-root distance, and the surgical requirements, i.e., the surgical accessibility and the micro-implant nail implantation angle requirements, a neural network or a heuristic algorithm based on anatomical rules is used to plan the optimal implant path vector including the position and direction, and to visualize it as an implant feasibility score heat map. The heat map score is mapped in the alveolar region, and the data format of the path data and the heat map data is JSON or 3D grid format.

[0037] Based on the fused three-dimensional digital dental model, the following calculations are performed:

[0038] The tooth root distance map + bone thickness map, combined with the approach angle requirements, tooth root / vessel exclusion zones, form a surgical accessibility score;

[0039] A neural network or a heuristic algorithm based on anatomical rules is used to generate the optimal implant path and the candidate path set;

[0040] The score is mapped to a heat map visualization layer and exported as JSON / 3D grid along with the path for rendering and interactive fine-tuning.

[0041] The registration and virtual-real registration module 103 uses at least one registration anchor to complete the virtual-real coordinate alignment. The registration anchor is preferably a incisal edge registration marker clip clamped in the incisal edge area of the maxillary central incisor. The top is provided with a standard image recognition pattern for MR device image recognition and pose solving.

[0042] As shown in Figure 4 and Figure 5 The incisal edge registration marker clip is used as at least one anchor point, and a real-time image stream is obtained through the MR device. AprilTag / checkerboard recognition (such as OpenCV / ARToolkit) is used to extract pattern corner points. EPnP+RANSAC is used to complete 6DoF pose solving to obtain a registration matrix, and Kabsch / ICP is used to complete virtual-real geometric fine registration to obtain a spatial transformation matrix. After TMR registration, the model is constructed in the MR space, and the "virtual dental model + AI path / heat map" is mapped to the real / camera coordinate system.

[0043] The end-cut registration mark clamp mentioned in the present application is a U-shaped structure, the clamp arms cover the buccal and lingual surfaces of the teeth and are embedded with anti-slip pads, the top is provided with an identifiable pattern platform, the pattern is an AprilTag or a checkerboard pattern, the pattern platform is raised by 3mm from the end-cut and is inclined forward by 10 degrees to improve the recognition accuracy. Its design can be stably clamped in the end-cut area of the maxillary central incisor, the top is provided with a standard image recognition pattern (such as a checkerboard pattern, an AprilTag, a two-dimensional code, etc.), which is suitable for three-dimensional model space registration in a mixed reality system. The clamp structure is light, symmetrical and reusable, and is compatible with different patients, different treatment methods (bracket correction, invisible correction), different surgical areas (buccal / lingual), and different oral opening states, and provides a reliable registration anchor point on the central axis of the dentition, which significantly improves the accuracy, efficiency and universality of the space registration of the MR / VR system.

[0044] Input virtual dental model + AI path, heat map (visual layer), MR device image stream (for recognition), the virtual dental model is obtained by collecting CBCT data and intraoral scanning data, and the AI path is the best implantation path obtained by the virtual dental model through an AI algorithm, so the virtual dental model serves as a "target model" for registration / registration and is rigidly aligned with the camera / world system obtained from the pattern of the mark clamp; and plays a superimposition and guidance role: after being mapped to the real coordinate through the alignment matrix, it serves as the main body of the intraoperative navigation superimposition display (optimal implantation path, reference guide line, depth scale, heat map safety zone / risk zone); the AI path and the heat map mainly play a safety and quality control role: the heat map / path is used for avoidance checking (tooth root, thin bone area) and intraoperative deviation evaluation after registration, and serves as a comparison reference to generate an accuracy report after surgery (data closed loop); these input data are not only the target of alignment, but also the content superimposed for the doctor to see, and also play the role of safety constraints and postoperative evaluation reference. At least one of the AprilTag, ArUco, checkerboard or two-dimensional code recognition algorithms (open source libraries such as OpenCV and ARToolkit) is used to detect the pattern on the top of the clamp; the pose registration algorithm (PnP: Perspective-n-Point, pose estimation through perspective imaging of n points) is performed through the recognized pattern pose; a registration matrix / space conversion matrix (4x4 transformation) and a real-time navigation view after mapping (virtual-real superimposition) are obtained, a registered model in the MR space is constructed, and the alignment of the physical model to the coordinate system of the virtual model is realized. In this embodiment, the PnP algorithm + ICP algorithm is used to complete the 6DoF (6 degrees of freedom) pose solution of the clamp pattern in space, and the virtual model is affinely registered to the real coordinate system. The PnP algorithm obtains the initial external parameter, i.e. the registration matrix, from the relationship between the 3D corners of the pattern and the 2D pixels of the image; then, the ICP / Kabsch algorithm is used to perform fine registration based on the geometric consistency of the real side point cloud / surface to the virtual model, and the space conversion matrix is obtained.

[0045] Where PnP (registration matrix) is the known camera intrinsic , pattern three-dimensional corner and pixel coordinate :

[0046]

[0047] Where K is the camera intrinsic matrix (3x3), including focal length, principal point, etc.

[0048] X i is the three-dimensional homogeneous coordinates of the pattern corner;

[0049] is the two-dimensional pixel coordinates (homogeneous form) of the corner on the image;

[0050] s i is the projection scale factor (perspective depth scaling);

[0051] R is the rotation matrix, representing the rigid body rotation;

[0052] t is the translation vector, representing the rigid body translation;

[0053] π() is the perspective projection operator, mapping three-dimensional points to two-dimensional pixel planes;

[0054] Get

[0055]

[0056] Where T pnp is the registration matrix (4x4 homogeneous rigid body transformation) obtained by the PnP algorithm;

[0057] ICP / Kabsch (space conversion matrix) is to make rigid alignment between the source point set and the target point set :

[0058]

[0059] Where p i is the point in the source point set (such as the camera observation point);

[0060] q i is the point in the target point set (such as the virtual model point);

[0061] σ(i) is the point correspondence, the source point p i corresponding to the target point index;

[0062] is the centroid of the source point set;

[0063] is the centroid of the target point set;

[0064] H is a covariance matrix, used for ICP solution;

[0065] is a singular value decomposition (SVD) of H;

[0066] R icp is a rotation matrix obtained by ICP;

[0067] t icp is a translation vector obtained by ICP;

[0068] T icp is a spatial transformation matrix (4x4) obtained by ICP;

[0069] The total transformation for superimposition is obtained by combination

[0070]

[0071] In the formula, T MR is the total transformation matrix finally used for MR superimposition;

[0072] Map the virtual side (dental model, optimal implant path, heat map, etc.) to the real / camera coordinates to achieve strict alignment.

[0073] The virtual-real superimposition of the navigation view is to superimpose the virtual dental model, the optimal implant path, and the reference guide line, etc. layers on the real scene after multiplying TMR before rendering, which can form a virtual-real superimposed view. The superimposed elements include the path, the angle / depth scale, the direction guide line, the heat map, etc.

[0074] The intraoperative navigation module 104 superimposes the implant path and the spatial reference guide line according to the MR space registered model in real time, assists the doctor in controlling the direction, depth and angle, and supports the comparison and correction of the implant tool posture. During the navigation process, the system updates the virtual path superimposed position in real time by tracking the MR device position change. Through the patient posture sensor such as the MR device camera or the inertial measurement unit IMU, the real-time position change of the user's head is collected. The purpose of collection is to convert the doctor's head posture change at each moment into stable and accurate MR superimposition and quantitative feedback.

[0075] Real-time rendering in HoloLens / Meta (including Quest 3 color perspective) on Unity or Unreal Engine: three-dimensional reconstruction model, optimal implantation path, spatial reference guide line and depth scale line / heat map layer; receive registration results and path data through Socket protocol or local API interface, and continuously collect head pose changes of MR device camera / IMU to update the superimposed view; support comparison with implant tool posture, output intraoperative visual navigation prompts such as direction / angle / depth correction prompts, and real-time deviation value and calibration feedback. Gesture / voice triggered switching and fine tuning are completed by the user interaction and tracking module (105), which is responsible for real-time updating of rendering and navigation guidance.

[0076] The auxiliary functions of this module specifically include:

[0077] Visual superposition: real-time superposition of optimal path, spatial reference guide line and scale (depth / direction) on real scene to ensure that the doctor can directly see the target approach and angle.

[0078] Quantitative feedback / correction: the system continuously collects the pose changes of the head-mounted camera or IMU to maintain alignment; and supports comparison with the implant tool posture (implant handle direction) to give direction / angle / depth deviation prompts to assist the doctor in correcting the operation (such as angle difference, entry and endpoint deviation, etc.).

[0079] Real-time update: with head position changes and scene motion, the navigation view and path superposition will be updated in real time to maintain virtual-real consistency and ensure effective guidance. Through posture comparison and deviation prompts, the system provides dynamic decision support for the doctor in terms of direction, angle and depth.

[0080] User interaction and tracking module 105 uses LeapMotion or voice control for model switching and path adjustment, real-time tracking of head position or clip offset to maintain registration consistency. The system supports cooperation with the implant tool to achieve direction and depth correction prompts by comparing its posture with the recommended path. Specifically, it includes: real-time use of spatial tracking and posture update algorithms (such as ICP, Kalman filter) to dynamically adjust the navigation interface; compare the implant handle direction with the preset path angle and provide deviation prompts through an angle difference feedback algorithm; display depth scale, direction guide line, heat map area marker and other information. Intraoperative navigation errors are recorded and used for postoperative path trajectory comparison and precision analysis.

[0081] Model switching is to switch display layers / versions on the MR side, such as only the crown layer, only the tooth root / alveolar bone layer, the fusion model, different viewing angles / transparency / heat map superposition layers; it can also switch between multiple candidate path schemes, the purpose being to allow the operator to quickly compare and decide among different information views.

[0082] Path adjustment is based on the AI-generated optimal implant path (multiple candidates may exist), allowing local fine-tuning, such as: entry point micro-shift (in the direction of the alveolar ridge or gum margin); incident angle fine-tuning (± several degrees, subject to "minimum safety distance from the tooth root / thin bone area" and heat map threshold); target depth fine-tuning (within the bone thickness and anatomical safety window).

[0083] The interaction mode is LeapMotion gesture / voice triggered "fine-tuning / switching", and the system updates the path in real time under the threshold of heat map safety area and geometric constraints (not allowed to cross into the risk area), and immediately overlaps to the MR view.

[0084] After registering the model under the MR space (matrix has been established), the tool axis and path tangent vectors are compared in time series to obtain angle error, entry point position error, depth error, and lateral deviation:

[0085] The calculation formula of angle error (direction deviation) is:

[0086]

[0087] where, is the unit vector of the tool axis, is the unit tangent vector of the path;

[0088] Δθ(t) is the angle error: the included angle between the tool direction and the reference path direction, unit ° or radian;

[0089] n tool (t) is the axial unit vector (direction) of the tool at time t;

[0090] n path (s(t)) is the tangent unit vector of the reference path at arc length position s(t);

[0091] The calculation formula of entry point position error (lateral deviation, mm) is:

[0092]

[0093] where

[0094] p tool_entry is the tool entry point coordinate;

[0095] p path_entry is the reference path entry point coordinate;

[0096] e entry is the entry point position error: the spatial deviation of the entry point, unit mm;

[0097] The calculation formula of depth error (projection difference along the path tangent, mm) is:

[0098]

[0099] where p tool (t) is the spatial position coordinate of the tool tip at time t;

[0100] p path (s(t)) is the spatial position coordinate of the reference path at arc length position s(t);

[0101] e depth (t) is the depth error: the difference of the tool projection in the path tangential direction, unit mm;

[0102] The calculation formula of the lateral deviation (perpendicular to the path plane) is:

[0103]

[0104] e ⊥ (t) is the lateral deviation: the offset of the tool in the plane perpendicular to the path direction, unit mm;

[0105] I is a unit matrix;

[0106] s(t) is the path parameter, corresponding to the arc length position at time t.

[0107] The data closed loop and postoperative review module 106 automatically uploads the operation data to the cloud system, which can be used for postoperative precision analysis, case reproduction and resident teaching. According to the actual operation path record (compared with the AI path), the intraoperative deviation data, the navigation image snapshot and the spatial trajectory data, the comparison and analysis algorithm (such as trajectory similarity matching, angle / position error analysis) is used to perform postoperative precision analysis, and form the postoperative operation accuracy report and visual operation record.

[0108] In a specific embodiment of the application, as shown in Figure 2 and Figure 3 The incisal edge registration marker clip adopts a U-shaped structure design, and is formed by TPU or medical PLA 3D printing; the upper and lower arms are wrapped around the upper and lower edges of the central incisor incisal edge, respectively, and the inner wall of the clip body is provided with a silica gel or foam pad to enhance the friction force to prevent slipping. In order to adapt to the actual clinical situation, especially when the patient's teeth have been bonded with brackets or other accessories (such as invisible and bracketless appliance accessories), the incisal edge registration marker clip is designed to have a single-arm height (buccal side and lingual side) close to the entire crown height (about 9-11mm), extending from the incisal edge to the edge of the gum margin 0.5-1mm, and the clip arm height covers more than 2 / 3 of the crown; thereby obtaining sufficient clamping area and force arm length without affecting the brackets or accessories, and enhancing stability.

[0109] The top pattern recognition platform area is ≥8x8mm, raised about 3mm from the incisal edge, and tilted forward by 10°, used for installing image marks such as chessboard patterns or AprilTag, to improve the recognition accuracy of mixed reality navigation systems.

[0110] The clip body is 3D printed from elastic thermoplastic material (such as TPU), which can be elastically deformed to adapt to the tooth surface morphology and the presence of brackets, achieving the dual requirements of avoidance and stable clamping.

[0111] The inside of the clip arm is provided with an embedded silica gel film or foam pad to enhance friction and wearing comfort, and to prevent loosening or slipping during surgery.

[0112] The top of the incisal edge registration marker clip is designed as a flat or slightly forward tilted platform, used for direct embedding or printing of image recognition patterns, such as AprilTag pattern areas. The pattern platform is raised about 3mm from the incisal edge to avoid obstruction by structures such as veneers and brackets.

[0113] The central incisor is the standard site for adaptation, and it can also be compatible with lateral incisor / lower incisor combination clamping. During wearing, it does not need to block the buccal / tongue surface brackets or attachments, and it is suitable for both invisible orthodontic patients and traditional bracket orthodontic patients. After wearing, the head-mounted mixed reality device can quickly recognize the pattern to achieve virtual-real registration, mainly applied to orthodontic micro-implant navigation, bracket bonding navigation, and secondarily considered for implant surgery guidance, restoration preparation display, and patient preoperative demonstration in multiple scenarios. The specific innovative points of this universal incisal edge registration marker clip include:

[0114] (1) The first design of a registration anchor device based on central incisor incisal edge clamping, not dependent on the bite plate structure.

[0115] (2) Universal structure suitable for various oral operation scenarios, preoperative, intraoperative, and postoperative;

[0116] (3) Equipped with high-identification pattern platform (supporting AprilTag / chessboard patterns, etc.) and high-friction foam / silicone anti-slip inner lining;

[0117] (4) The clip body is 3D printed from elastic thermoplastic material (such as TPU), which can be elastically deformed to adapt to the tooth surface morphology and the presence of brackets, achieving the dual requirements of avoidance and stable clamping. It is convenient for production, replacement, disinfection and low-cost promotion;

[0118] (5) The clip body is elastic and not affected by brackets, suitable for fixed / invisible orthodontic systems;

[0119] (6) The pattern platform is tilted forward and raised to facilitate recognition and enhance pose recognition stability.

[0120] The orthodontic micro-implant AI-MR navigation and marker registration system adopts a registration and identification algorithm including AprilTag or chess grid + OpenCV image recognition, pose solution using EPnP + RANSAC noise filtering, and spatial registration using Kabsch or ICP method; the path planning includes implant path generation based on tooth root distance map + bone thickness analysis, calculation of accessibility score, output as a three-dimensional trajectory model with heat map superposition, and import into the Unity platform for rendering; the rendering and interaction includes real-time rendering of three-dimensional reconstruction model, navigation path, depth scale line and other contents in HoloLens / Meta device using Unity or Unreal Engine, interface communication using Socket protocol or local API integration, and support for gesture triggered navigation switching, voice positioning control and other functions.

[0121] The above-mentioned technology can realize preoperative modeling, intelligent path planning, registration and registration, intraoperative navigation and interactive feedback of the closed loop, and is suitable for orthodontic navigation, implant assistance, teaching demonstration and other scenes. The following are several specific embodiments of the application:

[0122] Example 1: After the patient receives CBCT and IOS oral scanning, the data is imported into 3Shape / Mimics fusion modeling, the AI module generates an implant path heat map, and the heat map is imported into the Unity platform and fused with the CT three-dimensional reconstruction model. During the operation, the cutting end recognition clip is worn, the pattern is recognized by the MR head-mounted device, and the virtual model is aligned. During the navigation process, the system real-time tracks the change of the visual angle, renders the dynamic guide path and angle calibration feedback.

[0123] Example 2: When the patient's mouth is limited or without brackets, the traditional bite plate cannot be installed, and the cutting end clip in the application can still stably hold the teeth without the need for occlusal state intervention, and is suitable for palatal or buccal navigation.

[0124] Example 3: For resident physician training, historical paths can be played back, multiple cases can be compared, and navigation errors and operation trajectories can be analyzed.

[0125] The cutting end registration marker clip has the advantages of light structure, stable position and high registration accuracy, and can be widely used in the following multiple oral digital navigation operation scenes: orthodontic scenes, implant scenes, extraoral / surgical scenes, teaching and display scenes.

[0126] For orthodontic scenes, the application direction is micro-screw anchor implant navigation and bracket bonding accuracy positioning. In the micro-screw anchor implant navigation, the MR navigation guides the direction and depth of the nail, which is suitable for buccal or palatal anchor nail implantation, and assists junior doctors in standard operation; in the bracket bonding accuracy positioning, the MR preview fitting site and attitude are previewed before the bracket is bonded on the tooth surface, which improves the bonding accuracy and reduces the error rate.

[0127] For planting scenarios, the application direction is intraoperative positioning of implants, alveolar ridge evaluation and virtual implant demonstration, and navigation assistance for transpterygoid and transpterygoid implantation. Intraoperative positioning of implants, MR navigation guides the implant into the designed path, especially suitable for single anterior tooth aesthetic area without guide; In the evaluation of alveolar ridge and virtual implant demonstration, the preoperative superposition of virtual implant model and anatomical structure is compared, which is convenient for patient communication and surgical planning verification; Navigation assistance for transpterygoid and transpterygoid implantation can guide the direction and depth of the implant through clip registration + MR navigation when the traditional guide cannot fully perform.

[0128] For extraoral / surgical scenarios, the application direction is buried tooth extraction guidance, tooth trauma reconstruction or replantation simulation navigation, and MR surgery demonstration or residency teaching. In the extraction guidance of buried teeth (such as maxillary canine / third molar), the CBCT image is superimposed with the virtual root path before surgery, and the MR navigation can determine the windowing and traction path; In the simulation navigation of tooth trauma reconstruction or replantation, the original anatomical position can be restored as a reference through the registration system when the dislocated tooth is replanted; In the MR surgery demonstration or residency teaching, the recognition clip + MR equipment can be used for simulation navigation teaching to improve spatial understanding and operation visibility.

[0129] For patient teaching and demonstration scenarios, the application direction is preoperative MR demonstration of patients, medical record display, and simulation reexamination, and oral popularization and teaching demonstration. In the preoperative MR demonstration of patients, the clip is worn to superimpose the three-dimensional navigation image of the personalized surgical plan, which is convenient for preoperative communication and explanation of the consent form; In the simulation navigation of tooth trauma reconstruction or replantation, the MR system loads the historical model and locates the intraoral position through the recognition clip, which can track and follow up the tooth movement or repair progress; In the MR surgery demonstration or residency teaching, the three-dimensional interactive demonstration registration mechanism and clinical surgical path are used in exhibitions, in-hospital propaganda or digital courses in medical colleges.

[0130] The application also provides an orthodontic micro-implant AI-MR navigation and marker clip registration method, which comprises:

[0131] CBCT data and intraoral scanning data are collected, and a three-dimensional digital dental model is constructed using an image fusion algorithm;

[0132] According to the three-dimensional digital dental model, an AI algorithm is used to analyze the shape of the tooth root and alveolar bone, and an optimal implantation path is planned;

[0133] Use the incisal registration marker clip as at least one registration anchor, and clamp it on the incisal end of the incisor (preferably the incisal end of the maxillary central incisor). The standard image recognition pattern set on the top of the incisal registration marker clip is recognized by the MR head-mounted device, the three-dimensional pose is recognized, and the MR space registration model is constructed;

[0134] According to the MR space registration post-model, real-time tracking of MR head-mounted device position changes and updating of virtual path superposition positions are performed for navigation guidance throughout the entire implantation process.

[0135] It should be noted that, for the foregoing method embodiments, in order to simply describe, they are all expressed as a series of action combinations, but those skilled in the art should know that the present application is not limited to the action sequence described, because according to the present application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily necessary for the present application.

[0136] The present application also provides a storage medium, which is a computer readable storage medium, and a computer program is stored on the storage medium, and the computer program can execute part or all steps of the method shown in the specification when running. The storage medium can include a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc. The storage medium can also include a non-volatile memory or a non-transitory memory, etc.

[0137] The above embodiments can be realized all or partially by software, hardware, firmware or other any combination. When realized by software, the above embodiments can be realized all or partially in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network or other programmable device. The computer instructions can be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, the computer instructions can be transmitted from one website, computer, server or data provider to another website, computer, server or data provider through wired or wireless manner.

[0138] The above has carried out the detailed introduction to the embodiment of the application, the application has carried on the elaboration to the application in this article, the above example explanation is only for helping understanding the method and system of the application, it is only a part of the embodiment of the application, and is not all the embodiment. Based on the embodiment in the application, all other embodiments obtained by the ordinary skilled person in the art without making creative labor should belong to the protection scope of the application, and the content of the description should not be understood as limiting the application. Therefore, any modification, equivalent replacement, improvement, etc. within the spirit and principles of the application should be included in the protection scope of the application.

Claims

1. An orthodontic micro-implant AI-MR navigation and marker clip registration system, characterized in that, The system comprises: a data fusion module that collects cone-beam computed tomography data and intraoral scanning data, uses an image fusion algorithm to construct a three-dimensional digital dental model containing crowns, roots, and alveolar bone; an intelligent path planning module that analyzes the morphology of roots and alveolar bone based on the three-dimensional digital dental model, and generates path data containing heat map scores and optimal implantation paths; a registration and virtual-real registration module that uses at least one registration anchor to complete virtual-real coordinate alignment, the registration anchor being a incisal edge registration marker clamp clamped in the incisal edge region of the maxillary central incisor, a standard image recognition pattern provided on the top of the incisal edge registration marker clamp being recognized by a mixed reality (MR) device, and the 6 degrees of freedom (DoF) pose of the standard image recognition pattern being estimated to obtain a registration matrix and a spatial conversion matrix, and the path data and the three-dimensional digital dental model being mapped to a real coordinate system; an intraoperative navigation module that superimposes the path data and reference guide information in real time in a real scene based on the MR device, and compares with the implant tool pose to obtain correction prompts for output direction, angle, and depth; a data closed loop module for recording intraoperative deviations and implant tool trajectories, and performing postoperative precision analysis and visual report generation.

2. The orthodontic micro-implant Al-MR navigation and landmark clip registration system according to claim 1, wherein, The cone-beam computed tomography data collected by the data fusion module includes crown, root, and alveolar bone morphology, and the intraoral scanning data collected by the intraoral scanner includes tooth surface, i.e., crown and gingival morphology; the data fusion module performs medical image processing on the cone-beam computed tomography data, performs oral scanning modeling on the intraoral scanning data, uses an image fusion algorithm to reconstruct and fuse a three-dimensional model, and obtains the three-dimensional digital dental model including roots, alveolar bone, and crowns in a unified scale and coordinate system.

3. The orthodontic micro-implant AI-MR navigation and landmark clip registration system according to claim 1, wherein, The intelligent path planning module uses a neural network or a heuristic algorithm based on anatomical rules to plan an optimal implantation path including position and direction and visualize it as a heat map score distribution based on alveolar bone thickness, interradicular distance measurement, and requirements for surgical accessibility and micro-implant implantation angle.

4. The orthodontic micro-implant Al-MR navigation and landmark clip registration system according to claim 3, wherein, The intelligent path planning module supports interactive fine-tuning of entry point, incidence angle, and target depth, and is constrained by safety thresholds. When the thresholds are exceeded, a prompt is given and updating is refused.

5. The orthodontic micro-implant AI-MR navigation and landmark clip registration system according to claim 3, wherein, The registration and virtual-real registration module uses at least one of AprilTag, ArUco, checkerboard, or two-dimensional code recognition algorithms to recognize the standard image recognition pattern, uses PnP algorithm + ICP algorithm to complete 6DoF pose solving of the standard image recognition pattern in space, completes three-dimensional pose recognition, obtains a registration matrix and a spatial conversion matrix, and a real-time navigation view of virtual-real superposition after mapping, constructs an MR space registered model, and completes alignment of the physical model to the virtual model coordinate system.

6. The orthodontic micro-implant Al-MR navigation and landmark clip registration system according to claim 5, wherein, The intraoperative navigation module uses the MR space registered model, and the MR device overlays virtual paths and spatial reference guide lines in real time to assist doctors in controlling direction, depth and angle, and supports comparison and correction with the posture of the implantation tool; the MR device camera or inertial measurement unit (IMU) collects real-time changes in the user's head position.

7. The orthodontic micro-implant Al-MR navigation and landmark clip registration system according to claim 1, wherein, The system also includes a user interaction and tracking module, which uses LeapMotion gesture recognition devices or voice control for model switching and path adjustment, tracks the offset of the head position or the incision registration marker clip in real time to maintain registration consistency, and works with the implantation tool to provide direction and depth correction prompts by comparing the posture of the implantation tool with the virtual path. Specifically, this includes: dynamically adjusting the navigation interface in real time using spatial tracking and posture update algorithms; comparing the direction of the implantation tool with the preset path angle and providing deviation prompts through an angle difference feedback algorithm; displaying depth scales, direction guide lines, and heat map area marking information; and recording intraoperative navigation errors for postoperative path trajectory comparison and accuracy analysis.

8. An incisal registration marker clip for oral digitalization navigation, characterized in that, The incisal edge registration marker clip is used in the orthodontic micro-implant AI-MR navigation and marker clip registration system as described in claim 1. The clip body is a U-shaped clamping structure, with its buccal and lingual clamping arms used to cover the upper and lower edges of the incisal edge of the tooth and equipped with anti-slip pads. The anti-slip pads are made of silicone or foam and can be replaced once or disinfected and reused. A pattern recognition platform is set on the top of the clip body to install or integrally form coded planar markers. The clip body material is elastic to adapt to different crown shapes and the presence of brackets or accessories. The pattern recognition platform is raised relative to the incisal edge and has a forward tilt angle to improve the visibility and posture resolution stability of the machine-readable pattern.

9. The resected registration landmark clip for oral digitalization navigation of claim 8, wherein, The pattern recognition platform is raised 2–5 mm from the cut end and tilted forward 5–15°. The effective recognition area of ​​the pattern recognition platform has a side length of ≥8 mm.

10. A method of path planning for an orthodontic micro-implant, characterized in that, The method includes: constructing a three-dimensional digital dental model; calculating root distance maps and bone thickness maps and generating feasible regions by combining restricted structures and surgical accessibility constraints; outputting a candidate path set and scoring heatmap using a neural network or an anatomical rule-based heuristic algorithm; and interactively fine-tuning the entry point, incident angle, and target depth under threshold constraints and outputting the optimal implantation path.

Images