Oral and maxillofacial surgery registration method and multi-scale registration auxiliary device

Abstract

The application discloses an oral and maxillofacial surgery registration method and a multi-scale registration auxiliary device, relates to the technical field of medical surgery navigation, and integrates spatial information from different scales and sources by using the multi-scale registration auxiliary device, realizes high-precision and robust registration of fine internal oral anatomical structures, and can further quickly and accurately superimpose and register virtual information such as a preoperative planning surgery path and key tissue structures into a real maxillofacial surgery field of a patient, so as to provide intuitive and accurate intraoperative augmented reality navigation for doctors.

Description

Technical Field

[0001] This invention relates to the field of medical surgical navigation technology, and in particular to a multi-scale registration method for oral and maxillofacial surgery and a multi-scale registration auxiliary device. Background Technology

[0002] The application of augmented reality in surgical navigation in medical scenarios is showing a continuous growth trend. By overlaying virtual models of preoperative planned paths, important organs and tissues onto the real environment in real time, it can expand the doctor's field of vision during surgery and provide the doctor with intraoperative visual guidance.

[0003] Augmented reality requires precise alignment between virtual models and targets in real-world scenes, which involves the registration of the three-dimensional structure of patient tissues in virtual and real spaces. Currently, augmented reality navigation systems used in medicine mainly rely on optical devices to acquire surface point clouds of patient organs, and then register the three-dimensional models of patient organs segmented from CT scans onto the acquired surface point clouds to achieve the augmented reality scene.

[0004] However, in oral and maxillofacial surgery, these external optical devices mainly rely on point cloud information of the body surface or external visible areas, making it difficult to obtain high-precision three-dimensional structures of the obscured oral cavity, thus limiting their application in corresponding surgeries.

[0005] On the other hand, although oral scanners can reconstruct the three-dimensional structure inside the oral cavity with high precision, their field of view is limited to the inside of the oral cavity and they lack a direct correlation channel with the absolute spatial pose, thus making it difficult to provide reliable intraoperative navigation and limiting clinical application scenarios. Summary of the Invention

[0006] In view of the above problems, the present invention provides a method for registration in oral and maxillofacial surgery and a multi-scale registration auxiliary device for overcoming or at least partially solving the above problems.

[0007] This invention provides the following solution: A method for registration in oral and maxillofacial surgery includes: S1: Acquire CT image data of the patient's maxillofacial region, and perform dentition segmentation and three-dimensional reconstruction on the CT images to obtain a CT dentition model; S2: The multi-scale registration auxiliary device is fixedly installed at the corresponding position of the patient's oral and maxillofacial surface. The multi-scale registration auxiliary device includes an intraoral imaging reference part, an extraoral tracking reference part, and a rigid connection structure for connecting the intraoral imaging reference part and the extraoral tracking reference part. S3: Use an oral scanner to scan the target dentition area where the intraoral imaging reference unit is installed, and obtain intraoral scanning point cloud data including the intraoral imaging reference unit and the actual dentition. S4: Process the intraoral scanning point cloud data to segment it into the real dental arch point cloud and the intraoral imaging reference point cloud; S5: Register the CT dental arch model with the real dental arch point cloud to obtain the first transformation matrix; S6: Register the point cloud of the intraoral imaging reference section with its preset three-dimensional model to obtain the second transformation matrix; S7: Real-time tracking to obtain the real-time pose of the extraoral tracking reference part in the world coordinate system; based on the known and fixed spatial geometric relationship between the extraoral tracking reference part and the intraoral imaging reference part determined by the rigid connection structure, combined with the first transformation matrix and the second transformation matrix, calculate the final transformation relationship for registering the CT dental arch model to the world coordinate system; S8: Using the final transformation relationship, the CT dental arch model is spatially mapped and displayed in real time with the patient's actual anatomical structure in the navigation system.

[0008] Preferably, in step S2, the intraoral imaging reference portion is rigidly fixed to the patient's dentition by means of bonding, mechanical snapping, or dental resin fixation.

[0009] Preferably, in step S4, the intraoral scan point cloud data is segmented using a neural network.

[0010] Preferably, step S4 specifically includes: The intraoral scan point cloud data is dimensionality-reduced and mapped to a 2D plane using the UV unfolding algorithm to construct an oral cavity unfolding map; The features on the oral cavity unfolding diagram are identified by a neural network, and the 2D regions corresponding to the real dentition and the intraoral imaging reference part are segmented to obtain the actual dentition. The 2D region is inversely transformed to a 3D point cloud space to obtain the segmented real dental arch point cloud and the intraoral imaging reference point cloud.

[0011] Preferably, in step S5 and / or step S6, the registration adopts an algorithm that cascades coarse point cloud registration and fine point cloud registration.

[0012] Preferably, the coarse registration algorithm for point clouds includes the Super4PCS algorithm, and the fine registration algorithm for point clouds includes the GICP algorithm.

[0013] Preferably, in step S7, the pose of the external tracking reference part is obtained in real time by an external optical positioning system.

[0014] Preferably, in step S8, the CT image is superimposed onto the 2D image using the final transformation relationship to construct an augmented reality scene to guide the surgical procedure.

[0015] A multi-scale registration aid for implementing the above-mentioned registration method in oral and maxillofacial surgery, the device comprising: An external tracking reference unit has a first optical configuration, which is configured to be recognized and tracked by an external optical positioning system. The intraoral imaging reference section has a second optical configuration, which is configured to be imaged by an oral scanner and form a feature point cloud. A rigid connection structure has a preset rigid geometry. The first end of the rigid connection structure is fixedly connected to the external tracking reference part, and the second end of the rigid connection structure is fixedly connected to the intraoral imaging reference part, thereby establishing a known and fixed spatial geometric relationship between the external tracking reference part and the intraoral imaging reference part.

[0016] Preferably, the geometric parameters of the rigid connection structure are precisely calibrated so that the average tracking error between the external tracking reference part and the intraoral imaging reference part is no greater than 0.5 mm.

[0017] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects: This application provides a method for registering oral and maxillofacial surgical procedures and a multi-scale registration auxiliary device. By utilizing the multi-scale registration auxiliary device, spatial information from different scales and sources is integrated to achieve high-precision and robust registration of fine anatomical structures inside the oral cavity. This enables the rapid and accurate overlay and registration of virtual information such as the preoperatively planned surgical path and key tissue structures onto the patient's real maxillofacial surgical field, providing doctors with intuitive and precise intraoperative augmented reality navigation.

[0018] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0019] 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. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0020] Figure 1 This is a flowchart of a method for registration in oral and maxillofacial surgery provided by an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of the intraoral tracking reference section provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of the external tracking reference section provided in an embodiment of the present invention; Figure 4This is a schematic diagram of the rigid connection structure provided in an embodiment of the present invention.

[0021] In the figure: intraoral imaging reference part 10, extraoral tracking reference part 20, rigid connection structure 30. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention are within the scope of protection of the present invention.

[0023] See Figure 1 This invention provides a method for registering patients in oral and maxillofacial surgery, as exemplified by [example of the invention]. Figure 1 As shown, the method may include the following steps: S1: Acquire CT image data of the patient's maxillofacial region, and perform dentition segmentation and 3D reconstruction on the CT images to obtain a CT dentition model; perform thin-slice CT scans of the patient's maxillofacial region. Using medical image processing software (such as ITK-SNAP, Mimics) or deep learning algorithms, segment 3D models of the maxilla, mandible, and dentition. For navigation, focus is placed on the dentition model, as its surface morphology will serve as a reliable basis for subsequent registration. Simultaneously, complete virtual surgical planning (such as implant placement and osteotomy lines) on the CT data.

[0024] S2: The multi-scale registration auxiliary device is fixedly installed at the corresponding position on the patient's oral and maxillofacial surface. The multi-scale registration auxiliary device includes an intraoral imaging reference part 10, an extraoral tracking reference part 20, and a rigid connection structure 30 for connecting the intraoral imaging reference part 10 and the extraoral tracking reference part 20. In specific implementation, the embodiments of this application can provide that the intraoral imaging reference part 10 is rigidly fixed to the patient's dentition by means of bonding, mechanical snap-fit, or dental resin fixation.

[0025] For example, the multi-scale registration aid can be rigidly fixed to the patient's dentition by bonding. The geometric parameters of the rigid connection structure 30 are precisely calibrated, and the average tracking error between the extraoral tracking reference part 20 and the intraoral imaging reference part 10 is ≤0.5mm.

[0026] S3: Use an oral scanner to scan the target dentition area where the intraoral imaging reference unit 10 is installed, and obtain intraoral scanning point cloud data including the intraoral imaging reference unit 10 and the real dentition; In practical applications, use an oral scanner to scan the target dentition area where the device is installed (usually the part that overlaps with the CT dentition model, such as the occlusal surface of the crown, buccal and lingual surfaces), and obtain intraoral scanning point cloud data including the intraoral imaging reference unit 10 of the registration aid device and the real dentition.

[0027] When scanning the target dentition region where the intraoral imaging reference unit 10 is installed using an oral scanner, the obtained point cloud contains complete and clear color intraoral scanning point cloud data of the intraoral imaging reference unit 10 of the registration aid device and the actual dentition.

[0028] S4: Process the intraoral scanning point cloud data to segment it into a real dentition point cloud and an intraoral imaging reference point cloud 10; In specific implementation, embodiments of this application may provide a neural network for segmenting the intraoral scanning point cloud data.

[0029] The preferred approach is neural network-based segmentation. A U-Net-like network is trained, with input including various view renderings of the point cloud or multi-channel feature maps. The output is a prediction of whether each point belongs to "teeth," "baseline," or "background" (such as gums). To improve efficiency, the 3D point cloud can first be parametrically unfolded and mapped to a 2D planar image. Fast semantic segmentation using a CNN is then performed in the image domain, and the segmentation results are mapped back to 3D space to obtain preliminary point cloud labels.

[0030] Furthermore, the intraoral scan point cloud data is dimensionality-reduced and mapped to a 2D plane using a UV unfolding algorithm to construct an oral cavity unfolding map; The features on the oral cavity unfolding diagram are identified by a neural network, and the 2D regions corresponding to the real dentition and the intraoral imaging reference part 10 are segmented to obtain the actual dentition. The 2D region is inversely transformed to a 3D point cloud space to obtain the segmented real dental arch point cloud and the intraoral imaging reference 10 point cloud.

[0031] Segmentation can also be based on geometric features. Since the design of the intraoral imaging reference unit 10 has known geometric priors (such as size range and curvature features), it can be extracted from the scene using region growing, clustering (such as Euclidean clustering) combined with shape recognition algorithms (such as RANSAC fitting of basic geometry). In the remaining part, teeth and soft tissues can be further separated by features such as curvature and color.

[0032] S5: Register the CT dental arch model with the real dental arch point cloud to obtain the first transformation matrix; the registration adopts an algorithm that cascades coarse point cloud registration and fine point cloud registration.

[0033] The goal of this step is to obtain the transformation matrix T1 that best aligns the CT dental arch model with the real dental arch point cloud. This can be achieved using a cascaded combination of coarse and fine point cloud registration algorithms, or alternative algorithms with equivalent functionality. The transformation relationship between the CT dental arch model and the real dental arch point cloud is accurately calculated to obtain the first transformation matrix (T1).

[0034] In specific implementation, for example, the Super4PCS point cloud registration algorithm can be used to roughly estimate the transformation relationship between the CT dental arch model and the real dental arch point cloud. After testing, it is expected that the registration convergence can be ensured when the overlap rate is 0.8. The GICP point cloud registration algorithm is used to iteratively and accurately calculate the transformation relationship between the CT dental arch model and the real dental arch point cloud to obtain the first transformation matrix (T1).

[0035] Coarse registration: Since the initial positions of the two may differ significantly, a fast global registration algorithm is first used, such as Super4PCS (an accelerated version of the 4-point consistency set algorithm) or FPFH features + RANSAC. The goal at this stage is to find a preliminary alignment with sufficient overlap; extremely high precision is not required.

[0036] Fine registration: Based on the coarse registration results, the iterative nearest point algorithm and its variants are used for optimization. The GICP (Generalized ICP) algorithm considers the distribution of local surface features while minimizing point-to-point distances, and has better robustness to noise and partial overlap, making it the preferred solution. Iteration continues until convergence (the change in transformation parameters is less than the threshold or the error no longer decreases significantly), outputting a high-precision T1. At this point, the CT dental arch model has been transformed to the "oral scan local coordinate system" where the real dental arch is located through T1.

[0037] S6: Register the point cloud of the intraoral imaging reference section 10 with its preset three-dimensional model to obtain the second transformation matrix; similarly, a coarse point cloud registration algorithm, a fine point cloud registration algorithm, or an alternative algorithm with equivalent function can be used in a cascaded combination. Accurately calculate the transformation relationship between the point cloud of the intraoral imaging reference section 10 and its preset three-dimensional model to obtain the second transformation matrix (T2).

[0038] The coarse registration + fine registration process can also be used. However, since the 10 point cloud of the intraoral imaging reference part is acquired as part of the device, its initial pose relative to the CT dental arch model can be roughly estimated by the device installation position, and sometimes even the coarse pose can be calculated directly from the point cloud identification features. Therefore, the coarse registration step may be simplified or omitted.

[0039] After precise registration, T2 is obtained. This transformation establishes the relationship between the coordinate system of the actual dental arch point cloud (i.e., the coordinate system of the point cloud of the intraoral imaging reference section 10) and the coordinate system of the CAD model of the intraoral imaging reference section 10 itself.

[0040] For example, in a specific implementation, the Super4PCS point cloud registration algorithm can be used to roughly estimate the transformation relationship between the 10 point clouds of the intraoral imaging reference section and its preset three-dimensional model. After testing, it is expected that the registration convergence can be ensured when the overlap rate is 0.2. The GICP point cloud registration algorithm is used to iteratively and accurately calculate the transformation relationship between the 10 point clouds of the intraoral imaging reference section and its preset three-dimensional model to obtain the second transformation matrix (T2).

[0041] S7: Real-time tracking to obtain the real-time pose of the extraoral tracking reference unit 20 in the world coordinate system; based on the known and fixed spatial geometric relationship between the extraoral tracking reference unit 20 and the intraoral imaging reference unit 10 determined by the rigid connection structure 30, combined with the first transformation matrix and the second transformation matrix, calculate the final transformation relationship for registering the CT dental model to the world coordinate system. In practice, the pose of the extraoral tracking reference unit 20 can be obtained in real time through an extraoral optical positioning system. Specifically, using an extraoral optical positioning system, or a similar positioning device that can achieve the same function, the real-time pose (P) of the extraoral tracking reference unit 20 in the world coordinate system is obtained in real time; based on the known and fixed spatial geometric relationships within the multi-scale registration auxiliary device, the first transformation matrix, and the second transformation matrix, the final transformation relationship for registering the CT dental arch model to the world coordinate system is calculated.

[0042] S8: Using the final transformation relationship, the CT dental arch model is spatially mapped and displayed in real time with the patient's actual anatomical structure in the navigation system. Specifically, embodiments of this application can provide the use of the final transformation relationship to overlay CT images onto 2D images to construct an augmented reality scene to guide surgical procedures.

[0043] The navigation system acquires a real-time video stream of the surgical area (from the navigation system's built-in camera or the surgeon's head-mounted AR device camera). Appropriate tools can be used to render all pre-operatively planned virtual models (such as the mandibular canal, implants, osteotomy planes, etc.) to their corresponding spatial positions in the current video frame. During rendering, visual effects that blend virtual and real elements must be considered, such as semi-transparency, highlighting of outlines, and proper handling of depth occlusion. The surgeon can then see the virtual surgical plan "fitted" onto the real patient's anatomy through a monitor or AR glasses, thereby precisely guiding the angle, depth, and extent of surgical instrument insertion.

[0044] The oral and maxillofacial surgical registration method provided by this invention can be widely applied to various oral and maxillofacial surgeries requiring high-precision navigation, including but not limited to: dental implant surgery, orthognathic surgery (Le Fort I osteotomy, sagittal split osteotomy, etc.), maxillofacial trauma repair surgery, tumor resection, temporomandibular joint surgery, etc. It can be integrated with existing commercial surgical navigation systems (such as Brainlab, Stryker, Medtronic, etc.) to enhance their functionality in this specialty.

[0045] As can be seen, the oral and maxillofacial surgical registration method provided in this application, by utilizing a multi-scale registration auxiliary device, can quickly and accurately overlay CT images and preoperative surgical plans onto the patient's maxillofacial region to guide the surgeon, expanding the surgeon's field of vision and enabling the surgeon to complete the surgical procedure quickly. "Multi-scale registration" refers to a spatial registration technique that integrates spatial positioning, dental microstructures, and medical imaging features.

[0046] In summary, the oral and maxillofacial surgical registration method provided in this application utilizes the multi-scale registration auxiliary device to integrate spatial information from different scales and sources, achieving high-precision and robust registration of fine anatomical structures inside the oral cavity. This enables the rapid and accurate overlay and registration of virtual information such as the preoperatively planned surgical path and key tissue structures onto the patient's real maxillofacial surgical field, providing doctors with intuitive and precise intraoperative augmented reality navigation.

[0047] The core of this application lies in constructing a seamless spatial transformation chain: CT coordinate system → real dental arch coordinate system → intraoral imaging reference system 10 coordinate system → extraoral tracking reference system 20 coordinate system → world coordinate system (optical tracking system coordinate system). The provided multi-scale registration auxiliary device is the key physical carrier connecting the latter three coordinate systems.

[0048] This device forms the physical basis for the high-precision spatial bridging achieved in this invention, and its design must balance mechanical precision, biocompatibility, intraoperative stability, and manufacturability. Therefore, embodiments of this application may also provide a multi-scale registration auxiliary device for implementing the aforementioned oral and maxillofacial surgical registration method, comprising: The extraoral tracking reference unit 20 has a first optical configuration configured to be recognized and tracked by the extraoral optical positioning system to provide the real-time pose of the device in space. Its core is the first optical configuration, typically a rigid array of three or more non-collinearly arranged passive spherical reflectors or active infrared light-emitting diodes (IREDs). This first optical configuration must meet the recognition requirements of the optical positioning system used and have a sufficient spatial angle to ensure that at least some marker points are visible at any angle within the surgical field. Its base is securely connected to the rigid connection structure 30.

[0049] The intraoral imaging reference unit 10 has a second optical configuration, which is configured to be imaged by an oral scanner and form a feature point cloud. The core of this configuration is the second optical configuration, where "optical" mainly refers to the response characteristics to the light source (such as blue light or white light) of the oral scanner. The design principle is to enable it to generate unique, stable features in the oral scanner point cloud that are easy for algorithms to automatically recognize and segment.

[0050] The rigid connection structure 30 has a preset rigid geometry. A first end of the rigid connection structure 30 is fixedly connected to the external tracking reference part 20, and a second end of the rigid connection structure 30 is fixedly connected to the intraoral imaging reference part 10, thereby establishing a known and fixed spatial geometric relationship between the external tracking reference part 20 and the intraoral imaging reference part 10. In specific implementations, embodiments of this application can provide that the geometric parameters of the rigid connection structure 30 are precisely calibrated so that the average tracking error between the external tracking reference part 20 and the intraoral imaging reference part 10 is no greater than 0.5 mm.

[0051] The rigid connection structure 30 is the "skeleton" of the device, and its rigidity and dimensional stability directly determine the accuracy of the spatial relationship between the extraoral and intraoral reference parts. It is usually made of medical-grade stainless steel, titanium alloy, or high-performance engineering plastics (such as PEEK). Its shape can be designed as a rod, arc, or more complex shape as needed to bypass the lips and cheeks, adapt to the oral cavity space, and ensure that the extraoral parts are within the optimal field of view of the optical tracking system.

[0052] After manufacturing, the transformation relationship from the center of the reflective sphere of the external tracking reference unit 20 to the feature geometric center (or a specific set of feature points) of the internal imaging reference unit 10 can be precisely calibrated using a high-precision coordinate measuring machine or a laser tracker, and this fixed transformation matrix is ​​input into the navigation system as a known constant. In a preferred embodiment, the overall calibration error of the entire device should be controlled below 0.5 mm.

[0053] Before surgery begins, the sterilized multi-scale registration aid is securely fixed to the opposing or adjacent stable dentition in the surgical area via a connecting base. For example, during mandibular surgery, the device can be fixed to the maxillary dentition to avoid interfering with the surgical area.

[0054] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0055] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of this application.

[0056] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for system or system embodiments, since they are basically similar to method embodiments, the description is relatively simple, and relevant parts can be referred to the descriptions in the method embodiments. The systems and system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0057] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A method for registration in oral and maxillofacial surgery, characterized in that, The method includes the following steps: S1: Acquire CT image data of the patient's maxillofacial region, and perform dentition segmentation and three-dimensional reconstruction on the CT images to obtain a CT dentition model; S2: The multi-scale registration auxiliary device is fixedly installed at the corresponding position of the patient's oral and maxillofacial surface. The multi-scale registration auxiliary device includes an intraoral imaging reference part, an extraoral tracking reference part, and a rigid connection structure for connecting the intraoral imaging reference part and the extraoral tracking reference part. S3: Use an oral scanner to scan the target dentition area where the intraoral imaging reference unit is installed, and obtain intraoral scanning point cloud data including the intraoral imaging reference unit and the actual dentition. S4: Process the intraoral scanning point cloud data to segment it into the real dental arch point cloud and the intraoral imaging reference point cloud; S5: Register the CT dental arch model with the real dental arch point cloud to obtain the first transformation matrix; S6: Register the point cloud of the intraoral imaging reference section with its preset three-dimensional model to obtain the second transformation matrix; S7: Real-time tracking to obtain the real-time pose of the extraoral tracking reference part in the world coordinate system; based on the known and fixed spatial geometric relationship between the extraoral tracking reference part and the intraoral imaging reference part determined by the rigid connection structure, combined with the first transformation matrix and the second transformation matrix, calculate the final transformation relationship for registering the CT dental arch model to the world coordinate system; S8: Using the final transformation relationship, the CT dental arch model is spatially mapped and displayed in real time with the patient's actual anatomical structure in the navigation system.

2. The oral and maxillofacial surgical registration method according to claim 1, characterized in that, In step S2, the intraoral imaging reference part is rigidly fixed to the patient's dentition by means of bonding, mechanical snapping or dental resin fixation.

3. The oral and maxillofacial surgical registration method according to claim 1, characterized in that, In step S4, the intraoral scan point cloud data is segmented using a neural network.

4. The oral and maxillofacial surgical registration method according to claim 3, characterized in that, Step S4 specifically includes: The intraoral scan point cloud data is dimensionality-reduced and mapped to a 2D plane using the UV unfolding algorithm to construct an oral cavity unfolding map; The features on the oral cavity unfolding diagram are identified by a neural network, and the 2D regions corresponding to the real dentition and the intraoral imaging reference part are segmented to obtain the actual dentition. The 2D region is inversely transformed to a 3D point cloud space to obtain the segmented real dental arch point cloud and the intraoral imaging reference point cloud.

5. The oral and maxillofacial surgical registration method according to claim 1, characterized in that, In step S5 and / or step S6, the registration adopts an algorithm that cascades coarse point cloud registration and fine point cloud registration.

6. The oral and maxillofacial surgical registration method according to claim 5, characterized in that, The coarse point cloud registration algorithm includes the Super4PCS algorithm, and the fine point cloud registration algorithm includes the GICP algorithm.

7. The oral and maxillofacial surgical registration method according to claim 1, characterized in that, In step S7, the pose of the external tracking reference part is obtained in real time by an external optical positioning system.

8. The oral and maxillofacial surgical registration method according to claim 1, characterized in that, In step S8, the CT images are superimposed onto the 2D images using the final transformation relationship to construct an augmented reality scene to guide the surgical procedure.

9. A multi-scale registration auxiliary device for implementing the oral and maxillofacial surgical registration method according to any one of claims 1-8, characterized in that, The device includes: An external tracking reference unit has a first optical configuration, which is configured to be recognized and tracked by an external optical positioning system. The intraoral imaging reference section has a second optical configuration, which is configured to be imaged by an oral scanner and form a feature point cloud. A rigid connection structure has a preset rigid geometry. The first end of the rigid connection structure is fixedly connected to the external tracking reference part, and the second end of the rigid connection structure is fixedly connected to the intraoral imaging reference part, thereby establishing a known and fixed spatial geometric relationship between the external tracking reference part and the intraoral imaging reference part.

10. The apparatus according to claim 9, characterized in that, The geometric parameters of the rigid connection structure are precisely calibrated so that the average tracking error between the external tracking reference and the intraoral imaging reference is no greater than 0.5 mm.

Images