Bone defect bionic customized body generation method and system, terminal and storage medium

Abstract

The invention provides a bone defect bionic customized body generation method and system, a terminal and a storage medium, and the method comprises the steps: obtaining a bone defect model and a target defect repair bionic customized body model based on the defect-free bone data of a patient; on the basis of the point cloud completion model and the point cloud optimization model, completion training and optimization training are sequentially carried out on the bone defect marking point cloud, and an optimized bone defect bionic customized body point cloud is obtained; sequentially performing three-dimensional model reconstruction and reconstruction model optimization on the optimized bone defect bionic customized body point cloud to obtain a first bone defect bionic customized body model; and according to the first bone defect bionic customized body model, generating a personalized bone defect bionic customized body through an entity bionic customized body generation technology. The method solves the problem that in the prior art, preoperative design difficulty of bone increment is large, steps are tedious, and the prior art depends on experience of doctors, so that a bone defect complementing technology is limited. Through the simplified bone defect bionic customized body generation method, the limitation of a bone defect completion technology is reduced.

Description

Technical Field

[0001] The present invention relates to the field of bone defect repair, and in particular to a method, system, terminal and readable storage medium for generating a bone defect bionic customized body. Background Art

[0002] Bone defects caused by trauma, tumors and inflammation are common clinical diseases. Various types of bone defects affect the patient's appearance and function to varying degrees, and bring psychological and physiological obstacles to patients. Human bones vary in appearance and morphology, and can be divided into flat bones, long bones, short bones, irregular bones and sesamoid bones. According to the location of the defective bone, it can be divided into different types of defects such as skull defects, femoral defects, maxillary defects and alveolar bone defects. Different types of bone defects have different treatment principles, and bone transplantation is one of the main methods for various types of bone defects. In view of the limitations of autologous bone grafts, such as limited dosage, the need to open a second surgical area and difficulty in shaping, the development of bone substitute materials for repairing various types of bone defects is still a hot issue in the field of bone regeneration.

[0003] At present, computer-aided design combined with 3D printing technology has become one of the main methods for designing bone defect bionic customized bodies. However, the preoperative design process of this method is complex and highly dependent on experience. In addition, the design is difficult when there is no reference to the contralateral bone structure (such as alveolar bone defects and jaw defects across the midline), which greatly limits the applicable population and field of this method. In addition, the three-dimensional model completion method based on deep learning has been envisioned for the generation design of skull bone defect completion and repair, large-scale maxillofacial bone defect completion and repair, and cartilage defect repair. For example, by artificially creating defects to obtain defect data sets and corresponding true value data sets, input deep learning networks for training and testing, and obtain a three-dimensional model completion algorithm suitable for a single defect type; or directly complete CT images. However, the above-mentioned three-dimensional model completion method is difficult to be applied to the completion and repair of bone defects of multiple types or parts at the same time, forming a universal bone defect completion and repair method. It can be seen that the main defects of the existing technology are high design difficulty, strong dependence on experience and poor applicability. Summary of the invention

[0004] Based on the above problems, the present invention proposes a method, system, terminal and readable storage medium for generating a bionic customized bone defect body, which solves the problem that the prior art bone augmentation technology has limitations due to the difficulty of preoperative design, cumbersome steps and reliance on doctor's experience.

[0005] To achieve the above-mentioned purpose, an embodiment of the present invention provides a method for generating a bionic customized body for bone defect, comprising: based on the defect-free bone data of a patient, obtaining a bone defect model and a target defect repair bionic customized body model by simulating and removing the defect on the patient's defect-free bone model; wherein the target defect repair bionic customized body model is a true value model corresponding to the bone defect model; based on a point cloud completion model and a point cloud optimization model, respectively, performing completion training and optimization training on the bone defect annotated point cloud in sequence to obtain an optimized bone defect bionic customized body point cloud; wherein the bone defect annotated point cloud is obtained by merging the marked point clouds corresponding to the bone defect model and the target defect repair bionic customized body model; obtaining a first bone defect bionic customized body model by successively performing three-dimensional model reconstruction and reconstruction model optimization on the optimized bone defect bionic customized body point cloud; and generating a personalized bone defect bionic customized body through a solid bionic customized body generation technology according to the first bone defect bionic customized body model.

[0006] The embodiment of the present invention proposes a method for generating a bionic customized bone defect body, which simplifies the technical steps of generating a bionic customized bone defect body designed before bone augmentation surgery into: simulating a defect with a bone model without defect to obtain a bone defect model, then performing point cloud design and a first optimization on the bone defect model by using point cloud technology, and finally achieving the generation of a bionic customized bone defect body by performing three-dimensional reconstruction and a second optimization on the point cloud of the optimized bone defect bionic customized body. The technical difficulty of generating a bionic customized bone defect body designed before bone augmentation surgery is reduced by the process steps, and at the same time, because the point cloud technology is used to optimize the bone defect model, the point cloud design and a first optimization are performed on the bone defect model. The patient's bone defect model and the target defect repair model were designed with point cloud, completed and optimized to restore the patient's defect as close to the true value as possible and ensure the accuracy of the patient's bone defect data acquisition. The optimized bone defect bionic customized body point cloud was then reconstructed and optimized in three dimensions to ensure the reliability of the bone defect bionic customized body generation results for preoperative design of bone augmentation, thereby reducing the dependence on the doctor's experience. The simplified bone defect bionic customized body generation method was used to lower the design threshold of preoperative design of bone augmentation and reduce the technical limitations of bone defect completion.

[0007] Furthermore, based on the defect-free bone data of the patient, the defect simulation and removal of the defect on the defect-free bone model of the patient is performed to obtain a bone defect model and a bionic customized body model for target defect repair, specifically: the defect-free bone data of the patient is obtained, and a three-dimensional model is constructed for the defect-free bone data of the patient to obtain a bone model of the patient; a defect site model is obtained by setting the defect site of the patient's bone model; based on the defect-free bone contour retained by the defect site model, a true defect simulation is performed on the defect site model to obtain a bone defect model; and the target defect repair bionic customized body model is obtained by sequentially performing Boolean subtraction and post-processing on the defect site model and the bone defect model.

[0008] An embodiment of the present invention proposes a method for generating a bionic customized body for bone defects, which can accurately obtain the patient's bone defect data by sequentially setting defect sites and simulating true defects on a patient's non-defective bone model, and processing through algorithms such as Boolean subtraction and post-processing, so that the subsequent bionic customized body generation step can accurately generate the patient's defective bionic customized body.

[0009] Furthermore, based on the point cloud completion model and the point cloud optimization model, the bone defect annotated point cloud is respectively completed and optimized to obtain an optimized bone defect bionic customized body point cloud, specifically: the bone defect model and the target defect repair bionic customized body model are respectively point cloud labeled to obtain the labeled point clouds corresponding to the bone defect model and the target defect repair bionic customized body model; the bone defect annotated point cloud is obtained by merging the labeled point clouds corresponding to the bone defect model and the target defect repair bionic customized body model; based on the preset first quantitative index and the first qualitative index, the point cloud completion model is model trained and verified, and the bone defect annotated point cloud is completed and trained through the trained point cloud completion model to obtain the bone defect bionic customized body point cloud; based on the preset first quantitative index and the first qualitative index, the point cloud optimization model is model trained and verified, and the bone defect bionic customized body point cloud is optimized and trained through the trained point cloud optimization model to obtain the optimized bone defect bionic customized body point cloud.

[0010] An embodiment of the present invention proposes a method for generating a bionic customized body for bone defects. Point cloud design is performed on a bone defect model and a target defect repair bionic customized body model, and then an optimized bone defect bionic customized body point cloud is obtained through completion training and optimization training. In this way, the patient's bone defect state can be visualized in the form of a point cloud, which is convenient for subtle adjustments. The point cloud completion model and the point cloud optimization model are also trained and verified by presetting a first quantitative indicator and a first qualitative indicator, which is conducive to accurately obtaining the patient's bone defect data.

[0011] Furthermore, the first bionic customized bone defect model is obtained by successively performing three-dimensional model reconstruction and reconstruction model optimization on the optimized bone defect bionic customized body point cloud, specifically: three-dimensional modeling is performed on the optimized bone defect bionic customized body point cloud through a preset three-dimensional data processing model to obtain a second bionic customized bone defect model; according to the model optimization algorithm, the first bionic customized bone defect model is obtained by successively performing contour smoothing and simulated defect filling on the second bionic customized bone defect model.

[0012] An embodiment of the present invention proposes a method for generating a bionic customized body for bone defects. By performing three-dimensional reconstruction on the optimized point cloud of the bionic customized body for bone defects, and then processing it by means of contour smoothing and simulated defect filling, the first bionic customized body model for bone defects is made more accurate and more fitted to the bone defect position, which is conducive to printing out bone blocks with better retention.

[0013] Furthermore, according to the first bone defect bionic customized body model, a personalized bone defect bionic customized body is generated through a solid bionic customized body generation technology, specifically: based on a preset second quantitative indicator and a second qualitative indicator, the first bone defect bionic customized body model is evaluated to obtain an evaluation result of the first bone defect bionic customized body model; when the evaluation result of the first bone defect bionic customized body model meets the preset requirements, according to the patient's bone defect data, the defect position of the first bone defect bionic customized body model is adjusted and a defect simulation application is performed to obtain bone defect bionic customized body data; according to the bone augmentation material, the first bone defect bionic customized body model is printed through bio-printing technology to generate a personalized bone defect bionic customized body.

[0014] An embodiment of the present invention proposes a method for generating a bionic customized body for bone defects, wherein a first bionic customized body model for bone defects is evaluated by presetting a second quantitative indicator and a second qualitative indicator. The bionic customized body is printed only when the first bionic customized body model for bone defects meets the preset requirements, thereby determining that the first bionic customized body model for bone defects is trained in place and is more accurately close to the actual bone defect situation, thereby further improving the accuracy of bone defect generation.

[0015] The embodiment of the present invention also provides a bone defect bionic customized body generation system, comprising: a bone defect model construction module, a point cloud marking and optimization module, a model reconstruction and optimization module and a bone defect bionic customized body generation module; the bone defect model construction module is used to obtain a bone defect model and a target defect repair bionic customized body model by performing defect simulation removal on the patient's non-defective bone model based on the patient's non-defective bone data; wherein the target defect repair bionic customized body model is a true value model corresponding to the bone defect model; the point cloud marking and optimization module is used to respectively perform bone defect repair on the bone defect based on the point cloud completion model and the point cloud optimization model. The annotated point cloud is sequentially subjected to completion training and optimization training to obtain an optimized bone defect bionic customized body point cloud; wherein, the bone defect annotated point cloud is obtained by merging the marked point clouds corresponding to the bone defect model and the target defect repair bionic customized body model; the model reconstruction and optimization module is used to sequentially perform three-dimensional model reconstruction and reconstruction model optimization on the optimized bone defect bionic customized body point cloud to obtain a first bone defect bionic customized body model; the bone defect bionic customized body generation module is used to generate a personalized bone defect bionic customized body according to the first bone defect bionic customized body model through the solid bionic customized body generation technology.

[0016] The embodiment of the present invention proposes a bone defect bionic customized body generation system, which simplifies the technical steps of generating a bone defect bionic customized body designed before bone augmentation surgery into: simulating the defect of a non-defective bone model through a bone defect model construction module to obtain a bone defect model, then using a point cloud marking and optimization module to perform point cloud design and a first optimization on the bone defect model using point cloud technology, and performing three-dimensional reconstruction and a second optimization on the optimized bone defect bionic customized body point cloud through a model reconstruction and optimization module, and finally realizing the generation of a bone defect bionic customized body through a bone defect bionic customized body generation module; the process steps reduce the time and effort of the bionic customized body design of the bone defect before bone augmentation surgery. At the same time, the point cloud marking and optimization module is used to perform point cloud design, point cloud completion and point cloud optimization on the patient's bone defect model and the target defect repair model, so as to restore the patient's defect situation as close to the true value and ensure the accuracy of the patient's bone defect data acquisition. The model reconstruction and optimization module is then used to perform three-dimensional reconstruction and optimization of the optimized bone defect bionic customized body point cloud, which ensures the reliability of the bone defect bionic customized body generation results designed before bone augmentation surgery, thereby reducing the dependence on doctors' experience; and through the simplified bone defect bionic customized body generation method, the design threshold of bone augmentation preoperative design is lowered, and the technical limitations of bone defect completion are reduced.

[0017] Furthermore, the bone defect model construction module is used to obtain a bone defect model and a target defect repair bionic customized body model by performing defect simulation removal on the patient's non-defective bone model based on the patient's non-defective bone data, and also includes: a three-dimensional reconstruction unit, a defect site model construction unit, a bone defect true value simulation unit and a Boolean processing unit;

[0018] The three-dimensional reconstruction unit is used to obtain the patient's non-defective bone data, and to construct a three-dimensional model of the patient's non-defective bone data to obtain a patient's non-defective bone model;

[0019] The defect site model construction unit is used to obtain a defect site model by setting a defect site on the patient's non-defective bone model;

[0020] The bone defect true value simulation unit is used to obtain a bone defect model by performing true value defect simulation on the defect site model based on the defect-free bone contour retained by the defect site model;

[0021] The Boolean processing unit is used to obtain a target defect repair bionic customized body model by performing Boolean subtraction and post-processing on the defect site model and the bone defect model in sequence.

[0022] Furthermore, the point cloud labeling and optimization module is used to perform completion training and optimization training on the bone defect annotated point cloud in sequence based on the point cloud completion model and the point cloud optimization model to obtain an optimized bone defect bionic customized body point cloud, and also includes: a first point cloud set acquisition unit, a second point cloud set acquisition unit, a completion training unit and an optimization training unit;

[0023] The first point cloud set acquisition unit is used to perform point cloud marking on the bone defect model and the target defect repair bionic customized body model respectively, to obtain marked point clouds corresponding to the bone defect model and the target defect repair bionic customized body model;

[0024] The second point cloud set acquisition unit is used to obtain a bone defect annotated point cloud by merging the marked point clouds corresponding to the bone defect model and the target defect repair bionic customized body model;

[0025] The completion training unit is used to perform model training and verification on the point cloud completion model based on the preset first quantitative index and the first qualitative index, and to perform completion training on the bone defect annotated point cloud through the trained point cloud completion model to obtain the bone defect bionic customized volume point cloud;

[0026] The optimization training unit is used to perform model training and verification on the point cloud optimization model based on the preset first quantitative index and the first qualitative index, and optimize the bone defect bionic customized body point cloud through the trained point cloud optimization model to obtain the optimized bone defect bionic customized body point cloud.

[0027] Further, the model reconstruction and optimization module is used to obtain a first bone defect bionic customized body model by sequentially performing three-dimensional model reconstruction and reconstruction model optimization on the optimized bone defect bionic customized body point cloud, and further includes: a three-dimensional model construction unit and a smooth filling unit;

[0028] The three-dimensional model building unit is used to perform three-dimensional modeling on the optimized bone defect bionic customized body point cloud through a preset three-dimensional data processing model to obtain a second bone defect bionic customized body model;

[0029] The smoothing and filling unit is used to obtain the first bone defect bionic customized body model by sequentially smoothing the contour and simulating defect filling on the second bone defect bionic customized body model according to the model optimization algorithm.

[0030] Furthermore, the bone defect bionic customized body generation module is used to generate a personalized bone defect bionic customized body according to the first bone defect bionic customized body model through a solid bionic customized body generation technology, and also includes: a model evaluation unit, a defect simulation application unit and a bionic customized body printing unit;

[0031] The model evaluation unit is used to evaluate the first bone defect bionic customized body model based on the preset second quantitative index and the second qualitative index to obtain an evaluation result of the first bone defect bionic customized body model;

[0032] The defect simulation application unit is used to obtain bone defect bionic customized body data by adjusting the defect position of the first bone defect bionic customized body model and performing defect simulation application according to the patient's bone defect data when the evaluation result of the first bone defect bionic customized body model meets the preset requirements;

[0033] The bionic customized body printing unit is used to print the first bone defect bionic customized body model to generate a personalized bone defect bionic customized body through bio-printing technology according to the bone augmentation material.

[0034] A terminal for generating a bionic customized bone defect body comprises a processor, a memory and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, a method for generating a bionic customized bone defect body is implemented.

[0035] A storage medium for generating a bionic customized bone defect body, the storage medium comprising: a computer-readable storage medium and a stored computer program, wherein when the computer program is run, the device where the computer-readable storage medium is located is controlled to execute the method for generating a bionic customized bone defect body. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A schematic diagram of the steps of a method for generating a bone defect bionic customized body provided by a certain embodiment of the present invention;

[0037] Figure 2 A schematic diagram of an oral three-dimensional model of a method for generating a bone defect bionic customized body provided by a certain embodiment of the present invention;

[0038] Figure 3 A module structure diagram of a bone defect bionic custom body generation system provided by a certain embodiment of the present invention;

[0039] Figure 4 A structural diagram of a bone defect model building module of a bone defect bionic customized body generation system provided by a certain embodiment of the present invention;

[0040] Figure 5 A point cloud marking and optimization module structure diagram of a bone defect bionic custom body generation system provided by a certain embodiment of the present invention;

[0041] Figure 6 A structural diagram of a model reconstruction and optimization module of a bone defect bionic custom body generation system provided by a certain embodiment of the present invention;

[0042] Figure 7 A structural diagram of a bone defect bionic customized body generation module of a bone defect bionic customized body generation system provided in one embodiment of the present invention;

[0043] Figure 8 An example of a method for generating a bionic custom-made bone defect body provided by a certain embodiment of the present invention is shown in FIG. 1 ;

[0044] Fig. 9 A schematic diagram of application of Example 2 of a method for generating a bionic custom-made bone defect body provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] Bone is the support of soft tissue and the scaffold for the exercise of physiological functions. Bone replacement materials need to have a bionic appearance structure similar to the original bone structure to provide a basis for the restoration of soft tissue appearance and function. In addition, it is also necessary to have a contact surface that fits the bone defect site well to facilitate the fixation of bone replacement materials and subsequent bone healing and new bone formation. With the development of various bone replacement materials and additive manufacturing (3D printing) technology, bionic bone replacement materials customized according to the characteristics of the patient's bone defect (referred to as "bionic customized body for bone defect") have been regarded as an important strategy for solving different types of bone defect repair. In recent years, computer-aided design combined with 3D printing technology has become one of the main methods for the design of bionic customized body for bone defect, but the preoperative design process of this method is complex and highly dependent on experience, and the design is difficult when there is no reference to the contralateral bone structure (such as alveolar bone defect and mandibular defect across the midline), which greatly limits the applicable population and field of this method.

[0047] In recent years, with the vigorous development of artificial intelligence (AI) technology, the three-dimensional model completion algorithm based on deep learning can quickly complete and repair defects in different parts of the three-dimensional model or across the midline, bringing new ideas to the generation of various bone defect completions and bone defect bionic customized bodies, which is expected to reduce the difficulty of customized body design and facilitate large-scale clinical promotion.

[0048] In the medical field, deep learning-based 3D model completion methods have been envisioned for the generation and design of skull bone defect completion and repair, large-scale maxillofacial bone defect completion and repair, and cartilage defect repair. For example, by artificially creating defects to obtain defect data sets and corresponding true value data sets, input deep learning networks for training and testing, and obtain 3D model completion algorithms suitable for a single defect type; or directly complete CT images. However, the above-mentioned 3D model completion methods are difficult to be applied to the completion and repair of bone defects of multiple types or parts at the same time, forming a universal bone defect completion and repair method. There are mainly the following reasons: 1. The 3D model completion algorithm needs to learn a specific type of 3D model in order to complete the 3D model of this type. However, a public mixed multi-type bone database for deep learning needs to be further developed and established; in addition, the imaging acquisition methods, equipment, data formats and data protocols used for bone data in different parts are often different, and the universal bone defect completion and repair method requires all data to be preprocessed and normalized in advance. 2. In existing studies on completion and repair, the defect is often created by using the Boolean subtraction method between a complete bone model and a regular cube such as a sphere or column, which is still far from the actual clinical defect situation. 3. Bone defects include both large and regular bone defects such as the skull or femur, and small and irregular bone defects such as alveolar bone defects. The general bone defect completion method requires a fitting contact surface and a bionic bone shape contour, which places high requirements on the accuracy of the three-dimensional model completion algorithm.

[0049] According to the three-dimensional data expression form, the current three-dimensional model completion algorithm can be based on voxels, point clouds, grids, implicit expressions, structured representations, two-dimensional images or depth images for deep learning. Among them, point clouds are usually the original output form of three-dimensional acquisition equipment, which can be converted into various three-dimensional data formats; in addition, point cloud data can accurately represent three-dimensional shapes, and has the advantages of efficient and concise data structure suitable for deep learning, and storage consumes less memory. The three-dimensional model completion algorithm based on point clouds is also relatively mature. It can be applied to large-volume scene completion such as cultural relics restoration and three-dimensional scene completion, and can also be applied to small-volume irregular defect completion such as component incomplete completion. The three-dimensional model completion algorithm based on point clouds has great potential in the clinical application of various types of bone defect completion and repair.

[0050] Dental implant restoration has become one of the most important treatments for tooth defects. Alveolar bone remodeling after tooth extraction, periodontal disease, trauma and tumors may cause alveolar bone defects, affecting the implantation and long-term stability of the implant. Bone augmentation surgery is required before implantation to restore the width and height of the alveolar bone. Among them, implantation in the aesthetic zone has higher requirements for the repair of alveolar bone defects. In addition to obtaining sufficient bone volume around the implant, it is also necessary to form a good alveolar bone contour to provide a basis for aesthetic restoration of soft tissue. Therefore, in this embodiment, the application of a bone defect bionic customized body generation method proposed in the present invention in the oral field is exemplified, and no other limitations are made here and will not be repeated below. See. Figure 2 , Figure 2 A schematic diagram of an oral three-dimensional model of a method for generating a bone defect bionic customized body provided by a certain embodiment of the present invention; Figure 2 As shown, A is an example of a 3D model dataset of a non-defective jaw; B is an example of a 3D model dataset of a toothless jaw; C is an example of a jaw model dataset with alveolar bone defects and the corresponding ideal bone defect repair bionic customized body model dataset; D is an example of an annotated point cloud dataset of alveolar bone defects.

[0051] Example 1

[0052] See also Figure 1 , Figure 1 A schematic diagram of the steps of a method for generating a bone defect bionic customized body according to an embodiment of the present invention. Figure 1 As shown, the embodiment of the present invention provides a method for generating a bone defect bionic customized body, including steps 101 to 104, each step is specifically as follows:

[0053] Step 101, based on the patient's defect-free bone data, a bone defect model and a target defect repair bionic customized body model are obtained by performing defect simulation removal on the patient's defect-free bone model; wherein the target defect repair bionic customized body model is a true value model corresponding to the bone defect model;

[0054] As an example of this embodiment, the patient's defect-free bone data is obtained, and a three-dimensional model is constructed for the patient's defect-free bone data to obtain the patient's defect-free bone model; a defect site model is obtained by setting the defect site of the patient's defect-free bone model; based on the defect-free bone contour retained by the defect site model, a true defect simulation is performed on the defect site model to obtain a bone defect model; and a target defect repair bionic customized body model is obtained by performing Boolean subtraction and post-processing on the defect site model and the bone defect model in sequence.

[0055] As another example of this embodiment, oral CBCT image data (patients without defective bone data) that meet the inclusion criteria are screened and obtained, and exported and stored in a medical image format to obtain a bionic jaw defect database, wherein the inclusion criteria include but are not limited to no missing teeth in the maxillary anterior teeth area, periodontal health in the maxillary anterior teeth area, no craniomaxillofacial fractures and other lesions; the oral CBCT image data are imported through imaging processing software, the horizontal reference plane and the vertical reference plane are unified, the jaw and the corresponding alveolar bone and teeth in the area of ​​interest are segmented, the jaw in the above-mentioned area of ​​interest is three-dimensionally reconstructed, and exported in a 3D model file format to obtain a defect-free jaw three-dimensional model and construct a certain scale of defect-free jaw three-dimensional model data set (equivalent to the patient's defect-free bone model). Referring to the actual clinical situation of alveolar bone defects, specific teeth on the defect-free jaw three-dimensional model are removed to obtain a edentulous jaw three-dimensional model (equivalent to the defect site model), which has a defect-free alveolar bone contour at the tooth-missing site. Furthermore, a vertical alveolar bone defect is artificially created at the tooth-missing site to obtain a jaw model of the alveolar bone defect (equivalent to a bone defect model), and the defect is exported in a 3D model file format. Optionally, the tooth-missing site and the corresponding alveolar bone defect range include but are not limited to a single tooth alveolar bone defect, multiple consecutive tooth alveolar bone defects, unilateral alveolar bone defects or alveolar bone defects across the midline, etc., and the types of alveolar bone defects include but are not limited to vertical bone defects, horizontal bone defects, mixed bone defects (L-shaped bone defects), etc.; Optionally, the jaw and the corresponding alveolar bone and teeth in the area of ​​interest may come from the maxilla or mandible, and the range may include the entire dentition or only part of the dentition. A specific implementation method is: import the oral CBCT image data through Mimics, select the orbitoauricular plane and the human body midline as the horizontal reference plane and the vertical reference plane, use the segmentation function of the software to segment the maxillary anterior teeth area, reconstruct the healthy maxillary anterior teeth area in three dimensions, and export it in STL format to obtain a three-dimensional model of the maxillary anterior teeth area without defects. Specifically, the three-dimensional model of the maxillary anterior teeth area without defects includes eight teeth in the maxillary anterior teeth area and a healthy and clear alveolar bone structure, with its upper boundary to the bottom of the nose, the lower boundary including all the maxillary anterior teeth but not the mandibular teeth, the left and right boundaries to the outer edge of the first premolar, the front boundary including all the anterior teeth, and the rear boundary to the rear edge of the first premolar. This step collects and constructs a data set of 150 healthy maxillary anterior teeth 3D models; imports the intact maxillary anterior teeth 3D model into Geomagic Wrap, removes teeth 11 and 21 from the healthy maxillary anterior teeth 3D model, and artificially creates a vertical alveolar bone defect at the tooth position to obtain a maxillary anterior teeth vertical alveolar bone defect jaw model, which is then exported in STL format.Furthermore, a three-dimensional data processing software is used to perform Boolean subtraction operations on the three-dimensional model of the edentulous jaw and the alveolar bone defect jaw model, and a clinical physician with a background in oral medicine performs post-processing on the obtained model to obtain an ideal bone defect repair bionic customized body model (equivalent to the target defect repair bionic customized body model), that is, a true value model corresponding to the alveolar bone defect model, and export it in a 3D model file format. For the alveolar bone defect model, the ideal bone defect repair bionic customized body model has an ideal volume size, physiological contour and stable retention, which meets the needs of 3D printing and oral implant clinical bone transplantation. Optionally, the post-processing steps include but are not limited to smoothing, deleting, partially filling and deleting nail-shaped objects on the surface and edges of the model.

[0056] Step 102, based on the point cloud completion model and the point cloud optimization model, the bone defect annotated point cloud is sequentially completed and optimized to obtain an optimized bone defect bionic customized volume point cloud; wherein the bone defect annotated point cloud is obtained by merging the marked point clouds corresponding to the bone defect model and the target defect repair bionic customized volume model;

[0057] As an example of this embodiment, the bone defect model and the target defect repair bionic customized body model are respectively marked with point clouds to obtain the marked point clouds corresponding to the bone defect model and the target defect repair bionic customized body model; the bone defect annotated point cloud is obtained by merging the marked point clouds corresponding to the bone defect model and the target defect repair bionic customized body model; based on the preset first quantitative index and the first qualitative index, the point cloud completion model is trained and verified, and the bone defect annotated point cloud is completed and trained by the trained point cloud completion model to obtain the bone defect bionic customized body point cloud; specifically, the bone defect annotated point cloud is input into the preset point cloud completion model for downsampling to obtain several bone defect annotated point clouds with different resolutions; by performing feature extraction on the bone defect annotated point cloud data with different resolutions, the potential vectors mapped by the several bone defect annotated point clouds are obtained; by performing multi-layer perceptual mapping on the potential vectors mapped by the several bone defect annotated point clouds, a final feature vector is generated; by decoding the final feature vector, the bone defect bionic customized body point cloud is obtained. Based on the preset first quantitative index and the first qualitative index, the point cloud optimization model is trained and verified, and the bone defect bionic customized body point cloud is optimized and trained by the trained point cloud optimization model to obtain an optimized bone defect bionic customized body point cloud; specifically, the bone defect bionic customized body point cloud is scored point by point by a preset point cloud denoising model to obtain a score value for each point; the optimization training evaluation index is obtained by calculating the loss between the vector difference and the score value of the nearest point in the noise-free point cloud corresponding to each point of the bone defect bionic customized body point cloud; optionally, the point cloud optimization model can also choose to perform a point cloud upsampling model, specifically, the bone defect bionic customized body point cloud to be upsampled is input into the point cloud upsampling model constructed based on the deep learning model for optimization training, and optimized bone defect bionic customized body point clouds of different multiples and corresponding optimization training evaluation indicators are obtained; according to the optimization training evaluation index, the position of each point of the bone defect bionic customized body point cloud is updated until the preset requirements are met, so as to obtain an optimized bone defect bionic customized body point cloud.

[0058] As another example of this embodiment, the alveolar bone defect model and the ideal bone defect repair bionic custom body model are respectively imported into the point cloud mesh processing software and converted into point clouds. Further, the alveolar bone defect jaw model point cloud and the ideal bone defect repair bionic custom body model point cloud are marked respectively. Use the number "0" to mark the ideal bone defect repair bionic custom body model point cloud (equivalent to the marked point cloud corresponding to the target defect repair bionic custom body model). Use the number "1" to mark the alveolar bone defect model point cloud (equivalent to the marked point cloud corresponding to the bone defect model). Further, the alveolar bone defect model point cloud and the ideal bone defect repair bionic custom body model point cloud are merged to obtain the alveolar bone defect annotated point cloud and export it in point cloud format. Perform the above operations on all models to obtain the alveolar bone defect annotated point cloud data set. Specific point cloud marking can be used: CloudCompare.

[0059] As another example of this embodiment, one possible implementation method of the point cloud completion step is: normalize the point cloud of the alveolar bone defect model, the point cloud representing the true ideal bone defect repair bionic customized body model, and the point cloud of the ideal bone defect repair bionic customized body model with the mass point as the center. The point cloud of the ideal bone defect repair bionic customized body model is uniformly downsampled to the same number of points, and the deep learning-based point cloud completion model extracts features from the input point cloud, predicts and completes the defect site, and finally outputs a point cloud of the bone defect bionic customized body representing the missing area. Specifically, the alveolar bone defect annotated point cloud data set is input into a point cloud completion model based on deep learning, and a variety of training strategies are used to train, verify and optimize the point cloud completion model; after training, the alveolar bone defect model point cloud is input, and the bone defect bionic customized body point cloud representing the missing area is output; optionally, the network architecture method used by the deep learning-based point cloud completion model includes but is not limited to point-based methods, CNN-based methods, image-based methods, view-based methods, GAN-based methods and Transformer-based methods, etc., and the specific point cloud completion model includes but is not limited to PF-Net, PMP-Net, GRNet, ViPC and PointTr, etc.; further, a post-processing model for optimizing the bone defect bionic customized body point cloud is constructed and trained to obtain an optimized bone defect bionic customized body point cloud. Optionally, the post-processing model based on deep learning includes but is not limited to a point cloud denoising model and a point cloud upsampling model. The point cloud completion model of the post-processing model for optimizing the bone defect bionic customized body point cloud is trained and verified. After training, the point cloud of the bionic customized bone defect volume is input to obtain the optimized point cloud of the bionic customized bone defect volume;

[0060] One specific implementable method: One implementable method of point cloud completion is: the vertical alveolar bone defect model of the tooth is sampled as 8192, and the ideal bone defect repair imitation bio-customized model of the point cloud representing the true value and the ideal bone defect repair imitation bio-customized model of the point cloud are both 2048. The point cloud completion model based on the improved PF-Net specifically includes: a multi-resolution encoder module: the multi-resolution encoder structure of PF-Net is adopted to transform the point cloud of the upper front teeth vertical alveolar bone defect model through a multi-resolution encoder. The transformed point cloud of the upper front teeth vertical alveolar bone defect model is downsampled by an iterative farthest point sampling algorithm to obtain three point clouds of the upper front teeth vertical alveolar bone defect model with different resolution labels, and the point clouds of the upper front teeth vertical alveolar bone defect model with different resolutions are mapped into three latent vectors through three independent combined multi-layer perceptions, and these three latent vectors are connected to obtain a latent feature map, and further, the latent feature map is integrated into the final feature vector using multi-layer perception. Specifically, the input point cloud is copied three times and then downsampled to 1, 2, and 8 times, respectively, that is, [8192, 8192, 8192] is downsampled to [8192, 4096, 1024]. The downsampled point clouds are subjected to feature extraction to generate a potential feature vector group [1920, 1920, 1920], and the final feature vector

[1920] is generated using multi-layer perceptual mapping. Point pyramid decoder module: The point pyramid decoder of PF-Net is used, the final feature vector is input, and the generated bone defect bionic customized body point cloud M representing the missing area is output. Loss function module: The multi-stage completion loss and adversarial loss (Loss G, Loss D, Loss CD) of PF-Net, as well as F-score, replusion and uniformity are used as evaluation indicators. Point cloud completion model based on improved PF-Net built using Adam optimizer on PyTorch. The vertical alveolar bone defect annotated point cloud dataset is input into the point cloud completion model based on the improved PF-Net, and a variety of training strategies are used to train, verify and optimize the point cloud completion model. Specifically, the initial learning rate is 0.0001, the batch size is 36; k=2 is set in the multi-resolution encoder module, the number of points of the output point cloud is 1024, and the training strategy of alternating iterative training is adopted. After training, the point cloud of the vertical alveolar bone defect model of the upper anterior teeth is input, and the model can output the bone defect bionic customized body point cloud M' predicted by the model, with 1024 points.

[0061] One possible implementation method of point cloud optimization is: obtain the open source PU-Net dataset and perform Poisson disk sampling operations and randomly add noise to obtain a training noise point cloud set, a test noise point cloud set, and a corresponding noise-free point cloud set. Construct and train a score-based point cloud denoising model. The generated bone defect bionic customized body point cloud M' is input into the point cloud denoising model to obtain an optimized bone defect bionic customized body point cloud M1. Specifically, the point cloud denoising model includes: constructing a score estimation network: performing point-by-point score estimation on the noise point cloud set on a local basis to obtain a score function in a point-by-point neighborhood space, specifically including: constructing a score estimation network, a feature extraction unit, a score estimation unit, and an objective function module. Furthermore, the Adam optimizer is used, and L2 Loss is used as the evaluation index. The batch size and learning rate of the experiment on each dataset are set. After multiple iterative training, it is achieved to fit the training set. Specifically, L2 Loss represents the loss between the vector difference and score of each point in the input point cloud set and its corresponding nearest point in the noise-free point cloud; the learning rate is 1e-4weight_decay 0, the batch size is 32, and the number of iterations is 60000. Point cloud denoising algorithm module based on scoring: using the previously trained model and combining it into an integrated scoring function, the position of each point is updated through gradient ascent to obtain the denoised point cloud, that is, the optimized bone defect bionic customized body point cloud M1.

[0062] One possible implementation method of point cloud optimization is: obtain an open source PU-GAN dataset and perform Poisson disk sampling on the test set to obtain a training point cloud set and a test point cloud set and their corresponding label sets. Construct and train an arbitrary scale point cloud upsampling model based on deep learning, input the generated bone defect bionic customized body point cloud M' to be upsampled into the point cloud upsampling model, and obtain different multiples of bone defect bionic customized body point cloud M2. Specifically, the arbitrary scale point cloud upsampling model based on deep learning specifically includes: using midpoint generation and farthest point sampling to implement upsampling operations on point clouds, decoupling point generation from model training, and obtaining a midpoint interpolation point cloud set. Construct a distance network model based on P2PNet and use the above midpoint interpolation point cloud set and its corresponding label set for model training. Adopt the Adam optimizer, use L1 Loss as the evaluation index, and after multiple iterative training, set the batch size and learning rate of the experiment on each data set to achieve a good fit on the training set. Specifically, L1 Loss represents minimizing the loss between the predicted distance and the true distance; the batch size is 64, the learning rate is 1e-3weight_decay 0, the beta smoothing L1 Loss is 0.01, and the epoch is 110.

[0063] Based on the above point cloud completion model and point cloud optimization model, according to the preset first quantitative index and the first qualitative index, the optimized bone defect bionic customized volume point cloud is quantitatively and qualitatively analyzed with reference to the preset true value bone defect bionic customized volume point cloud; wherein the preset first quantitative index includes: CD value score, F-score score, replusion score and uniformity score; the first qualitative index includes but is not limited to doctor score, etc.;

[0064] Step 103, obtaining a first bone defect bionic customized body model by sequentially performing three-dimensional model reconstruction and reconstruction model optimization on the optimized bone defect bionic customized body point cloud;

[0065] As an example of this embodiment, the optimized bone defect bionic customized body point cloud is three-dimensionally modeled through a preset three-dimensional data processing model to obtain a second bone defect bionic customized body model; according to the model optimization algorithm, the second bone defect bionic customized body model is successively subjected to contour smoothing and simulated defect filling to obtain the first bone defect bionic customized body model.

[0066] As another example of this embodiment, the bone defect bionic customized body point cloud is imported into the three-dimensional data processing software, and the encapsulation function of the software is used for three-dimensional reconstruction to obtain a preliminary bone defect bionic customized body model (equivalent to the second bone defect bionic customized body model); further, the preliminary bone defect bionic customized body model is processed using the basic functions such as filling holes and smoothing provided by the software to obtain a fine bone defect bionic customized body model (equivalent to the first bone defect bionic customized body model) to meet the requirements of 3D printing. Optionally, the edges and contours of the fine bone defect bionic customized body model are further processed using the three-dimensional data processing software to make it more closely fit the corresponding alveolar bone defect model, so that the subsequent 3D printed bone blocks can obtain better retention. Specifically, Geomagic Wrap can be used.

[0067] Step 105: Generate a personalized bone defect bionic customized body according to the first bone defect bionic customized body model by using a solid bionic customized body generation technology.

[0068] As an example of this embodiment, based on the preset second quantitative index and the second qualitative index, the first bone defect bionic customized body model is evaluated to obtain an evaluation result of the first bone defect bionic customized body model; when the evaluation result of the first bone defect bionic customized body model meets the preset requirements, according to the patient's bone defect data, by adjusting the defect position of the first bone defect bionic customized body model and performing defect simulation implantation, the bone defect bionic customized body data is obtained; according to the bone augmentation material, the first bone defect bionic customized body model is printed by bioprinting technology to generate a personalized bone defect bionic customized body.

[0069] As another example of this embodiment, according to the preset second quantitative index, the first bone defect bionic customized body model is quantitatively analyzed with reference to the preset true value bone defect repair bionic customized body model to obtain a quantitative analysis result; wherein the preset second quantitative index includes: maximum deviation, minimum deviation, average deviation, standard deviation, RMS, discreteness, percentage within tolerance and three-dimensional deviation diagram; according to the quantitative analysis result and the second qualitative analysis index, the first bone defect bionic customized body model is qualitatively analyzed to obtain a qualitative analysis result; by using the preset second qualitative index, the first bone defect bionic customized body model is scored, and the score of the first bone defect bionic customized body model is updated until the preset requirements are met to obtain a qualitative analysis result. Specifically, with reference to the ideal bone defect repair bionic customized body model, the refined bone defect bionic customized body model is qualitatively analyzed. The qualitative analysis includes but is not limited to doctor scoring, such as fitting the fine bone defect bionic customized body model to the corresponding alveolar bone defect jaw model through three-dimensional data processing software, simulating the effect of transplanting the fine bone defect bionic customized body model after 3D printing to the corresponding patient's alveolar bone defect position, and exporting pictures of different views. The fine bone defect bionic customized body model is imported into the corresponding CBCT image data (equivalent to the bone defect bionic customized body data) through imaging processing software, the cylindrical simulated implant is loaded, the defect is simulated and implanted at the site, and different view pictures are exported. Further, the above two pictures are handed over to senior oral implant clinicians for scoring; a specific implementation method is to import the fine bone defect bionic customized body model into the corresponding CBCT image data through Mimics, load the 3.3mm*12mm cylindrical simulated implant, simulate implantation at sites 11 and 21, and export different view pictures in JPEG format.

[0070] According to the patient's bone defect data, by adjusting the defect position of the superior bone defect model and performing defect simulation implantation, the bone defect bionic customized body data is obtained, the bone augmentation material slurry is prepared, the slurry is poured into the slurry tank of the sinking DLP light-curing 3D printer and the printing parameters are set. Further, the bone defect bionic customized body data is imported, the bone augmentation material is used as the raw material, and the sinking DLP bioprinting technology is used for printing, and the washing, drying, degreasing and sintering are performed to finally obtain a 3D printed personalized bone defect bionic customized body. Optionally, the bioprinting technology can also use high-precision 3D printing technology.

[0071] The embodiment of the present invention proposes a method for generating a bionic customized bone defect body, which simplifies the technical steps of generating a bionic customized bone defect body designed before bone augmentation surgery into: simulating a defect with a bone model without defect to obtain a bone defect model, then performing point cloud design and a first optimization on the bone defect model by using point cloud technology, and finally achieving the generation of a bionic customized bone defect body by performing three-dimensional reconstruction and a second optimization on the point cloud of the optimized bone defect bionic customized body. The technical difficulty of generating a bionic customized bone defect body designed before bone augmentation surgery is reduced by the process steps, and at the same time, because the point cloud technology is used to optimize the bone defect model, the point cloud design and a first optimization are performed on the bone defect model. The patient's bone defect model and the target defect repair model were designed with point cloud, completed and optimized to restore the patient's defect as close to the true value as possible and ensure the accuracy of the patient's bone defect data acquisition. The optimized bone defect bionic customized body point cloud was then reconstructed and optimized in three dimensions to ensure the reliability of the bone defect bionic customized body generation results for preoperative design of bone augmentation, thereby reducing the dependence on the doctor's experience. The simplified bone defect bionic customized body generation method was used to lower the design threshold of preoperative design of bone augmentation and reduce the technical limitations of bone defect completion.

[0072] Example 2

[0073] See also Figure 3 , Figure 3 The module structure diagram of a bone defect bionic customized body generation system provided by a certain embodiment of the present invention. Figure 3 As shown, the present invention proposes a bone defect bionic customized body generation system, including: a bone defect model construction module 301, a point cloud marking and optimization module 302, a model reconstruction and optimization module 303 and a bone defect bionic customized body generation module 304;

[0074] The bone defect model construction module 301 is used to obtain a bone defect model and a target defect repair bionic customized volume model based on the patient's defect-free bone data by performing defect simulation removal on the patient's defect-free bone model; wherein the target defect repair bionic customized volume model is a true value model corresponding to the bone defect model;

[0075] As an example of this embodiment, see Figure 4 , Figure 4 FIG. 1 is a structural diagram of a bone defect model building module of a bone defect bionic customized body generation system provided in an embodiment of the present invention; Figure 4 As shown, the bone defect model construction module 301 is used to simulate and remove the defect of the preset non-defective bone model through the patient's bone defect data to obtain the bone defect model; wherein the preset non-defective bone model is composed of a plurality of non-defective bone three-dimensional models, and also includes: a three-dimensional reconstruction unit 401, a defect site model construction unit 402, a bone defect true value simulation unit 403 and a Boolean processing unit 404;

[0076] The three-dimensional reconstruction unit 401 is used to obtain the patient's non-defective bone data, and construct a three-dimensional model for the patient's non-defective bone data to obtain a patient's non-defective bone model;

[0077] The defect site model construction unit 402 is used for obtaining a defect site model by setting a defect site on the non-defective bone model of the patient;

[0078] The bone defect true value simulation unit 403 is used for obtaining a bone defect model by performing true value defect simulation on the defect site model based on the defect-free bone contour retained by the defect site model;

[0079] The Boolean processing unit 404 is used to obtain a target defect repair bionic customized body model by performing Boolean subtraction and post-processing on the defect site model and the bone defect model in sequence.

[0080] The point cloud labeling and optimization module 302 is used to perform completion training and optimization training on the bone defect annotated point cloud in sequence based on the point cloud completion model and the point cloud optimization model, so as to obtain an optimized bone defect bionic customized volume point cloud; wherein the bone defect annotated point cloud is obtained by merging the labeled point clouds corresponding to the bone defect model and the target defect repair bionic customized volume model;

[0081] As an example of this embodiment, see Figure 5 , Figure 5 A point cloud marking and optimization module structure diagram of a bone defect bionic customized body generation system provided by a certain embodiment of the present invention, such as Figure 5 As shown, the point cloud marking and optimization module 302 is used to perform completion training and optimization training on the bone defect annotation point cloud in sequence through the point cloud completion model according to a preset completion processing algorithm to obtain an optimized bone defect bionic customized body point cloud; wherein the bone defect annotation point cloud is obtained by performing point cloud marking on the bone defect model, and further includes: a first point cloud set acquisition unit 501, a second point cloud set acquisition unit 502, a completion training unit 503 and an optimization training unit 504;

[0082] The first point cloud set acquisition unit 501 is used to perform point cloud labeling on the bone defect model and the target defect repair bionic customized body model respectively, to obtain the labeled point clouds corresponding to the bone defect model and the target defect repair bionic customized body model;

[0083] The second point cloud set acquisition unit 502 is used to obtain a bone defect annotated point cloud by merging the marked point clouds corresponding to the bone defect model and the target defect repair bionic customized body model;

[0084] The completion training unit 503 is used to perform model training and verification on the point cloud completion model based on the preset first quantitative index and the first qualitative index, and to perform completion training on the bone defect annotated point cloud through the trained point cloud completion model to obtain the bone defect bionic customized volume point cloud;

[0085] The optimization training unit 504 is used to perform model training and verification on the point cloud optimization model based on the preset first quantitative index and the first qualitative index, and optimize the bone defect bionic customized body point cloud through the trained point cloud optimization model to obtain the optimized bone defect bionic customized body point cloud.

[0086] The model reconstruction and optimization module 303 is used to obtain a first bone defect bionic customized body model by sequentially performing three-dimensional model reconstruction and reconstruction model optimization on the optimized bone defect bionic customized body point cloud;

[0087] As an example of this embodiment, see Figure 6 , Figure 6 A structural diagram of a model reconstruction and optimization module of a bone defect bionic customized body generation system provided by a certain embodiment of the present invention, such as Figure 6 As shown, the model reconstruction and optimization module 303 is used to obtain a first bone defect bionic customized body model by sequentially performing three-dimensional model reconstruction and reconstruction model optimization on the optimized bone defect bionic customized body point cloud, and further includes: a three-dimensional model construction unit 601 and a smooth filling unit 602;

[0088] The three-dimensional model building unit 601 is used to perform three-dimensional modeling on the optimized bone defect bionic customized body point cloud through a preset three-dimensional data processing model to obtain a second bone defect bionic customized body model;

[0089] The smoothing and filling unit 602 is used to obtain the first bone defect bionic customized body model by performing contour smoothing and simulated defect filling on the second bone defect bionic customized body model in sequence according to the model optimization algorithm.

[0090] The bone defect bionic customized body generation module 304 is used to generate a personalized bone defect bionic customized body according to the first bone defect bionic customized body model by using a solid bionic customized body generation technology.

[0091] As an example of this embodiment, as an example of this embodiment, see Figure 7 , Figure 7 A structural diagram of a bone defect bionic customized body generation module of a bone defect bionic customized body generation system provided in a certain embodiment of the present invention, such as Figure 7As shown, the bone defect bionic customized body generation module 304 is used to generate a personalized bone defect bionic customized body according to the first bone defect bionic customized body model through a solid bionic customized body generation technology, and also includes: a model evaluation unit 701, a defect simulation implantation unit 702 and a bionic customized body printing unit 703;

[0092] The model evaluation unit 701 is used to evaluate the first bone defect bionic customized body model based on the preset second quantitative index and the second qualitative index to obtain an evaluation result of the first bone defect bionic customized body model;

[0093] The defect simulation implantation unit 702 is used to obtain bone defect bionic customized body data by adjusting the defect position of the first bone defect bionic customized body model and performing defect simulation implantation according to the patient's bone defect data when the evaluation result of the first bone defect bionic customized body model meets the preset requirements;

[0094] The bionic customized body printing unit 703 is used to print the first bone defect bionic customized body model to generate a personalized bone defect bionic customized body according to the bone augmentation material through bio-printing technology.

[0095] The embodiment of the present invention proposes a bone defect bionic customized body generation system, which simplifies the technical steps of generating a bone defect bionic customized body designed before bone augmentation surgery into: simulating the defect of a non-defective bone model through a bone defect model construction module to obtain a bone defect model, then using a point cloud marking and optimization module to perform point cloud design and a first optimization on the bone defect model using point cloud technology, and performing three-dimensional reconstruction and a second optimization on the optimized bone defect bionic customized body point cloud through a model reconstruction and optimization module, and finally realizing the generation of a bone defect bionic customized body through a bone defect bionic customized body generation module; the process steps reduce the time and effort of the bionic customized body design of the bone defect before bone augmentation surgery. At the same time, the point cloud marking and optimization module is used to perform point cloud design, point cloud completion and point cloud optimization on the patient's bone defect model and the target defect repair model, so as to restore the patient's defect situation as close to the true value and ensure the accuracy of the patient's bone defect data acquisition. The model reconstruction and optimization module is then used to perform three-dimensional reconstruction and optimization of the optimized bone defect bionic customized body point cloud, which ensures the reliability of the bone defect bionic customized body generation results designed before bone augmentation surgery, thereby reducing the dependence on doctors' experience; and through the simplified bone defect bionic customized body generation method, the design threshold of bone augmentation preoperative design is lowered, and the technical limitations of bone defect completion are reduced.

[0096] Example 3

[0097] See also Fig. 9 , Fig. 9 A schematic diagram of an example 2 of a method for generating a bone defect bionic customized body provided by a certain embodiment of the present invention, as shown in FIG. Fig. 9The embodiment of the present invention also proposes a bone defect bionic customized body generation process using another point cloud completion network PMP-Net as an example.

[0098] Based on steps 101 and 102 of a method for generating a bionic customized body for bone defect proposed in an embodiment of the present invention, a plurality of bone defect models, target defect repair bionic customized body models and bone defect annotation point clouds are respectively obtained, and the point cloud of the bone defect model, the point cloud representing the true ideal bone defect repair bionic customized body model and the point cloud of the ideal bone defect repair bionic customized body model are normalized with the mass point as the center. For the specific acquisition method, see Example 1, which will not be repeated here;

[0099] Based on the preset first quantitative index and the first qualitative index, the point cloud completion model is trained and verified, and the bone defect annotated point cloud is completed and trained by the trained point cloud completion model to obtain the bone defect bionic customized volume point cloud; specifically, the PMP-Net point cloud completion network is used as the backbone, such as Fig. 9 The alveolar bone defect model point cloud (A) shown in the figure is used as input and the output is as follows: Fig. 9 The completed bone model point cloud (B) shown in the figure; the bone defect bionic customized body point cloud is directly completed on the basis of the alveolar bone defect model point cloud, so that the alveolar bone defect model point cloud directly has the ideal bone defect repair contour;

[0100] Based on the preset first quantitative index and the first qualitative index, the point cloud optimization model is trained and verified, and the bone defect bionic customized body point cloud is optimized and trained through the trained point cloud optimization model; specifically, the bone defect bionic customized body point cloud is optimized and trained through the optimized trained point cloud optimization model to obtain the optimized bone defect bionic customized body point cloud. The specific optimization training process is shown in Example 1, which will not be repeated here;

[0101] By reconstructing the three-dimensional model and optimizing the reconstructed model on the optimized and completed bone model point cloud, we can obtain Fig. 9 The first bone-complementing model (C) shown in the figure can be used for the auxiliary design of bone augmentation procedures such as personalized titanium mesh and non-absorbable barrier membrane, 3D printing, and doctor-patient communication. The specific three-dimensional model reconstruction and reconstruction model optimization implementation methods can be found in Example 1, which will not be repeated here;

[0102] By performing Boolean operations and model optimization on the first bone filling model and the alveolar bone defect model, the following Fig. 9 The first bone defect bionic customized body model (D) is shown. The first bone defect bionic customized body model can be used for 3D printing of bone defect bionic customized body.

[0103] As another example of this embodiment, Fig. 9The completed bone model point cloud (B) and Fig. 9 The alveolar bone defect model point cloud (A) shown in the figure is subjected to Boolean operation and post-processing to obtain a third bone defect bionic customized body model. Based on the preset first quantitative index and the first qualitative index, the point cloud optimization model is trained and verified, and the third bone defect bionic customized body model is optimized and trained by the trained point cloud optimization model to obtain the following: Fig. 9 The optimized bone defect bionic customized body point cloud (E) corresponding to the third bone defect bionic customized body model is shown. According to step 103 of a bone defect bionic customized body generation method proposed in an embodiment of the present invention, the first bone defect bionic customized body model is obtained by sequentially performing three-dimensional model reconstruction and reconstruction model optimization on the optimized bone defect bionic customized body point cloud corresponding to the third bone defect bionic customized body model.

[0104] A method for generating a bionic customized bone defect body proposed in an embodiment of the present invention performs point cloud design, point cloud completion and point cloud optimization on a patient's bone defect model and a target defect repair model, so as to restore the patient's defect condition as close to the true value as possible, thereby ensuring the accuracy of the patient's bone defect data acquisition, and then performs three-dimensional reconstruction and optimization on the optimized bone defect bionic customized body point cloud, thereby ensuring the reliability of the bone defect bionic customized body generation result designed before bone augmentation surgery.

[0105] Example 4

[0106] In this embodiment, the actual situation of a patient is processed by encapsulating a bone defect bionic customized body generation method proposed in an embodiment of the present invention into the alveolar bone defect intelligent auxiliary completion software. Figure 8 , Figure 8 The following is a schematic diagram of an example of a method for generating a bone defect bionic customized body according to an embodiment of the present invention. Figure 8 As shown,

[0107] This embodiment takes a patient with vertical alveolar bone defect of the upper front teeth as an example to demonstrate the clinical application process of this method: Description of the patient's condition: The patient, a male, has been missing his upper front teeth for 4 years. Because of the impact on chewing and aesthetics, he came to the hospital for consultation and implant restoration. Specialist examination: 11-21 is missing, and the alveolar bone volume is poor. CBCT shows: 11-21 absorbs the middle 1 / 2 of the root, showing vertical bone absorption. After implant simulation and patient communication, it is planned to perform personalized bone block transplantation, and a personalized bone block with a length of about 12mm, a height of about 6mm, and a width of about 7.5mm needs to be implanted.

[0108] Personalized bionic customized body design and application process:

[0109] (1) According to the method described in step 101 of Example 1, the oral CBCT image data of the patient is imported through Mimics, the maxillary anterior teeth area is segmented and three-dimensional reconstruction is performed, and further, the alveolar bone defect jaw model and the alveolar bone defect model point cloud are directly exported in stl format and txt format.

[0110] (2) Input the alveolar bone defect model point cloud into the upper anterior teeth vertical alveolar bone defect intelligent auxiliary completion software to perform the point cloud completion and point cloud optimization described in step 102 of Example 1 to obtain the predicted bone defect bionic customized body point cloud (equivalent to optimizing the bone defect bionic customized body point cloud).

[0111] (3) According to the method described in step 103 of Example 1, the point cloud of the bone defect bionic customized body is three-dimensionally reconstructed to obtain a fine bone defect bionic customized body model (equivalent to the first bone defect bionic customized body model).

[0112] (4) According to the method described in step 104 of Example 1, a resin 3D printer is used to print a fine bone defect bionic customized body model and a corresponding alveolar bone defect model, and a 3D printed fine bone defect bionic customized body resin model and an alveolar bone defect resin model are obtained as diagnostic models for communicating surgical plans with patients.

[0113] (5) After confirming the surgical plan with the patient, the bone augmentation material is used as the raw material, and the sinking DLP bioprinting technology is used to print a fine bone defect bionic customized body model for personalized bone transplantation surgery. Specifically, the fine bone defect bionic customized body model is imported into Materialise Magics, support is added, and the fine bone defect bionic customized body model with support is exported in STL format. The fine bone defect bionic customized body model with support is imported into Autodesk Netfebb for slicing, and the appropriate printing layer thickness, scaling size and ratio are set to obtain the printing file of the fine bone defect bionic customized body model. Prepare the bone augmentation material slurry, pour the slurry into the sinking DLP light-curing 3D printer slurry tank and set the printing parameters. Further, import the printing file, use the bone augmentation material as the raw material, use the sinking DLP bioprinting technology for printing, and rinse, dry, degrease and sinter, and finally obtain a 3D printed personalized bone defect bionic customized body.

[0114] Artificial intelligence is used for the preoperative design of bone augmentation surgery, and an AI-assisted alveolar bone defect repair full-process paradigm of "data collection-network construction-3D reconstruction and optimization-results analysis-application" is proposed, which can be used for the auxiliary design and 3D printing of bone augmentation procedures such as bone defect bionic customized bodies, personalized titanium meshes and non-absorbable barrier membranes, simplifying the preoperative design process, reducing the difficulty of preoperative design, and enabling inexperienced junior doctors and grassroots doctors to perform personalized bone augmentation design; a three-dimensional database of oral teeth and alveolar bones is constructed, and a deep learning-based point cloud completion model and a post-processing model for optimizing the point cloud of a bionic customized body for bone defects are proposed, which can efficiently and automatically generate and optimize the point cloud of a bionic customized body for bone defects; the deep learning-based point cloud completion model and the post-processing model for optimizing the point cloud of a bionic customized body for bone defects are integrated and a visualization module is added to further lower the threshold for use of this method, and the intelligent completion effect can be viewed in real time.

[0115] As another example of this embodiment, based on the application of the method proposed in this embodiment in the auxiliary design and 3D printing of bone augmentation procedures such as personalized titanium mesh and non-absorbable barrier membrane, personalized titanium mesh is used as a specific example as one of the explanations of this embodiment, specifically:

[0116] (1) A first bone defect bionic customized body model is obtained by using a bone defect bionic customized body generation method proposed in this embodiment.

[0117] A method for obtaining a first bone defect bionic customized body model can be referred to in Example 1, wherein the defect position of the first bone defect bionic customized body model is adjusted according to the patient's bone defect data so that it fits the patient's bone defect model. Boolean operations and post-processing are performed on the patient's bone defect model and the first bone defect bionic customized body model to obtain the first bone defect bionic customized body model.

[0118] Another method for obtaining the first bone defect bionic customized body model can be referred to Example 3, in which the first bone defect bionic customized body model is obtained by changing the backbone network of the point cloud completion model, verifying and training the point cloud completion model and the point cloud optimization model.

[0119] (2) Printing the first bone defect bionic customized body model through bio-printing technology to generate a first bone defect bionic customized body model entity.

[0120] (3) The titanium mesh is placed on the first bone defect bionic customized body model entity for cutting and pre-bending to obtain a pre-bent personalized titanium mesh that can achieve ideal bone augmentation. Compared with shaping the titanium mesh during surgery, pre-bending the titanium mesh using the first bone defect bionic customized body model entity in advance is conducive to reducing the operation time and difficulty.

[0121] An embodiment of the present invention also proposes a terminal for generating a bionic customized bone defect body, comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements the method for generating a bionic customized bone defect body when executing the computer program.

[0122] An embodiment of the present invention also proposes a storage medium for generating a bionic customized bone defect body, the storage medium comprising: a computer-readable storage medium and a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the method for generating a bionic customized bone defect body.

[0123] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

[0124] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example" or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine different embodiments or examples described in this specification and the features of different embodiments or examples, unless they are contradictory.

[0125] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

Claims

1. A method for generating a bionic customized bone defect body, characterized in that: include: Based on the patient's non-defective bone data, a bone defect model and a target defect repair bionic customized body model are obtained by performing defect simulation removal on the patient's non-defective bone model; wherein the target defect repair bionic customized body model is a true value model corresponding to the bone defect model; Based on the point cloud completion model and the point cloud optimization model, the bone defect annotated point cloud is respectively completed and optimized to obtain an optimized bone defect bionic customized volume point cloud; wherein the bone defect annotated point cloud is obtained by merging the marked point clouds corresponding to the bone defect model and the target defect repair bionic customized volume model; The first bone defect bionic customized body model is obtained by sequentially performing three-dimensional model reconstruction and reconstruction model optimization on the optimized bone defect bionic customized body point cloud; According to the first bone defect bionic customized body model, a personalized bone defect bionic customized body is generated by using a solid bionic customized body generation technology.

2. The method for generating a bone defect bionic customized body according to claim 1, characterized in that: Based on the patient's non-defective bone data, the defect simulation removal is performed on the patient's non-defective bone model to obtain a bone defect model and a target defect repair bionic customized body model, specifically: Acquiring the patient's non-defective bone data, and constructing a three-dimensional model of the patient's non-defective bone data to obtain a non-defective bone model of the patient; The defect site model is obtained by setting the defect site on the non-defective bone model of the patient; Based on the defect-free bone contour retained by the defect site model, a bone defect model is obtained by performing a true defect simulation on the defect site model; The target defect repair bionic customized body model is obtained by sequentially performing Boolean subtraction and post-processing on the defect site model and the bone defect model.

3. The method for generating a bone defect bionic customized body according to claim 1, characterized in that: Based on the point cloud completion model and the point cloud optimization model, the bone defect annotated point cloud is sequentially completed and optimized to obtain the optimized bone defect bionic customized volume point cloud, specifically: Performing point cloud marking on the bone defect model and the target defect repair bionic customized body model respectively, to obtain marked point clouds corresponding to the bone defect model and the target defect repair bionic customized body model; By merging the marked point clouds corresponding to the bone defect model and the target defect repair bionic customized body model, a bone defect marked point cloud is obtained; Based on the preset first quantitative index and the first qualitative index, the point cloud completion model is trained and verified, and the bone defect annotated point cloud is completed and trained by the trained point cloud completion model to obtain the bone defect bionic customized volume point cloud; Based on the preset first quantitative index and first qualitative index, the point cloud optimization model is trained and verified, and the bone defect bionic customized body point cloud is optimized and trained through the trained point cloud optimization model to obtain the optimized bone defect bionic customized body point cloud.

4. The method for generating a bone defect bionic customized body according to claim 1, characterized in that: The first bone defect bionic customized body model is obtained by sequentially performing three-dimensional model reconstruction and reconstruction model optimization on the optimized bone defect bionic customized body point cloud, specifically: The optimized bone defect bionic customized body point cloud is subjected to three-dimensional modeling by using a preset three-dimensional data processing model to obtain a second bone defect bionic customized body model; According to the model optimization algorithm, the first bone defect bionic customized body model is obtained by successively performing contour smoothing and simulated defect filling on the second bone defect bionic customized body model.

5. The method for generating a bone defect bionic customized body according to claim 1, characterized in that: The step of generating a personalized bone defect bionic customized body by using a solid bionic customized body generation technology according to the first bone defect bionic customized body model is specifically as follows: Based on the preset second quantitative index and the second qualitative index, the first bone defect bionic customized body model is evaluated to obtain an evaluation result of the first bone defect bionic customized body model; When the evaluation result of the first bone defect bionic customized body model meets the preset requirements, according to the patient's bone defect data, the defect position of the first bone defect bionic customized body model is adjusted and a defect simulation application is performed to obtain the bone defect bionic customized body data; According to the bone augmentation material, the first bone defect bionic customized body model is printed by bio-printing technology to generate a personalized bone defect bionic customized body.

6. A bone defect bionic customized body generation system, characterized in that: include: Bone defect model construction module, point cloud marking and optimization module, model reconstruction and optimization module and bone defect bionic customized body generation module; The bone defect model construction module is used to obtain a bone defect model and a target defect repair bionic customized body model based on the patient's defect-free bone data by performing defect simulation removal on the patient's defect-free bone model; wherein the target defect repair bionic customized body model is a true value model corresponding to the bone defect model; The point cloud labeling and optimization module is used to perform completion training and optimization training on the bone defect annotated point cloud in sequence based on the point cloud completion model and the point cloud optimization model, so as to obtain an optimized bone defect bionic customized volume point cloud; wherein the bone defect annotated point cloud is obtained by merging the labeled point clouds corresponding to the bone defect model and the target defect repair bionic customized volume model; The model reconstruction and optimization module is used to obtain a first bone defect bionic customized body model by sequentially performing three-dimensional model reconstruction and reconstruction model optimization on the optimized bone defect bionic customized body point cloud; The bone defect bionic customized body generation module is used to generate a personalized bone defect bionic customized body according to the first bone defect bionic customized body model through a solid bionic customized body generation technology.

7. The bone defect bionic customized body generation system according to claim 6, characterized in that: The bone defect model construction module is used to obtain a bone defect model and a target defect repair bionic customized body model by performing defect simulation removal on the patient's non-defective bone model based on the patient's non-defective bone data, and also includes: Three-dimensional reconstruction unit, defect site model building unit, bone defect true value simulation unit and Boolean processing unit; The three-dimensional reconstruction unit is used to obtain the patient's non-defective bone data, and to construct a three-dimensional model of the patient's non-defective bone data to obtain a patient's non-defective bone model; The defect site model construction unit is used to obtain a defect site model by setting a defect site on the patient's non-defective bone model; The bone defect true value simulation unit is used to obtain a bone defect model by performing true value defect simulation on the defect site model based on the defect-free bone contour retained by the defect site model; The Boolean processing unit is used to obtain a target defect repair bionic customized body model by performing Boolean subtraction and post-processing on the defect site model and the bone defect model in sequence.

8. The bone defect bionic customized body generation system according to claim 6, characterized in that: The point cloud marking and optimization module is used to perform completion training and optimization training on the bone defect annotated point cloud in sequence based on the point cloud completion model and the point cloud optimization model to obtain the optimized bone defect bionic customized body point cloud, and also includes: A first point cloud set acquisition unit, a second point cloud set acquisition unit, a completion training unit and an optimization training unit; The first point cloud set acquisition unit is used to perform point cloud marking on the bone defect model and the target defect repair bionic customized body model respectively, to obtain marked point clouds corresponding to the bone defect model and the target defect repair bionic customized body model; The second point cloud set acquisition unit is used to obtain a bone defect annotated point cloud by merging the marked point clouds corresponding to the bone defect model and the target defect repair bionic customized body model; The completion training unit is used to perform model training and verification on the point cloud completion model based on the preset first quantitative index and the first qualitative index, and to perform completion training on the bone defect annotated point cloud through the trained point cloud completion model to obtain the bone defect bionic customized volume point cloud; The optimization training unit is used to perform model training and verification on the point cloud optimization model based on the preset first quantitative index and the first qualitative index, and optimize the bone defect bionic customized body point cloud through the trained point cloud optimization model to obtain the optimized bone defect bionic customized body point cloud.

9. A terminal for generating a bone defect bionic customized body, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, a method for generating a bone defect bionic customized body as claimed in any one of claims 1 to 5 is implemented.

10. A storage medium for generating a bone defect bionic customized body, characterized in that: The storage medium comprises: a computer-readable storage medium and a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute a method for generating a bone defect bionic customized body as described in any one of claims 1 to 5.

Images