WO2023185405A1 - 义齿3d打印支架的设计方法、装置及可存储介质 - Google Patents

义齿3d打印支架的设计方法、装置及可存储介质 Download PDF

Info

Publication number
WO2023185405A1
WO2023185405A1 PCT/CN2023/080208 CN2023080208W WO2023185405A1 WO 2023185405 A1 WO2023185405 A1 WO 2023185405A1 CN 2023080208 W CN2023080208 W CN 2023080208W WO 2023185405 A1 WO2023185405 A1 WO 2023185405A1
Authority
WO
WIPO (PCT)
Prior art keywords
dimensional
denture
model
orthotopic
bracket
Prior art date
Application number
PCT/CN2023/080208
Other languages
English (en)
French (fr)
Inventor
孙玉春
陈虎
李骋
唐宝
庞恩林
翟文茹
Original Assignee
南京前知智能科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 南京前知智能科技有限公司 filed Critical 南京前知智能科技有限公司
Publication of WO2023185405A1 publication Critical patent/WO2023185405A1/zh

Links

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B29—WORKING OF PLASTICS; WORKING OF SUBSTANCES IN A PLASTIC STATE IN GENERAL
    • B29C—SHAPING OR JOINING OF PLASTICS; SHAPING OF MATERIAL IN A PLASTIC STATE, NOT OTHERWISE PROVIDED FOR; AFTER-TREATMENT OF THE SHAPED PRODUCTS, e.g. REPAIRING
    • B29C64/00—Additive manufacturing, i.e. manufacturing of three-dimensional [3D] objects by additive deposition, additive agglomeration or additive layering, e.g. by 3D printing, stereolithography or selective laser sintering
    • B29C64/30—Auxiliary operations or equipment
    • B29C64/386—Data acquisition or data processing for additive manufacturing
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B33—ADDITIVE MANUFACTURING TECHNOLOGY
    • B33Y—ADDITIVE MANUFACTURING, i.e. MANUFACTURING OF THREE-DIMENSIONAL [3-D] OBJECTS BY ADDITIVE DEPOSITION, ADDITIVE AGGLOMERATION OR ADDITIVE LAYERING, e.g. BY 3-D PRINTING, STEREOLITHOGRAPHY OR SELECTIVE LASER SINTERING
    • B33Y50/00—Data acquisition or data processing for additive manufacturing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/90—Determination of colour characteristics
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Definitions

  • the present disclosure relates to the technical field of denture design, and in particular to a design method and device for a denture 3D printing bracket, as well as storage media and electronic equipment.
  • a design method for a denture 3D printing bracket including: obtaining a three-dimensional dental model, performing orthotopic processing on the three-dimensional dental model to obtain an orthotopic three-dimensional dental model; obtaining the three-dimensional dental model; A denture bracket boundary model corresponding to the orthotopic three-dimensional dental model, and extracting physiological anatomical features corresponding to the denture bracket boundary model according to the denture bracket boundary model; converting the physiological anatomical features into corresponding RGB color features; wherein, The RGB color features are Red, Green, and Blue, that is, the three color features of red, green, and blue; expand the orthotopic three-dimensional dental model to obtain the corresponding two-dimensional image and the topology of the three-dimensional model to the two-dimensional image relationship, and transfer the RGB color features and the denture bracket boundary model to the two-dimensional image to obtain a new two-dimensional image; construct a feature recognition deep neural network, and input the orthotopic three-dimensional dental model into The feature recognition deep neural network performs prediction to obtain
  • the specific process of converting the physiological anatomical features into RGB color features includes: converting the orthotopic three-dimensional dental model into corresponding RGB color features through a multi-viewpoint illumination feature empowerment method.
  • the specific process of the multi-view lighting feature empowerment method includes: setting a fixed light source distribution, based on the relative positions of the orthotopic three-dimensional dental model and the fixed light source distribution and the orthotopic three-dimensional dental model.
  • the normal direction and RGB characteristics of each vertex are calculated, and the ADSF component of each vertex is calculated; wherein the ADSF component refers to ambient light reflection, diffuse reflection, specular reflection, and Fresnel reflection; the ADSF component is reconciled, Obtain the corresponding RGB color features.
  • the specific process of obtaining the final denture 3D printing bracket boundaries and labels includes: using the three-dimensional dental model, the new two-dimensional image, and the topological relationship between the three-dimensional model and the two-dimensional image as learning Database; input the orthotopic three-dimensional dental model into the feature recognition deep neural network for learning, map the results through the topological relationship between the three-dimensional model and the two-dimensional image, and obtain the final denture 3D printing bracket boundary and label ;Use the fixed distance expansion method of the graphics area to complete the design of the boundaries and labels of the denture 3D printing bracket and export the design file.
  • the denture bracket boundary model includes: any one or more of the upper and lower jaw retention network edges, major connector edges, small connector edges, clasp edges, and support edges.
  • a device utilizing the design method of a denture 3D printing bracket described in any one of the above including an alignment module, an extraction module, an RGB feature generation module, an image generation module, and a recognition module connected in sequence. Modules and design modules;
  • the alignment module is used to obtain a three-dimensional dental model, and the three-dimensional dental model is subjected to orthotopic processing to obtain a three-dimensional dental model;
  • the extraction module is used to obtain the orthotopic three-dimensional dental model corresponding to Denture bracket boundary model;
  • the RGB feature generation module is used to convert the orthotopic three-dimensional dental model into corresponding RGB color features;
  • the image generation module is used to unfold the orthotopic three-dimensional dental model to obtain The topological relationship between the corresponding two-dimensional image and the three-dimensional model and the two-dimensional image, and transfer the RGB color features and the denture bracket boundary model to the two-dimensional image to obtain a new two-dimensional image;
  • the identification module Used to construct a feature recognition deep neural network, input the orthotopic three-dimensional dental model into the feature recognition deep neural network for prediction, and obtain the final denture 3D printing bracket boundaries and labels;
  • the design module is used to calculate the Design of denture 3D printed bracket boundaries and labels.
  • a computer-readable storage medium is provided.
  • a computer program is stored on the computer-readable storage medium.
  • the computer program is executed by a processor, a denture as described in any one of the above is implemented. Design method of 3D printed bracket.
  • an electronic device including: a memory; and a device coupled to the memory.
  • a processor of memory the processor being configured to perform the method as described above based on instructions stored in the memory.
  • Figure 1 is a flow chart of a design method of a denture 3D printed bracket according to some exemplary embodiments
  • Figure 2 is a schematic structural diagram of a design device for a 3D printed denture bracket according to some exemplary embodiments
  • Figure 3a is a schematic orthographic view of a three-dimensional dental model according to some exemplary embodiments.
  • Figure 3b is a schematic orthographic view of a three-dimensional dental model according to some exemplary embodiments.
  • FIG. 4 shows a schematic structural diagram of an electronic device suitable for implementing embodiments of the present disclosure.
  • the present disclosure provides a design method, device, storage medium and electronic denture 3D printing bracket.
  • Equipment the machine learning algorithm for instance segmentation/semantic segmentation of images is applied to the UV unfolded graphics of the three-dimensional model; a deep neural network dedicated to the recognition of physiological and anatomical features of the teeth and jaws is established; the multi-view illumination characteristics of the bionic dragonfly compound eye are established
  • the enabling algorithm Bionic Dragonfly Compound-eye Full-view Illumination Rendering, BDCFIR
  • the disclosed design method, device and storage medium for denture 3D printing brackets can reduce excessive manual interactive design operations in design software, quickly improve the operator's design level and design efficiency, and improve the operator's design accuracy.
  • an embodiment of the present disclosure discloses a design method for a denture 3D printing bracket, which includes the following steps:
  • Step 101 Obtain a three-dimensional tooth and jaw model, perform orthotopic processing on the three-dimensional tooth and jaw model, and obtain an orthotopic three-dimensional tooth and jaw model.
  • Step 102 Obtain the denture bracket boundary model corresponding to the orthotopic three-dimensional dental model, and extract the physiological anatomical features corresponding to the denture bracket boundary model according to the denture bracket boundary model.
  • Step 103 Convert physiological anatomical features into corresponding RGB color features.
  • Step 104 Expand the orthotopic three-dimensional dental model to obtain the corresponding two-dimensional image and the topological relationship between the three-dimensional model and the two-dimensional image, and transfer the RGB color characteristics and the denture bracket boundary model to the two-dimensional image to obtain a new 2D image.
  • Step 105 Construct a feature recognition deep neural network, input the orthotopic three-dimensional dental model into the feature recognition deep neural network for prediction, obtain the final denture 3D printing bracket boundaries and labels, and implement the design based on the denture 3D printing bracket boundaries and labels.
  • the dental and jaw model is aligned.
  • the specific process includes: using manual interaction or key point identification algorithm Determine three points, the three points are: the front end point Pa of the retention network, the last end points Pb and Pc at both ends, and use these three points to determine the positioning plane of the coordinates.
  • the direction of the line connecting points Pb and Pc is the x-axis.
  • the z-axis of the plane finally constructs a local coordinate system (x, y, z) with Po as the origin to orient the model.
  • the UV unfolding process is to parameterize the triangular mesh and establish a one-to-one mapping with the parameter plane.
  • Each vertex obtains a UV parameter value (texture coordinate), and the vertex coordinates and texture coordinates are indirectly connected through the triangular patch. Together, the three-dimensional to two-dimensional topological mapping relationship is obtained.
  • the specific process of converting physiological anatomical features into RGB color features includes:
  • the color characteristics of the RGB three channels are defined as: the height value of the vertex of the orthodontic posterior dental model is used as the R channel Input, where the height value is the average height of the dental model analyzed and normalized through big data analysis; the vertex normal x value is input as the G channel, and the vertex normal y value is input as the B channel, where the vertex normal xy value is The normal value (after normalization) of a certain vertex of the orthodontic posterior jaw model along the xy axis;
  • the orthotopic three-dimensional dental model is converted into corresponding RGB color features through a multi-view lighting feature empowerment method.
  • the specific process of the multi-view lighting feature empowerment method includes:
  • Observed-color is the corresponding RGB color feature obtained by final observation; A, B, C and D are all coefficients; diffuse-color is the color feature after rendering through diffuse reflection lighting, and specular-color is after rendering through specular reflection lighting.
  • the color characteristics of fresnel-color are the color characteristics after rendering through fresnel reflection lighting, and ambient-color is the color characteristics after rendering through ambient reflection lighting.
  • Diffuse reflection When the medium surface is not smooth enough, parallel incident light is reflected in scattered directions after contacting the medium surface.
  • Specular reflection When the medium surface is smooth enough, the parallel incident light will still be reflected parallelly after contacting the medium surface, such as specular reflection, water surface reflection, etc.
  • Fresnel reflection which means that the amount of light reflected from the medium surface depends on the angle of observation. When the line of sight is perpendicular to the surface of the observed point, the reflectivity is the lowest; when the line of sight is not perpendicular to the surface of the observed point, the line of sight is different from the observed point. The smaller the angle between the point surfaces, the higher the reflectivity.
  • Ambient reflection which represents the amount of light scattered by a light source onto an object. It can describe the global lighting effect in the environment. It has nothing to do with the incident angle. The ambient light can establish a coefficient as a parameter relative to the light source system, such as 0.1.
  • Three-dimensional lighting rendering of the RGB color features of the edentulous model using multi-view light sources can make up for the shortcomings of the inability to extract RGB color features due to lack of light reflection in the undercuts of the dental model rendered by a single light source, ensuring that the RGB color features of each part of the dental model are
  • the extraction of color features effectively eliminates the impact of shadows and blind areas produced by a single light source on machine learning.
  • the multi-view light source point is located above the dental model, which simulates the multi-view observation of the dental model by dental experts and further generates light and dark features, which is beneficial to further processing and accuracy improvement of subsequent algorithms.
  • the specific process of obtaining the final denture 3D printed bracket boundaries and labels includes:
  • the orthotopic 3D dental model is input into the feature recognition deep neural network for learning, and the results are mapped through the topological relationship between the 3D model and the 2D image to obtain the final denture 3D printing bracket boundaries and labels; through the fixed distance expansion method of the graphics area , complete the design of the boundaries and labels of the denture 3D printing bracket and export the design file.
  • the denture bracket boundary model includes: any one or more of the upper and lower jaw retention network edges, major connector edges, small connector edges, clasp edges, and support edges. .
  • an embodiment of the present disclosure also provides a device that utilizes the design method of a denture 3D printing bracket in any one of the above embodiments, including an alignment module 21, an extraction module 22, and an RGB feature generation module connected in sequence.
  • the alignment module 21 is used to obtain a three-dimensional dental model, and performs orthotopic processing on the three-dimensional dental model to obtain a three-dimensional dental model;
  • the extraction module 22 is used to obtain the denture bracket boundary model corresponding to the orthotopic three-dimensional dental model;
  • RGB The feature generation module 23 is used to convert the orthotopic three-dimensional dental model into corresponding RGB color features;
  • the image generation module 24 is used to expand the orthotopic three-dimensional dental model to obtain the corresponding two-dimensional image and the three-dimensional model into a two-dimensional image.
  • the recognition module 25 is used to build a feature recognition deep neural network, and input the orthotopic three-dimensional dental model to the feature Identify the deep neural network for prediction and obtain the final denture 3D printing bracket boundary and label;
  • the design module 26 is used to implement the design based on the denture 3D printing bracket boundary and label.
  • Embodiments of the present disclosure also provide a computer-readable storage medium.
  • a computer program is stored on the computer-readable storage medium.
  • the computer program is executed by a processor, a design method for a denture 3D printing bracket as in any one of the above embodiments is implemented. .
  • the process of designing the denture 3D printing bracket provided by the embodiment of the present disclosure is as follows:
  • Experts use CAD software to design the denture 3D printing bracket, or directly mark the data on the three-dimensional dental model to outline the boundaries of the 3D printing bracket.
  • the 3D printing bracket is automatically designed and the dental model is aligned.
  • Use manual interaction or key point identification algorithm to determine three points (the front end point Pa of the retention network, the last end points Pb and Pc at both ends), and use these three points to determine the positioning plane of the coordinates.
  • the direction of the line connecting points Pb and Pc is the x-axis.
  • the z-axis of the plane finally constructs a local coordinate system (x, y, z) with Po as the origin to orient the model.
  • the color features of the RGB three channels are defined as: the height value of the vertex of the orthodontic posterior dental model is input as the R channel, where the height value is the average height of the dental model analyzed and normalized through big data analysis; the vertex The normal x value is input as the G channel, and the vertex normal y value is input as the B channel.
  • the vertex normal xy value is the normal value (after normalization) of a certain vertex of the orthopedic posterior jaw model along the xy axis. .
  • BDCFIR is used to perform three-dimensional lighting rendering of light and dark features to generate the final RGB features (RGB color features + light and dark features).
  • the light source point is first located at the origin of the dental rotation lighting model, and the first lighting rendering is performed at this position; a circle with the origin as the center and a radius of 25mm is divided into 16 parts on average, and the light sources are placed at this position.
  • a circle with the origin as the center and a radius of 25mm is divided into 16 parts on average, and the light sources are placed at this position.
  • perform the 2nd to 17th lighting renderings divide the circle with the origin as the center and the radius of 50mm into 12 equal parts, and place the light sources at this position respectively, and perform the 18th to 29th lighting renderings; set the origin as the center of the circle and the radius
  • the 75mm circle is divided into 8 parts on average, and the light sources are placed at this position to perform the 30th-37th lighting rendering.
  • the light source moves 25mm upward along the z-axis, and the 38th lighting rendering is performed 25mm above the origin z-axis; the circle with a center point 25mm above the origin z-axis and a radius of 25mm is divided into 16 parts on average, and the light sources are placed here. position, perform the 39th to 54th lighting renderings; divide the circle with a center 25mm above the origin z-axis and a radius of 50mm into 12 equal parts, and place the light sources at this position respectively, and perform the 55th to 66th lighting renderings; The center of the circle is 25mm above the z-axis of the origin, and the circle with a radius of 75mm is divided into 8 parts on average.
  • the light sources are placed at this position for the 67th-74th lighting rendering. Then the light source moves 50mm upward along the z-axis, and the 75th lighting rendering is performed 50mm above the origin z-axis; the circle with a center of 50mm above the origin z-axis and a radius of 25mm is divided into 16 parts on average, and the light sources are placed here.
  • the origin of the dental rotation illumination model is defined as: 20 mm above the z-axis of the center point of the maxillary and mandibular dental models.
  • specific Ground the center point of the maxillary and mandibular dental model is defined as: the arithmetic mean of the three-point coordinates composed of the most front end point of the retention network of the maxillary and mandibular dental model and the last end points of both ends.
  • the process of generating the final RGB features is: based on the dental model, the relative position of the fixed light source (fixed position, pure white light source), and the normal phase of each vertex on the dental model And the RGB characteristic color assigned by it, calculates the ADSF component of each vertex in the scene, that is, the color value of ambient light reflection, diffuse reflection, and specular reflection.
  • the material formula is reconciled by the following material formula:
  • 0.5, 0.4, and 0.6 are the parameters used by A, B, and C in formula (1) respectively in specific implementation cases, where ambient-color takes the constant 0.1.
  • the final color of each vertex in the scene can be obtained, that is, the final RGB feature (RGB color feature + light and dark feature);
  • the UV expansion process is to parameterize the triangular mesh and establish a one-to-one mapping with the parameter plane.
  • Each vertex obtains the UV parameter value (texture coordinates), and the vertex coordinates and texture coordinates are indirectly linked through the triangle patch, that is, The three-dimensional to two-dimensional topological mapping relationship is obtained;
  • the post-processing module of the three-dimensional denture 3D printing bracket automatically performs post-processing of the adduction of the rear edge of the retention network, Post-processing such as outward expansion of the rear edge of the large connector; through the fixed distance expansion algorithm of the graphics area, the boundaries of each part of the stent can be automatically overlapped; finally, through preset parameters, the tissue termination line, tissue termination point, and upper/mandibular retention network are automatically generated , large connectors, small connectors, snap rings, supports, etc., and complete pattern engraving. Finally, the denture 3D printing bracket file is exported, which can be directly used for 3D printing manufacturing, realizing efficient and automatic digital design of dentures.
  • FIG. 4 shows an electronic device that implements an embodiment of the present disclosure.
  • the electronic device may include a memory 401, a processor 402, a communication interface 403, and a bus 404.
  • the memory 401 is used to store instructions, and the processor 402 is coupled to the memory 401.
  • the processor 402 is configured to execute the above-mentioned denture 3D based on the instructions stored in the memory 401. Design methods for printing scaffolds.
  • the memory 401 can be a high-speed RAM memory, a non-volatile memory (non-volatile memory), etc.
  • the memory 401 can also be a memory array.
  • the storage 401 may also be divided into blocks, and the blocks may be combined into virtual volumes according to certain rules.
  • the processor 402 may be a central processing unit CPU, or an Application Specific Integrated Circuit (ASIC), or one or more integrated circuits configured to implement the design method of the denture 3D printing bracket of the present disclosure.
  • ASIC Application Specific Integrated Circuit
  • the design method, device, storage medium and electronic equipment of the denture 3D printing bracket provided by the above embodiments have the following beneficial effects:
  • Reverse engineering is suitable for personalized denture design, which is based on three-dimensional scanning image data of teeth and jaws composed of massive triangular faces.
  • image instance segmentation methods assisted by deep neural networks and segmentation and recognition models based on three-dimensional convolutional neural networks have problems such that detailed features are easily lost and the misclassification rate of dental three-dimensional landmarks is high.
  • RGB three-channel high-resolution rendering assignment of the spatial pose of triangular patches with high-frequency curvature changes was clarified, and a new composition rule of the deep neural network for intelligent recognition of three-dimensional landmarks of dental and jaw physiological anatomy was revealed: feature generation module, 2- Three-dimensional topology mapping module, learning prediction module.
  • the bidirectional reversible mapping mechanism of dental three-dimensional graphical data Pobject and two-dimensional imaging data pixel is clarified, and an adaptive visual distance orthogonal projection method assisted by image information entropy is innovatively proposed to construct a standardized multi-source dental biometric database with high information content.
  • Data management platform >300,000 sets of data. Exploring a new generation of big data-driven intelligent reasoning of denture morphology The theory and implementation method effectively build a bridge between the physiological and anatomical characteristics data of individual teeth and the experience and knowledge of denture design and manufacturing.
  • a dual-discriminant adversarial learning network reasoning model based on dental big data was constructed to realize the transformation of bionic dentures from morphological bionic design to functional bionic design.
  • embodiments of the present disclosure may be provided as methods, systems, or computer program products. Accordingly, the present disclosure may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment that combines software and hardware aspects. Furthermore, the present disclosure may take the form of a computer program product embodied on one or more computer-usable non-transitory storage media (including, but not limited to, disk memory, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein. .
  • These computer program instructions may also be stored in a computer-readable memory that causes a computer or other programmable data processing apparatus to operate in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including the instruction means, the instructions
  • the device implements the functions specified in a process or processes of the flowchart and/or a block or blocks of the block diagram.
  • the methods and systems of the present disclosure may be implemented in many ways.
  • the methods and systems of the present disclosure may be implemented through software, hardware, firmware, or any combination of software, hardware, and firmware.
  • the above order for the steps of the methods is for illustration only, and the steps of the methods of the present disclosure are not limited to the order specifically described above unless otherwise specifically stated.
  • the present disclosure may also be implemented as programs recorded in recording media, and these programs include machine-readable instructions for implementing methods according to the present disclosure.
  • the present disclosure also covers recording media storing programs for executing methods according to the present disclosure.

Abstract

一种义齿3D打印支架的设计方法、装置、可存储介质和电子设备，应用于义齿设计技术领域，其中方法包括以下步骤：获取三维牙颌模型，将三维牙颌模型进行正位处理，得到正位三维牙颌模型；获取正位三维牙颌模型对应的义齿支架边界模型，并根据义齿支架边界模型提取义齿支架边界模型对应的生理解剖特征；将生理解剖特征转换为对应的RGB颜色特征；将正位三维牙颌模型进行展开，得到对应的二维图像及三维模型到二维图像的拓扑关系；构建特征识别深度神经网络，将正位三维牙颌模型输入至特征识别深度神经网络进行预测，得到最终的义齿3D打印支架边界及标签，并根据义齿3D打印支架边界及标签实现设计。

Description

义齿3D打印支架的设计方法、装置及可存储介质
相关申请的交叉引用
本公开是以CN申请号为202210345563.9申请日为2022年4月2日的申请为基础，并主张其优先权，该CN申请的公开内容在此作为整体引入本公开中。
技术领域
本公开涉及义齿设计技术领域，特别涉及一种义齿3D打印支架的设计方法、装置及可存储介质和电子设备。
背景技术
目前，随着大数据时代的到来，基于机器学习的人工智能技术，尤其是人工神经网络技术飞速发展，相关研究已涉及多个口腔医疗领域，特别在口腔三维解剖特征的自动化分割与识别方面具有巨大潜力，可辅助口腔医生及技师完成繁琐重复的手工劳动、消除主观误差，更加高效精确地完成诊断与诊疗计划的制订。但是，由于口内表面黏膜平缓光滑，生理解剖特征区域无明显分界且无明显曲率变化，机器学习分割与识别困难，因此，目前三维牙颌模型解剖特征的分割与识别仅集中于牙齿于牙龈的分割，其余部分主要依赖用户交互。在相关技术中，尚未有三维牙颌模型解剖特征的自动分割与识别神经网络，而三维牙颌模型的生理解剖边界为口腔正畸、修复、外科等高效全自动数字化设计的基础。
发明内容
根据本公开的第一方面，提供一种义齿3D打印支架的设计方法，包括：获取三维牙颌模型，将所述三维牙颌模型进行正位处理，得到正位三维牙颌模型；获取所述正位三维牙颌模型对应的义齿支架边界模型，并根据所述义齿支架边界模型提取所述义齿支架边界模型对应的生理解剖特征；将所述生理解剖特征转换为对应的RGB颜色特征；其中，所述RGB颜色特征为Red、Green、Blue，即红、绿、蓝三色颜色特征；将所述正位三维牙颌模型进行展开，得到对应的二维图像及三维模型到二维图像的拓扑关系，并将所述RGB颜色特征及所述义齿支架边界模型转移至所述二维图像上，得到新的二维图像；构建特征识别深度神经网络，将所述正位三维牙颌模型输入至所述特征识别深度神经网络进行预测，得到最终的义齿3D打印支架边界及标签，并根据所述义齿3D打印支架边界及标签实现设 计。
在一些实施例中，将所述生理解剖特征转换为RGB颜色特征的具体过程包括：通过多视角光照特征赋能方法将所述正位三维牙颌模型转换为对应的RGB颜色特征。
在一些实施例中，多视角光照特征赋能方法的具体过程包括：设置一固定光源分布，根据所述正位三维牙颌模型与固定光源分布的相对位置以及所述正位三维牙颌模型上各个顶点的法向及RGB特征，计算所述各个顶点的ADSF分量；其中，所述ADSF分量指的是环境光反射、漫反射、高光反射、菲涅耳反射；对所述ADSF分量进行调和，得到对应的所述RGB颜色特征。
在一些实施例中，得到最终的义齿3D打印支架边界及标签的具体过程包括：将所述三维牙颌模型、所述新的二维图像及所述三维模型到二维图像的拓扑关系作为学习数据库；将所述正位三维牙颌模型输入至所述特征识别深度神经网络进行学习，将结果通过所述三维模型到二维图像的拓扑关系进行映射，得到最终的义齿3D打印支架边界及标签；通过图形区域固定距离膨胀方法，完成义齿3D打印支架边界及标签的设计并导出设计文件。
在一些实施例中，所述义齿支架边界模型包括：上颌及下颌的固位网边缘、大连接体边缘、小连接体边缘、卡环边缘、支托边缘中的任一种或任几种。
根据本公开的第二方面，提供一种利用上述任一项所述的义齿3D打印支架的设计方法的装置，包括依次连接的正位模块、提取模块、RGB特征生成模块、图像生成模块、识别模块以及设计模块；
所述正位模块用于获取三维牙颌模型，将所述三维牙颌模型进行正位处理，得到正位三维牙颌模型；所述提取模块用于获取所述正位三维牙颌模型对应的义齿支架边界模型；所述RGB特征生成模块用于对所述正位三维牙颌模型转换为对应的RGB颜色特征；所述图像生成模块用于将所述正位三维牙颌模型进行展开，得到对应的二维图像及三维模型到二维图像的拓扑关系，并将所述RGB颜色特征及所述义齿支架边界模型转移至所述二维图像上，得到新的二维图像；所述识别模块用于构建特征识别深度神经网络，将所述正位三维牙颌模型输入至所述特征识别深度神经网络进行预测，得到最终的义齿3D打印支架边界及标签；所述设计模块用于根据所述义齿3D打印支架边界及标签实现设计。
根据本公开的第三方面，提供一种计算机可读存储介质，所述计算机可读存储介质上存储计算机程序，所述计算机程序被处理器执行时实现如上述任一项所述的一种义齿3D打印支架的设计方法。
根据本公开的第四方面，提供公开一种电子设备，包括：存储器；以及耦接至所述存 储器的处理器，所述处理器被配置为基于存储在所述存储器中的指令，执行如上所述的方法。
附图说明
为了更清楚地说明本公开实施例或相关技术中的技术方案，下面将对实施例或相关技术描述中所需要使用的附图作一简单地介绍，显而易见地，下面描述中的附图仅仅是本公开的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其他的附图。
图1为根据一些示例性实施例示出的义齿3D打印支架的设计方法的流程图；
图2为根据一些示例性实施例示出的义齿3D打印支架的设计装置的结构示意图；
图3a为根据一些示例性实施例示出的的三维牙颌模型正位示意图；
图3b为根据一些示例性实施例示出的三维牙颌模型正位示意图；
图4示出了适于用来实现本公开实施例的电子设备的结构示意图。
具体实施方式
下面参照附图对本公开进行更全面的描述，其中说明本公开的示例性实施例。下面将结合本公开实施例中的附图，对本公开实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本公开一部分实施例，而不是全部的实施例。基于本公开中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本公开保护的范围。下面结合各个图和实施例对本公开的技术方案进行多方面的描述。
在发明人所知晓的相关技术中，计算机辅助设计与制造、已广泛应用于牙科的各个领域，但是多为半自动的数字化流程，由模型扫描仪扫描石膏模型或口内扫描仪扫描患者牙列、获取三维数据后，由技工在CAD软件上进行手工交互设计，需要使用鼠标及键盘不断点击与输入。半自动的数字化流程虽然能够解决数字化设计的要求，但由于在设计软件中过多的手工交互设计操作，一是导致设计效率较低；二是过多的手工交互设计操作导致设计精度不高、设计效果过渡依赖操作者的经验和操作技巧，有时需要重新设计才能完成设计任务；三是专家经验需要较长时间积累，初级技工需要记忆大量案例才能设计合格的产品。因此，如何提供一种能够解决上述问题的义齿3D打印支架的设计方法是本领域技术人员亟需解决的问题。
有鉴于此，本公开提供了一种义齿3D打印支架的设计方法、装置及可存储介质和电子 设备，将用于图像的实例分割/语义分割的机器学习算法应用在了三维模型UV展开的图形上；建立了牙颌生理解剖特征识别专用深度神经网络；建立了仿生蜻蜓复眼的多视角光照特征赋能算法(Bionic Dragonfly Compound-eye Full-view Illumination Rendering，BDCFIR)模拟牙科专家肉眼多角度观察以进行牙模特征提取。
本公开的义齿3D打印支架的设计方法、装置及可存储介质，可减少设计软件中过多的手工交互设计操作，快速提高操作人员的设计水平与设计效率，提高操作人员的设计精度。
参见附图1所示，本公开的实施例公开了一种义齿3D打印支架的设计方法，包括以下步骤：
步骤101，获取三维牙颌模型，将三维牙颌模型进行正位处理，得到正位三维牙颌模型。
步骤102，获取正位三维牙颌模型对应的义齿支架边界模型，并根据义齿支架边界模型提取义齿支架边界模型对应的生理解剖特征。
步骤103，将生理解剖特征转换为对应的RGB颜色特征。
步骤104，将正位三维牙颌模型进行展开，得到对应的二维图像及三维模型到二维图像的拓扑关系，并将RGB颜色特征及义齿支架边界模型转移至二维图像上，得到新的二维图像。
步骤105，构建特征识别深度神经网络，将正位三维牙颌模型输入至特征识别深度神经网络进行预测，得到最终的义齿3D打印支架边界及标签，并根据义齿3D打印支架边界及标签实现设计。
在一些实施例中，参见附图3a-b所示，为了有利于牙齿模型在位置一致的条件下进行统一处理，对牙颌模型进行正位，具体过程包括：采用手动交互或关键点识别算法确定三点，三点为：固位网最前端点Pa、两端最后端点Pb及Pc，并以此三点确定坐标的定位平面。在此平面上，Pb与Pc点连线方向为x轴，计算Pa点到x轴线的垂线为y轴，交于一点Po，为局部坐标系的原点，进一步获得通过Po点且垂直于定位平面的z轴，最终构建出以Po为原点的局部坐标系(x，y，z)对模型进行正位。
在本公开的一些实施例中，UV展开过程即将三角网格参数化，与参数平面建立一一映射，每个顶点获得UV参数值(纹理坐标)，顶点坐标与纹理坐标通过三角面片间接联系起来，即获得了三维到二维的拓扑映射关系。
在本公开的一些实施例中，将生理解剖特征转换为RGB颜色特征的具体过程包括：
首先，RGB三通道的颜色特征定义为：正位后牙颌模型面片顶点的高度值作为R通道 输入，其中高度值为通过大数据分析的并归一化处理的牙颌模型平均高度；顶点法向x值作为G通道输入，顶点法向y值作为B通道输入，其中顶点法向xy值为正位后牙颌模型某顶点的法向值(归一化后)沿xy轴上的分量；
在本公开的一些实施例中，通过多视角光照特征赋能方法将正位三维牙颌模型转换为对应的RGB颜色特征。多视角光照特征赋能方法的具体过程包括：
设置一固定光源分布，根据正位三维牙颌模型与固定光源分布的相对位置以及正位三维牙颌模型上各个顶点的法向及RGB特征，计算各个顶点的ADSF分量；
对ADSF分量进行调和，得到对应的RGB颜色特征；
调和的具体公式为：
Observed-color＝A*diffuse-color+B*specular-color+C*fresnel-color+D*ambient-color    (1)；
其中，Observed-color为最终观察得到对应的RGB颜色特征；A、B、C和D都为系数；diffuse-color为通过diffuse reflection光照渲染后的颜色特征，specular-color为通过specular reflection光照渲染后的颜色特征，fresnel-color为通过fresnel reflection光照渲染后的颜色特征，ambient-color为通过ambient reflection光照渲染后的颜色特征。
Diffuse reflection：漫反射，当介质表面不够光滑时，平行入射光线在接触介质表面后被以分散的方向反射出去。
Specular reflection：高光反射，当介质表面足够光滑时，平行入射光线在接触介质表面后仍然被平行的反射出去，如镜面反射，水面反射等。
在同一环境条件下，specular reflection的高光小而亮，diffuse reflection的高光大而暗，但两种reflection的光总量相等。
Fresnel reflection：菲涅耳反射，表示从介质表面反射的光量取决于观察的角度，当视线垂直于被观测点表面时，反射率最低；当视线不垂直于被观测点表面时，视线与被观测点表面夹角越小，反射率越高。
Ambient reflection：环境光反射，表示光源散射到物体上光的量，可以描述环境中的全局照明效果，它和入射角度没有关系。环境光可以建立一个与光源系比值的系数作为参数，例如0.1。
通过多视角的光源对无牙颌模型RGB颜色特征进行三维光照渲染，可弥补单光源渲染的牙模倒凹处由于无光反射而无法提取RGB颜色特征的缺点，保证了牙颌模型各部位RGB 颜色特征的提取，有效地消除了单光源所产生的阴影和盲区对机器学习的影响。多视角的光源点位于牙颌模型的上方，模拟了牙科专家对牙颌模型多视角的观察，并进一步地生成明暗特征，有利于后续算法的进一步处理及精确度提高。
在本公开的一些实施例中，得到最终的义齿3D打印支架边界及标签的具体过程包括：
将三维牙颌模型、新的二维图像及三维模型到二维图像的拓扑关系作为学习数据库；
将正位三维牙颌模型输入至特征识别深度神经网络进行学习，将结果通过三维模型到二维图像的拓扑关系进行映射，得到最终的义齿3D打印支架边界及标签；通过图形区域固定距离膨胀方法，完成义齿3D打印支架边界及标签的设计并导出设计文件。
在本公开的一些实施例中，义齿支架边界模型包括：上颌及下颌的固位网边缘、大连接体边缘、小连接体边缘、卡环边缘、支托边缘中的任一种或任几种。
参见附图2所示，本公开实施例还提供一种利用上述实施例中任一项的义齿3D打印支架的设计方法的装置，包括依次连接的正位模块21、提取模块22、RGB特征生成模块23、图像生成模块24、识别模块25以及设计模块26；
正位模块21用于获取三维牙颌模型，将三维牙颌模型进行正位处理，得到正位三维牙颌模型；提取模块22用于获取正位三维牙颌模型对应的义齿支架边界模型；RGB特征生成模块23用于对正位三维牙颌模型转换为对应的RGB颜色特征；图像生成模块24用于将正位三维牙颌模型进行展开，得到对应的二维图像及三维模型到二维图像的拓扑关系，并将RGB颜色特征及义齿支架边界模型转移至二维图像上，得到新的二维图像；识别模块25用于构建特征识别深度神经网络，将正位三维牙颌模型输入至特征识别深度神经网络进行预测，得到最终的义齿3D打印支架边界及标签；设计模块26用于根据义齿3D打印支架边界及标签实现设计。
本公开实施例还提供一种计算机可读存储介质，计算机可读存储介质上存储计算机程序，计算机程序被处理器执行时实现如上述实施例中任一项的一种义齿3D打印支架的设计方法。
在一些实施例中，计算机程序被处理器执行时实现本公开实施例提供的义齿3D打印支架的设计方法过程如下：
(1)获取三维扫描获得上颌及下颌的3D牙颌模型(M)：
专家通过CAD软件进行义齿3D打印支架的设计，或直接在三维牙颌模型上进行数据标注勾画出3D打印支架的边界。导出上颌及下颌的3D牙颌模型及其专家标注的义齿3D打印支架各部分的边缘信息，包括但不限于上颌及下颌的固位网边缘、大连接体边缘、小 连接体边缘、卡环边缘、支托边缘等；
(2)对3D牙颌模型(M)进行正位处理：
在位置一致的条件下进行3D打印支架的自动设计，对牙颌模型进行正位。采用手动交互或关键点识别算法确定三点(固位网最前端点Pa、两端最后端点Pb及Pc)，并以此三点确定坐标的定位平面的方法。在此平面上，Pb与Pc点连线方向为x轴，计算Pa点到x轴的垂线为y轴，交于一点Po，为局部坐标系的原点，进一步获得通过Po点且垂直于定位平面的z轴，最终构建出以Po为原点的局部坐标系(x，y，z)对模型进行正位。
(3)对正位三维牙颌模型转换为对应的RGB颜色特征；
首先，RGB三通道的颜色特征定义为：正位后牙颌模型面片顶点的高度值作为R通道输入，其中高度值为通过大数据分析的并归一化处理的牙颌模型平均高度；顶点法向x值作为G通道输入，顶点法向y值作为B通道输入，其中顶点法向xy值为正位后牙颌模型某顶点的法向值(归一化后)沿xy轴上的分量。
其次，对正位后的带有RGB颜色特征的牙颌模型倒凹，通过BDCFIR进行三维光照渲染明暗特征，生成最终的RGB特征(RGB颜色特征+明暗特征)。
在一些实施例中，光源点首先位于牙科专用旋转光照模型原点，在此位置上进行第1次光照渲染；将原点为圆心，半径25mm的圆，平均分为16份，光源分别放置在此位置上，进行第2-17次光照渲染；将原点为圆心，半径50mm的圆，平均分为12份，光源分别放置在此位置上，进行第18-29次光照渲染；将原点为圆心，半径75mm的圆，平均分为8份，光源分别放置在此位置上，进行第30-37次光照渲染。然后光源沿着z轴向上方移动25mm，在原点z轴上方25mm处进行第38次光照渲染；将原点z轴上方25mm为圆心，半径25mm的圆，平均分为16份，光源分别放置在此位置上，进行第39-54次光照渲染；将原点z轴上方25mm为圆心，半径50mm的圆，平均分为12份，光源分别放置在此位置上，进行第55-66次光照渲染；将原点z轴上方25mm为圆心，半径75mm的圆，平均分为8份，光源分别放置在此位置上，进行第67-74次光照渲染。然后光源沿着z轴向上方移动50mm，在原点z轴上方50mm处进行第75次光照渲染；将原点z轴上方50mm为圆心，半径25mm的圆，平均分为16份，光源分别放置在此位置上，进行第76-91次光照渲染；将原点z轴上方50mm为圆心，半径50mm的圆，平均分为12份，光源分别放置在此位置上，进行第92-103次光照渲染；将原点z轴上方50mm为圆心，半径75mm的圆，平均分为8份，光源分别放置在此位置上，进行第104-111次光照渲染。
牙科专用旋转光照模型原点定义为：上颌及下颌牙颌模型中心点z轴上方20mm。具体 地，上颌及下颌牙颌模型中心点定义为：上颌及下颌牙颌模型固位网最前端点、两端最后端点组成的三点坐标的算术平均。
在一些实施例中，生成最终的RGB特征(RGB颜色特征+明暗特征)过程为：根据牙颌模型，固定光源(位置固定，纯白色光源)的相对位置，以及牙颌模型上各个顶点的法相和其赋予的RGB特征颜色，计算出在该场景下的各个顶点的ADSF分量，即环境光反射，漫反射，镜面反射的颜色值。
在一些实施例中，通过以下的材质公式调和，材质公式为：
Observed-color＝0.5*diffuse-color+0.4*specular-color+0.6*fresnel-color+0.1      (2)；
在公式(2)中，0.5,0.4,0.6分别是公式(1)中A、B、C在具体实施案例中采用的参数,其中ambient-color取常量0.1。
基于上述的公式能够得到该场景下的各个顶点最终颜色，即最终的RGB特征(RGB颜色特征+明暗特征)；
(4)将正位三维牙颌模型进行UV展开，得到对应的二维图像及三维模型到二维图像的拓扑关系，并将RGB颜色特征及义齿支架边界模型转移至二维图像上，得到新的二维图像；
在一些实施例中，uv展开过程即将三角网格参数化，与参数平面建立一一映射，每个顶点获得uv参数值(纹理坐标)，顶点坐标与纹理坐标通过三角面片间接联系起来，即获得了三维到二维的拓扑映射关系；
构建特征识别深度神经网络，将正位三维牙颌模型输入至特征识别深度神经网络进行预测，得到最终的义齿3D打印支架边界及标签；
通过上颌及下颌牙颌模型固位网最前端点、两端最后端点组成的三点定位坐标系，三维牙颌专用的义齿3D打印支架后处理模块自动进行固位网后缘内收后处理、大连接体后缘外扩等后处理；通过图形区域固定距离膨胀算法，可自动将支架各部分边界重叠；最终通过预设参数，自动生成组织终止线、组织终止点、上/下颌固位网、大连接体、小连接体、卡环、支托等，并完成花纹雕刻。最终导出义齿3D打印支架文件，能直接用于3D打印制造，实现了义齿的高效、自动数字化设计。
在一些实施例中，图4为实现本公开实施例的电子设备，该电子设备可包括存储器401、处理器402、通信接口403以及总线404。存储器401用于存储指令，处理器402耦合到存储器401，处理器402被配置为基于存储器401存储的指令执行实现上述的义齿3D 打印支架的设计方法。
存储器401可以为高速RAM存储器、非易失性存储器(non-volatile memory)等，存储器401也可以是存储器阵列。存储器401还可能被分块，并且块可按一定的规则组合成虚拟卷。处理器402可以为中央处理器CPU，或专用集成电路ASIC(Application Specific Integrated Circuit)，或者是被配置成实施本公开的义齿3D打印支架的设计方法的一个或多个集成电路。
上述实施例提供的义齿3D打印支架的设计方法、装置及可存储介质和电子设备具备如下有益效果：
(1)阐明了虚拟牙颌宏微观功能几何特征的归一化数学表达新原理
宏微观几何特征的充分必要数学表达，是个体牙颌生理解剖标志智能识别和提取的前提，却是以“经验+技巧”为基石的义齿经典设计理论研究的空白。研究发现基于齿间相切约束的二阶B样条函数可描述横纵向牙列曲线特征，单轨扫描曲面、UV放样曲面等函数可描述牙齿、牙龈、牙槽嵴的宏观形状特征。但牙尖、牙窝、牙沟、牙嵴、咬合磨耗曲面、邻接曲面等与义齿咀嚼等功能密切相关的细节几何特征，难以用宏观描述体系精准表达。研究发现“高度和法向”是牙颌三角面片(虚拟牙颌表面的最小构成单元，单颌模型面片数量通常>10万个)空间位、姿的最高权重特征向量，而RGB三通道色彩饱和度可对海量三角面片空间位姿进行快速精确表达，阐明了基于三通道细分色阶的牙颌功能几何特征的归一化数学表达原理。
(2)揭示了个性牙颌生理解剖标志智能识别深度神经网络的构成新法则
逆向工程适用于个性化义齿设计，其基础是海量三角面片构成的牙颌三维扫描图像数据。但针对相关深度神经网络辅助的图像实例分割方法和基于三维卷积神经网络的分割与识别模型存在细节特征易丢失、牙颌三维标志误分类率高的问题。创新性提出仿蜻蜓复眼的牙颌模型三维旋转光照渲染策略，模拟医学专家眼-脑-手的高效协同观测行为。进一步，阐明了高频曲率变化三角面片的空间位姿的RGB三通道高分辨率渲染赋值原理，揭示了牙颌生理解剖三维标志智能识别的深度神经网络全新构成法则：特征生成模块、二-三维拓扑映射模块、学习预测模块。
(3)阐明了多源异构大数据的知识表征驱动的仿生义齿结构推理新机制
阐明了牙颌三维图形化数据Pobject与二维图像化数据pixel的双向可逆映射机理，创新性提出图像信息熵辅助的自适应视距正交投影方法，构建规范高信息量多源牙齿生物特征大数据管理平台(>30万套数据)。探索了新一代大数据驱动的义齿形态智能化推理理 论和实现方法，有效搭建个体牙颌生理解剖特征数据与义齿设计制造经验知识之间的桥梁。构建基于牙颌大数据的双判别对抗学习网络推理模型，实现了仿生义齿从形态仿生设计到功能仿生设计的转变。
本说明书中各个实施例采用递进的方式描述，每个实施例重点说明的都是与其他实施例的不同之处，各个实施例之间相同相似部分互相参见即可。对于实施例公开的装置而言，由于其与实施例公开的方法相对应，所以描述的比较简单，相关之处参见方法部分说明即可。
本领域内的技术人员应当明白，本公开的实施例可提供为方法、系统、或计算机程序产品。因此，本公开可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且，本公开可采用在一个或多个其中包含有计算机可用程序代码的计算机可用非瞬时性存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本公开是参照根据本公开实施例的方法、设备(系统)和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
可能以许多方式来实现本公开的方法和系统。例如，可通过软件、硬件、固件或者软件、硬件、固件的任何组合来实现本公开的方法和系统。用于方法的步骤的上述顺序仅是为了进行说明，本公开的方法的步骤不限于以上具体描述的顺序，除非以其它方式特别说明。此外，在一些实施例中，还可将本公开实施为记录在记录介质中的程序，这些程序包括用于实现根据本公开的方法的机器可读指令。因而，本公开还覆盖存储用于执行根据本公开的方法的程序的记录介质。
对所公开的实施例的上述说明，使本领域专业技术人员能够实现或使用本公开。对这些实施例的多种修改对本领域的专业技术人员来说将是显而易见的，本文中所定义的一般 原理可以在不脱离本公开的精神或范围的情况下，在其它实施例中实现。因此，本公开将不会被限制于本文所示的这些实施例，而是要符合与本文所公开的原理和新颖特点相一致的最宽的范围。

Claims (8)

  1. 一种义齿3D打印支架的设计方法，包括：
    获取三维牙颌模型，将所述三维牙颌模型进行正位处理，得到正位三维牙颌模型；
    获取所述正位三维牙颌模型对应的义齿支架边界模型，并根据所述义齿支架边界模型提取所述义齿支架边界模型对应的生理解剖特征；
    将所述生理解剖特征转换为对应的RGB颜色特征；
    将所述正位三维牙颌模型进行展开，得到对应的二维图像及三维模型到二维图像的拓扑关系，并将所述RGB颜色特征及所述义齿支架边界模型转移至所述二维图像上，得到新的二维图像；
    构建特征识别深度神经网络，将所述正位三维牙颌模型输入至所述特征识别深度神经网络进行预测，得到最终的义齿3D打印支架边界及标签，并根据所述义齿3D打印支架边界及标签实现设计。
  2. 根据权利要求1所述的义齿3D打印支架的设计方法，所述将所述生理解剖特征转换为RGB颜色特征包括：
    通过多视角光照特征赋能方法将所述正位三维牙颌模型转换为对应的RGB颜色特征。
  3. 根据权利要求2所述的义齿3D打印支架的设计方法，多视角光照特征赋能方法的具体过程包括：
    设置一固定光源分布，根据所述正位三维牙颌模型与固定光源分布的相对位置以及所述正位三维牙颌模型上各个顶点的法向及RGB特征，计算所述各个顶点的ADSF分量；
    对所述ADSF分量进行调和，得到对应的所述RGB颜色特征。
  4. 根据权利要求1所述的义齿3D打印支架的设计方法，所述得到最终的义齿3D打印支架边界及标签包括：
    将所述三维牙颌模型、所述新的二维图像及所述三维模型到二维图像的拓扑关系作为学习数据库；
    将所述正位三维牙颌模型输入至所述特征识别深度神经网络进行学习，将结果通过所述三维模型到二维图像的拓扑关系进行映射，得到最终的义齿3D打印支架边界及标签；通过图形区域固定距离膨胀方法，完成义齿3D打印支架边界及标签的设计并导出设计文件。
  5. 根据权利要求1-4任一项所述的义齿3D打印支架的设计方法，其中，所述义齿支 架边界模型包括：上颌及下颌的固位网边缘、大连接体边缘、小连接体边缘、卡环边缘、支托边缘中的任一种或任几种。
  6. 一种利用权利要求1-5任一项所述的义齿3D打印支架的设计方法的装置，包括正位模块、提取模块、RGB特征生成模块、图像生成模块、识别模块以及设计模块；
    所述正位模块用于获取三维牙颌模型，将所述三维牙颌模型进行正位处理，得到正位三维牙颌模型；所述提取模块用于获取所述正位三维牙颌模型对应的义齿支架边界模型；
    所述RGB特征生成模块用于对所述正位三维牙颌模型转换为对应的RGB颜色特征；
    所述图像生成模块用于将所述正位三维牙颌模型进行展开，得到对应的二维图像及三维模型到二维图像的拓扑关系，并将所述RGB颜色特征及所述义齿支架边界模型转移至所述二维图像上，得到新的二维图像；
    所述识别模块用于构建特征识别深度神经网络，将所述正位三维牙颌模型输入至所述特征识别深度神经网络进行预测，得到最终的义齿3D打印支架边界及标签；
    所述设计模块用于根据所述义齿3D打印支架边界及标签实现设计。
  7. 一种计算机可读存储介质，所述计算机可读存储介质上存储计算机程序，所述计算机程序被处理器执行时实现如权利要求1至5中任一项所述的义齿3D打印支架的设计方法。
  8. 一种电子设备，包括：
    存储器；以及耦接至所述存储器的处理器，所述处理器被配置为基于存储在所述存储器中的指令，执行如权利要求1至5中任一项所述的方法。
PCT/CN2023/080208 2022-04-02 2023-03-08 义齿3d打印支架的设计方法、装置及可存储介质 WO2023185405A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210345563.9 2022-04-02
CN202210345563.9A CN114714626B (zh) 2022-04-02 2022-04-02 一种义齿3d打印支架的设计方法、装置及可存储介质

Publications (1)

Publication Number Publication Date
WO2023185405A1 true WO2023185405A1 (zh) 2023-10-05

Family

ID=82241434

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/080208 WO2023185405A1 (zh) 2022-04-02 2023-03-08 义齿3d打印支架的设计方法、装置及可存储介质

Country Status (2)

Country Link
CN (1) CN114714626B (zh)
WO (1) WO2023185405A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114714626B (zh) * 2022-04-02 2023-08-15 北京大学口腔医学院 一种义齿3d打印支架的设计方法、装置及可存储介质

Citations (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105411717A (zh) * 2014-09-19 2016-03-23 北京大学口腔医学院 一种计算机辅助设计的个体化牙弓夹板及其制作方法
CN109528330A (zh) * 2018-12-20 2019-03-29 成都登特牙科技术开发有限公司 可摘局部义齿支架的数字化设计方法
US20190105134A1 (en) * 2017-10-07 2019-04-11 Henry Schein, Inc. Denture fabrication
CN110251276A (zh) * 2019-06-03 2019-09-20 浙江工业大学 一种增减材复合加工成型的口腔赝复体支架制作方法
US20190336254A1 (en) * 2018-05-03 2019-11-07 Dentsply Sirona Inc. Methods of three-dimensional printing for fabricating a dental appliance
CN110490966A (zh) * 2018-05-14 2019-11-22 重庆工港致慧增材制造技术研究院有限公司 生物3d打印正向支架模型设计方法
CN111513881A (zh) * 2020-05-09 2020-08-11 北京大学口腔医学院 一种上颌骨缺损赝复体的制作方法及系统
CN111529105A (zh) * 2020-05-19 2020-08-14 北京联袂义齿技术有限公司 一种全口义齿立体支架的数字化制造方法
CN111941828A (zh) * 2020-08-12 2020-11-17 北京大学口腔医学院 咬合调整器的数字化实现方法、装置、设备及存储介质
CN112215065A (zh) * 2020-09-04 2021-01-12 北京大学口腔医学院 一种牙颌边界特征自动化识别方法
US20210153985A1 (en) * 2017-07-07 2021-05-27 Dio Corporation Digital denture manufacturing method and manufacturing system, and denture hole guider applied thereto and manufacturing method thereof
CN112895459A (zh) * 2021-01-14 2021-06-04 南京前知智能科技有限公司 一种基于设计组件信息的3d打印模型智能预处理方法及装置
CN113171188A (zh) * 2021-01-29 2021-07-27 正雅齿科科技(上海)有限公司 一种具有硬腭区域的数字化牙颌模型构建方法及系统
CN113288480A (zh) * 2021-06-25 2021-08-24 成都登特牙科技术开发有限公司 骨增量与种植义齿联合设计方法及骨增量模型的制造方法
CN113397742A (zh) * 2021-07-29 2021-09-17 维视医疗信息科技(山东)有限公司 可摘局部义齿支架模型自动生成方法及系统
CN113397743A (zh) * 2021-06-04 2021-09-17 珠海市三通陶齿有限公司 一种牙支持式可摘局部义齿制作方法
CN114714626A (zh) * 2022-04-02 2022-07-08 北京大学口腔医学院 一种义齿3d打印支架的设计方法、装置及可存储介质

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8442283B2 (en) * 2006-08-30 2013-05-14 Anatomage Inc. Patient-specific three-dimensional dentition model
CN105581849A (zh) * 2014-10-23 2016-05-18 北京大学口腔医学院 多色树脂义齿修复体一体化快速成型方法
US11049606B2 (en) * 2018-04-25 2021-06-29 Sota Precision Optics, Inc. Dental imaging system utilizing artificial intelligence
CN109223216A (zh) * 2018-09-27 2019-01-18 北京大学口腔医学院 一种可摘局部义齿的高效率数字设计方法
EP4096569A1 (en) * 2020-01-31 2022-12-07 James R. Glidewell Dental Ceramics, Inc. Teeth segmentation using neural networks
CN111261287B (zh) * 2020-02-26 2022-10-21 中国人民解放军第四军医大学 一种种植方案设计方法及系统、终端和计算机可读存储介质
WO2021210966A1 (ko) * 2020-04-16 2021-10-21 이마고웍스 주식회사 딥러닝을 이용한 3차원 의료 영상 데이터의 특징점 자동 검출 방법 및 장치, 치과용 3차원 데이터 위치 정렬 자동화 방법, 치과용 3차원 데이터 위치 정렬 자동화 방법, 치과용 3차원 스캔 데이터의 랜드마크 자동 검출 방법, 3차원 치과 ct 영상과 3차원 디지털 인상 모델의 정합 정확도 판단 방법 및 상기 방법들을 컴퓨터에서 실행시키기 위한 프로그램이 기록된 컴퓨터로 읽을 수 있는 기록 매체

Patent Citations (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105411717A (zh) * 2014-09-19 2016-03-23 北京大学口腔医学院 一种计算机辅助设计的个体化牙弓夹板及其制作方法
US20210153985A1 (en) * 2017-07-07 2021-05-27 Dio Corporation Digital denture manufacturing method and manufacturing system, and denture hole guider applied thereto and manufacturing method thereof
US20190105134A1 (en) * 2017-10-07 2019-04-11 Henry Schein, Inc. Denture fabrication
US20190336254A1 (en) * 2018-05-03 2019-11-07 Dentsply Sirona Inc. Methods of three-dimensional printing for fabricating a dental appliance
CN110490966A (zh) * 2018-05-14 2019-11-22 重庆工港致慧增材制造技术研究院有限公司 生物3d打印正向支架模型设计方法
CN109528330A (zh) * 2018-12-20 2019-03-29 成都登特牙科技术开发有限公司 可摘局部义齿支架的数字化设计方法
CN110251276A (zh) * 2019-06-03 2019-09-20 浙江工业大学 一种增减材复合加工成型的口腔赝复体支架制作方法
CN111513881A (zh) * 2020-05-09 2020-08-11 北京大学口腔医学院 一种上颌骨缺损赝复体的制作方法及系统
CN111529105A (zh) * 2020-05-19 2020-08-14 北京联袂义齿技术有限公司 一种全口义齿立体支架的数字化制造方法
CN111941828A (zh) * 2020-08-12 2020-11-17 北京大学口腔医学院 咬合调整器的数字化实现方法、装置、设备及存储介质
CN112215065A (zh) * 2020-09-04 2021-01-12 北京大学口腔医学院 一种牙颌边界特征自动化识别方法
CN112895459A (zh) * 2021-01-14 2021-06-04 南京前知智能科技有限公司 一种基于设计组件信息的3d打印模型智能预处理方法及装置
CN113171188A (zh) * 2021-01-29 2021-07-27 正雅齿科科技(上海)有限公司 一种具有硬腭区域的数字化牙颌模型构建方法及系统
CN113397743A (zh) * 2021-06-04 2021-09-17 珠海市三通陶齿有限公司 一种牙支持式可摘局部义齿制作方法
CN113288480A (zh) * 2021-06-25 2021-08-24 成都登特牙科技术开发有限公司 骨增量与种植义齿联合设计方法及骨增量模型的制造方法
CN113397742A (zh) * 2021-07-29 2021-09-17 维视医疗信息科技(山东)有限公司 可摘局部义齿支架模型自动生成方法及系统
CN114714626A (zh) * 2022-04-02 2022-07-08 北京大学口腔医学院 一种义齿3d打印支架的设计方法、装置及可存储介质

Also Published As

Publication number Publication date
CN114714626A (zh) 2022-07-08
CN114714626B (zh) 2023-08-15

Similar Documents

Publication Publication Date Title
US11642195B2 (en) Visual presentation of gingival line generated based on 3D tooth model
US11727607B2 (en) Generation of images with tooth color determined using depth information
US11317999B2 (en) Augmented reality enhancements for dental practitioners
US10945813B2 (en) Providing a simulated outcome of dental treatment on a patient
CN105447908B (zh) 基于口腔扫描数据和cbct数据的牙列模型生成方法
US11517272B2 (en) Simulated orthodontic treatment via augmented visualization in real-time
CN110200710A (zh) 一种基于三维成像与实时建模的口腔修复方法
CN114746952A (zh) 用于从二维草图创建三维牙科修复体的方法、系统和计算机可读存储介质
CN111784754B (zh) 基于计算机视觉的牙齿正畸方法、装置、设备及存储介质
TWI728374B (zh) 牙齒虛擬編輯方法、系統、電腦設備和儲存媒體
KR20140015239A (ko) 2ｄ 영상 장치
US20230206451A1 (en) Method for automatic segmentation of a dental arch
WO2021218724A1 (zh) 一种用于口腔数字印模仪的数字模型智能设计方法
WO2023185405A1 (zh) 义齿3d打印支架的设计方法、装置及可存储介质
US20240024076A1 (en) Combined face scanning and intraoral scanning
CN108992193A (zh) 一种牙齿修复辅助设计方法
US20230113425A1 (en) Oral Digital Impression Instrument, Digital Model for the Same, and Intelligent Method of Making and/or Using the Same
CN116749522A (zh) 一种用于正畸矫正工具的3d打印系统和方法
CN116211521A (zh) 获取缺失牙咬合关系的数字扫描方法、装置、设备及介质
CN117788747A (zh) 一种口扫数据与cbct融合的牙冠三维重建方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23777783

Country of ref document: EP

Kind code of ref document: A1