Method for flattening a digital 3D mesh representing a dental object

Abstract

This disclosure relates to a computer-implemented method for flattening a digital 3D mesh representing a dental object. The method includes obtaining a digital 3D mesh. The method includes determining a circumferential mesh boundary in the digital 3D mesh, wherein the mesh boundary comprises a plurality of boundary vertices surrounding a plurality of interior vertices. The method includes determining a target shape having a circumferential target boundary, wherein the target shape is planar. The method includes, for each boundary vertex, performing a consistent mapping to an anchor point on the target boundary corresponding to the respective vertex. The method includes, for each boundary vertex, translating the respective boundary vertex to its corresponding anchor point on the target boundary. The method includes flattening the interior vertices to the target shape based on the anchor points to generate a flattened mesh.

Description

Technical Field

[0001] This disclosure relates to a method for flattening a digital 3D mesh representing a dental object. The disclosure also relates to a dental scanning system. Background Technology

[0002] Digital dentistry has revolutionized dental care by integrating advanced digital technologies into traditional dental procedures. This innovation encompasses a wide range of tools and systems, including intraoral scanners, computer-aided design and manufacturing (CAD / CAM) systems, 3D printing, and digital imaging technologies. These advancements introduce significant advantages over traditional methods, improving the accuracy, efficiency, and overall quality of dental treatment.

[0003] A significant benefit is the improved patient experience. The digital workflow minimizes the discomfort associated with traditional methods, such as taking impressions of dental objects using alginate materials. Instead, intraoral scanners rapidly and non-invasively capture a digital three-dimensional (3D) representation of the dental object. This increases patient satisfaction and reduces anxiety during dental visits.

[0004] Digital 3D meshes typically take the form of 3D meshes composed of multiple facets (such as triangles) that represent the surface of a dental object. Using digital 3D meshes can significantly reduce treatment time. For example, CAD / CAM technology allows for the design and fabrication of in-clinic restorations (such as crowns, veneers, and bridges) during a single appointment, eliminating the need for temporary restorations and multiple appointments. Similarly, 3D printing technology accelerates the production of dental models, surgical guides, and restorations, streamlining clinical workflows. From a clinical perspective, the integration of digital technologies enhances treatment outcomes. Precise data obtained through digital tools enables highly customized and predictable treatment plans tailored to each patient's unique anatomy. This is particularly advantageous in complex procedures such as implant placement and orthodontic treatment, where accuracy is paramount.

[0005] In prosthodontics, dental restorations (such as crowns) are attached to existing teeth by reshaping them to match the inner contours of the restoration. This allows the restoration to bond to the reshaped tooth. The reshaped tooth is also called the restoration / preparation site and includes the marginal line, which is the precise boundary or edge where the prepared tooth structure contacts the restorative material (such as a crown, veneer, or filling). This is a crucial interface that affects the function and aesthetics of the restoration. Therefore, correctly determining the marginal line in digital data is essential for designing a suitable dental restoration based on that data, as errors in the determined marginal line can lead to a mismatch between the physical marginal line and the interface. However, correctly determining the marginal line from 3D data can be difficult because, from certain angles, the marginal line may be obscured by the rest of the prepared tooth.

[0006] Therefore, while processing 3D data offers many benefits, it can also present challenges, as the complexity of handling 3D information can be difficult for both computers and human users. Such challenges can be mitigated through mesh flattening, which allows a 3D mesh to be flattened into a two-dimensional (2D) shape. Mesh flattening is disclosed in the applicant's own US11321918B2, which discloses a method for manipulating 3D objects by flattening a 3D mesh (more specifically, a 3D mesh representing a dental crown). While this method works well for dental objects (such as dental crowns with continuous circumferential boundaries that do not bend on their own—that is, the radial angle only increases as you move from one boundary vertex to the next until a full circle around the boundary is completed), it is not always suitable for preparation sites (such as edge lines) because it assumes that the boundary orientation is monotonic to maintain order-preserving properties, which is not guaranteed for the boundaries of preparation sites.

[0007] Therefore, there is a need for an improved method for manipulating (e.g., flattening) 3D meshes representing dental objects. Summary of the Invention

[0008] A first aspect of this disclosure relates to a computer-implemented method for flattening a digital 3D mesh representing a dental object. The method includes obtaining a digital 3D mesh. The method includes determining circumferential mesh boundaries in the digital 3D mesh, wherein the mesh boundaries comprise a plurality of boundary vertices surrounding a plurality of interior vertices. The method includes determining a target shape having a circumferential target boundary, wherein the target shape is planar. The method includes, for each boundary vertex, performing a consistent mapping to an anchor point on the target boundary corresponding to the respective vertex. The method includes, for each boundary vertex, translating the respective boundary vertex to its corresponding anchor point on the target boundary. The method includes flattening the interior vertices to the target shape based on the anchor points to generate a flattened mesh.

[0009] A further aspect of this disclosure relates to a computer program product comprising instructions that, when executed by a computer, cause the computer to perform the method described in any of the first aspects.

[0010] A further aspect of this disclosure relates to a non-transient computer-readable medium comprising instructions that, when executed by a computer, cause the computer to perform the method described in any of the first aspects.

[0011] A further aspect of this disclosure relates to a dental system including a processing device configured to perform the method of the first aspect.

[0012] Given an open and connected polygonal mesh, this algorithm aims to map its ordered set of boundary vertices to a regular shape. In the case of multiple boundaries, all boundary vertices must belong to the same boundary. Here, "ordered" refers to the sequence of vertices navigating around the mesh boundary in a common orientation along the boundary edges.

[0013] Once the boundaries are fixed, the remaining vertices and edges are mapped to planar objects, thus flattening the initial 3D mesh into a planar mesh. Various embodiments may use different methods for this mapping. Initially, a matrix of connectivity between each vertex and other vertices can be created, and this matrix system can be solved to map the coordinates of each vertex to a plane (Tutte, William Thomas. “How to draw a graph.” Proceedings of the London Mathematical Society 3.1 (1963): 743-767). This connectivity can be weighted in different ways, resulting in different mappings. In various embodiments, these methods include, but are not limited to: uniform weighting (Tutte 1963), angle-based weighting (Floater, Michael S. and Ming-Jun Lai. “Polygonal spline spaces and the numerical solution of the Poisson equation.” SIAM Journal on Numerical Analysis 54.2 (2016):797-824), and cotangent-based weighting (Meyer, Mark et al., “Discrete differential-geometry operators for triangulated 2-manifolds.” Visualization and mathematics III. Springer, Berlin, Heidelberg, 2003. 35-57).

[0014] For an example of flattening the initial 3D scan into a planar mesh, see below.

[0015] The advantage of flattening the initial 3D mesh in this way is that the initial 3D mesh and the planar mesh are bijective, which means that each vertex can be mapped back and forth between the two meshes (i.e., the 3D mesh and the flattened mesh). Attached Figure Description

[0016] The above and other features and advantages of the present invention will readily become apparent to those skilled in the art from the following detailed description of exemplary embodiments of the invention with reference to the accompanying drawings, wherein:

[0017] Figure 1a An intraoral scanner used in the scanning system or method of the present invention is shown;

[0018] Figure 1b An intraoral scanner used in the scanning system or method of the present invention is shown;

[0019] Figure 2 A laboratory scanner used in the scanning system or method of the present invention is shown;

[0020] Figure 3 A block diagram of the scanning system of the present invention is shown;

[0021] Figure 4 A block diagram of the scanning system of the present invention is shown;

[0022] Figure 5 Point cloud data is shown;

[0023] Figure 6 A 3D representation including a 3D mesh is shown;

[0024] Figure 7a An example of a grid is shown;

[0025] Figure 7b An example of a grid is shown;

[0026] Figure 7c An example of a grid is shown;

[0027] Figure 7d An example of a grid is shown;

[0028] Figure 7e An example of a grid is shown;

[0029] Figure 8a A 3D mesh is shown;

[0030] Figure 8b The enlarged portion of the 3D mesh is shown;

[0031] Figure 8c The enlarged portion of the 3D mesh is shown;

[0032] Figure 9 A 3D mesh is shown;

[0033] Figure 10 The initial 3D mesh is shown;

[0034] Figure 11a The diagram illustrates the mapping rules of the existing technology;

[0035] Figure 11b The diagram illustrates the mapping rules of existing technologies that lead to invalid mappings;

[0036] Figure 12 A 3D mesh representing a dental object with edge lines is shown;

[0037] Figure 13 A flowchart of the method of the present invention is shown;

[0038] Figure 14 A flowchart of the method of the present invention is shown;

[0039] Figure 15a A 3D mesh with the identified regions of interest is shown;

[0040] Figure 15b A sub-section of a 3D mesh with an identified region of interest surrounded by a circumferential mesh boundary is shown;

[0041] Figure 15c This shows a sub-section of the flattened grid that has been flattened to the region of interest;

[0042] Figure 16a The sub-parts of the identified region of interest are shown;

[0043] Figure 16b This shows a sub-section of the flattened grid that has been flattened to the region of interest;

[0044] Figure 16c The edge lines in the flattened grid are shown;

[0045] Figure 17 It shows the graphical user interface and underlying image processing; and

[0046] Figure 18 It shows the graphical user interface and underlying image processing. Detailed Implementation

[0047] The detailed description below, taken in conjunction with the accompanying drawings, is intended to describe various configurations. The detailed description includes specific details intended to provide a thorough understanding of the various concepts. However, it will be apparent to those skilled in the art that these concepts can be practiced without these specific details. Several aspects of devices, systems, media, programs, and methods are described by various blocks, functional units, modules, components, circuits, steps, processes, algorithms, etc. (collectively, “elements”). Depending on the specific application, design constraints, or other reasons, these elements may be implemented using electronic hardware, computer programs, or any combination thereof.

[0048] A computer-implemented method for manipulating at least a portion of a digital 3D mesh representing a dental object is disclosed, the method comprising:

[0049] • Obtain a digital 3D mesh;

[0050] • Determine the circumferential mesh boundary in the digital 3D mesh, wherein the mesh boundary includes multiple boundary vertices and one or more internal vertices surrounded by the boundary vertices;

[0051] • Determine the shape of the target having a circumferential target boundary, wherein the target boundary is planar; and

[0052] • For each vertex on the boundary, perform a consistent mapping to the anchor point corresponding to the vertex on the target boundary.

[0053] By performing coherent mapping, the shortcomings of existing techniques are mitigated, namely: the result is rarely flat because the anchor points are not located on the same plane, and the flattened shape may have arbitrary boundary shapes because the boundary shape is maintained during flattening.

[0054] Choosing a planar target shape ensures that the result (flattened mesh) is flat, even if we use an inconsistent mapping.

[0055] Consistent mapping is used to ensure that the result does not lose information; for example, it does not create self-intersections that are not present in the unflattened mesh. In the disk example, if two adjacent boundary points are mapped (inconsistently) to opposite ends of the unit circle, this will create a patch across the center of the circle in the flattened mesh, which may obscure some internal patches and effectively reduce information.

[0056] In the prior art, the previous method of flattening a 3D mesh (such as a preparatory part) was to set anchor points by fixing the boundary vertices in place, thus taking into account any curvature of the preparatory part boundary, which is rarely flat, making it difficult to achieve a flat result.

[0057] This method can be used to flatten a 3D mesh into a flattened 2D mesh. It can also be used to identify the edge lines of dental objects within a 3D mesh. The use of consistency mapping has several applications. Firstly, it is applied when flattening a 3D mesh because it ensures the correct mapping from boundary vertices to anchor points, thus preventing self-intersections and unintended folding of the 3D mesh during the flattening step following consistency mapping. Flattening can further be a useful tool for identifying which vertices of the 3D mesh represent the edge lines of a dental object.

[0058] In 3D, edges can be difficult to identify for both humans and computer algorithms because, from most perspectives, they are at least partially occluded, i.e., hidden. This, in turn, makes it difficult to provide training material for training machine learning models to identify edges at the vertex level in 3D. This invention does not use a 3D mesh to identify edges, but instead leverages the technique of flattening a 3D mesh into a flattened mesh (i.e., a 2D mesh), thus making the entire edge visible, i.e., unoccluded. This works because of consistent mapping, as the 3D mesh must be correctly flattened to ensure a usable flattened mesh is obtained. After identifying edges in the flattened mesh, the flattening can be reversed to bring the flattened mesh back into 3D, or the vertices in the 3D mesh representing edges can be identified based on the corresponding vertices in the flattened mesh that are identified as edges.

[0059] In one embodiment of the invention, the consistency mapping satisfies the order-preservation criterion, such that the order of the anchor point around the target boundary is the same as the order of its corresponding boundary vertex around the mesh boundary, and / or, wherein the consistency mapping satisfies the uniqueness criterion, such that each unique boundary vertex maps to a unique anchor point.

[0060] In one embodiment of the present invention, the consistent mapping includes: selecting initial vertices of the boundary vertices and initial anchor points on the target boundary, and mapping the initial vertices to the initial anchor points. The selection of initial vertices and initial anchor points can be arbitrary. The important step is that subsequent boundary vertices must be consistently mapped to the corresponding anchor points on the target boundary, that is, the order of anchor points around the target boundary is the same as the order of boundary vertices around the circumferential mesh boundary.

[0061] In one embodiment of the present invention, the consistency mapping includes: for each vertex in the boundary vertices other than the initial vertex, determining the relative distance between the corresponding vertex and the initial vertex along the mesh boundary, and for each vertex in the boundary vertices other than the initial vertex, mapping the corresponding vertex to an anchor point whose relative distance from the initial anchor point along the target boundary is the same as the relative distance between the corresponding vertex and the initial vertex.

[0062] In one embodiment of the invention, the method further includes: for each boundary vertex, translating the corresponding boundary vertex to its corresponding anchor point on the target boundary. Once the anchor point is known according to the consistency mapping, the boundary vertices can be moved / translated to the target boundary of the target shape.

[0063] In one embodiment of the invention, the target shape is planar. In one embodiment, the target shape is circular. For applications involving flattening, the target shape will always be planar. However, other applications may also utilize the disclosed consistency mapping, for which the target shape may have other shapes, such as a cone.

[0064] In one embodiment of the invention, the method further includes flattening the internal vertices to a target shape based on anchor points to generate a flattened mesh. Known flattening methods are more likely to achieve correct flattening after the boundary vertices are consistently mapped to the target shape.

[0065] In one embodiment of the invention, determining the circumferential mesh boundary in a digital 3D mesh includes: coarsely identifying the edge lines of a dental object in the digital 3D mesh, and setting a circumferential mesh boundary around the coarsely identified edge lines. To flatten the 3D mesh, portions of the 3D mesh that constitute the edge lines (i.e., portions representing the edge lines of the dental object) are separated. For this purpose, the edge lines in the 3D mesh can be coarsely identified, i.e., not at the vertex level, but by separating portions of the 3D mesh that can be reliably determined to include the entire edge line. This can be done using a machine learning model trained to identify edge lines in 3D (also referred to as a first trained model). Alternatively, coarse edge line identification can be performed by identifying portions of the 3D mesh 600 representing the preparation site 632. Since the edge line 630 is part of the preparation site 632, it can be ensured that it will form part of the region of interest 642 and will fall entirely within the circumferential mesh boundary 640 set around the preparation site 632. Therefore, the first model can be trained to identify the preparatory part 632 in the 3D mesh 600, thereby roughly identifying the edge line 630, that is, identifying the region of interest 642 within which the entire edge line 630 is located.

[0066] In one embodiment of the invention, a buffer distance is provided between the circumferential mesh boundary 640 and the coarsely identified edge line, for example, by adding a buffer region around the preparatory portion 632 identified by the first model. To ensure that the entire edge line 630 is included by the determined circumferential mesh boundary 640, the buffer distance can be added to the coarsely identified edge line so that, for example, errors that may occur when the first trained model identifies the edge line 630 will not cause a portion of the edge line 630 to fall outside the circumferential mesh boundary 640 and thus be omitted from the flattened mesh 652.

[0067] It is worth noting that a sub-part of the 3D mesh defined by the circumferential mesh boundary 640 can be extracted from the 3D mesh for use in the following steps. However, each subsequent method step can be backtracked so that vertices or surface points (representing regions of interest, such as edge lines 630) later identified in the processed mesh (e.g., flattened mesh 652) can be used to identify their corresponding vertices or surface points in the 3D mesh, thereby identifying / determining the vertices in the 3D mesh that represent regions of interest (e.g., edge lines 630).

[0068] In one embodiment of the invention, the method further includes identifying edge lines 630 in the flattened mesh. In one embodiment of the invention, identifying edge lines in the flattened mesh includes determining vertices representing edge lines in the flattened mesh. Edge line identification can be performed at the vertex level after flattening a sub-part of the 3D mesh including the edge lines (i.e., the sub-part defined by the circumferential mesh boundary), because flattening ensures that edges are not occluded. Edge lines in the flattened mesh can be identified using a second trained model that is trained to identify edge lines in 2D. The success of identifying edge lines in the flattened mesh depends on the reliability of the mesh flattening. Therefore, the method of the present invention ensures that the mesh is correctly flattened through consistent mapping, thereby reducing the failure rate of edge line identification.

[0069] In one embodiment of the present invention, the method further includes: identifying vertices in the digital 3D mesh that correspond to vertices in the flattened mesh that are determined to represent edge lines. In another embodiment of the present invention, the method further includes: determining edge lines in the digital 3D mesh based on the identified vertices in the digital 3D mesh that correspond to the vertices of the determined flattened mesh.

[0070] In one embodiment, the steps to obtain a digital 3D mesh include:

[0071] o Obtain one or more 2D images of a dental object, for example, using an intraoral scanner or a laboratory scanner;

[0072] o Generates point cloud data based on one or more 2D images, wherein the point cloud data includes multiple points representing surface points of a dental object; and

[0073] o Generates a digital 3D mesh based on point cloud data, which includes multiple polygons representing the surface of a dental object.

[0074] In one embodiment, generating a digital 3D mesh includes: obtaining geometric information based on one or more 2D images, wherein the geometric information indicates the rate of change of one or more surface properties in multiple surface portions of a dental object.

[0075] In one embodiment, generating a digital 3D mesh includes determining the mesh resolution for each surface portion based on geometric information. In another embodiment, generating a digital 3D mesh includes generating a mesh for each surface portion, the resolution of which is determined based on the respective surface portion.

[0076] By segmenting the 3D representation into sub-parts representing corresponding surface portions of a dental object, a mesh can be generated at a resolution determined to be suitable for representing those surface portions. Therefore, the mesh accurately represents the object's surface without requiring excessive processing power from the scanning system. Consequently, meshes can be generated faster, saving time for both users and patients and improving the user experience.

[0077] In one embodiment, at least a portion of the geometric information associated with a corresponding surface portion is based on the point distribution of point cloud data corresponding to the surface portion. In one embodiment, at least a portion of the geometric information associated with a corresponding surface portion is based on ray tracing of light reflected at points in the point cloud data corresponding to the surface portion. In one embodiment, the geometric information includes information indicating the rate of change of the surface portion's contour. In one embodiment, at least a portion of the geometric information associated with a corresponding surface portion is based on color data corresponding to the surface portion. In one embodiment, the geometric information includes information indicating the rate of change of the surface portion's color.

[0078] Based on geometric information about color data from 2D images, it will be possible to provide geometric data indicating changes in the surface color of an object. This can then be used to determine the grid resolution required to accurately represent the color changes of individual surface portions of the object.

[0079] When generating a 3D mesh representing a preparatory area with edge lines, using geometric information that includes color and / or contour data is particularly useful because edge lines are defined by sharp edges (i.e., a high rate of change in the contour) and are also often defined by a high degree of color variation, as a portion of the edge line includes a portion of the original surface of the tooth (which has been exposed to the patient's oral environment and thus undergone color change) and a portion of newly exposed tooth material (due to the preparation process of grinding a portion of the tooth). Therefore, using geometric information to generate a 3D mesh means that a higher mesh resolution can be selected for the surface portion representing the edge lines, thus ensuring more accurate resolution of the edge lines.

[0080] In one embodiment, obtaining one or more 2D images of an object includes obtaining multiple 2D images of the object. In one embodiment, generating point cloud data includes: resolving a correspondence problem to determine which portions of a first image among the multiple 2D images correspond to which portions of a second image among the multiple 2D images. In one embodiment, generating point cloud data includes: registering points from a plurality of points obtained from each corresponding 2D image to a common coordinate system.

[0081] In one embodiment, generating a digital 3D mesh includes assigning a color value to each polygon based on one or more 2D images. This has the advantage that the 3D representation can represent not only the shape of an object but also its color. This allows for more detailed diagnosis when a user examines the 3D representation and also allows restorations made based on the 3D representation to match the color of adjacent teeth.

[0082] In one embodiment, obtaining one or more 2D images includes projecting light onto a dental object. In one embodiment, the light is projected in the form of a pattern. In one embodiment, the projected pattern is static over time. In one embodiment, the projected pattern changes over time. By projecting light in a pattern, the correspondence problem may be more easily solved when using multiple 2D images taken at different times and / or from different perspectives; that is, the pattern features can be used to identify which parts of one 2D image correspond to which parts of another 2D image.

[0083] In one embodiment, obtaining one or more 2D images of a dental object is done using an intraoral scanner. The advantage of an intraoral scanner is that it allows users (e.g., dentists or dental technicians) to quickly and accurately create digital 3D representations of a user's teeth without the need for physical impressions that can be uncomfortable for the patient.

[0084] In one embodiment, obtaining one or more 2D images of a dental object is performed using a laboratory scanner. Therefore, in situations where an intraoral scanner is not available, using a laboratory scanner to obtain 2D images of a patient's dental impressions may be a cost-effective method for generating digital 3D representations.

[0085] In one embodiment, generating point cloud data includes performing a first transformation to generate a coordinate dataset of at least a portion of an object in a scanning device coordinate system. In another embodiment, generating point cloud data includes performing a second transformation on the coordinate dataset of at least a portion of the object in the scanning device coordinate system to generate at least a portion of point cloud data in a real-world coordinate system.

[0086] In one embodiment, generating a digital 3D mesh further includes: generating an initial mesh based on point cloud data, wherein the initial mesh includes a plurality of initial polygons. In one embodiment, generating a digital 3D mesh is based on the initial mesh.

[0087] In one embodiment, obtaining geometric information about a surface portion includes comparing the normal vectors of adjacent initial polygons within the surface portion to obtain information about the rate of change of the surface portion's profile.

[0088] In one embodiment, the method further includes generating 3D dental restoration data representing a dental restoration configured for placement on a tooth with edge lines, based on one or both of a digital 3D mesh and vertices in the flattened mesh identified as representing edge lines or vertices in the digital 3D mesh identified as corresponding to edge lines. Using the mesh flattening method disclosed herein, after identifying edge lines in a dental object, a dental restoration can be designed and subsequently manufactured using the 3D mesh and 3D dental restoration data, making it less likely to suffer from interface defects due to inaccurate fit with the edge lines. Therefore, this method can improve the success rate of dental restoration treatment and reduce the likelihood of patients needing follow-up treatment or redoing. Finally, the improved edge line / restoration interface can increase the lifespan of the restoration, meaning the restoration can last longer.

[0089] The above-described embodiments and advantages can also be applied to the dental scanning system of the present invention.

[0090] Dental objects

[0091] Although referred to in the singular, a dental object can also be a set of dental features in the oral cavity of a subject (i.e., a patient), such as a set of teeth and the surrounding gingiva. Examples of dental objects include one or more of the following: a tooth / teeth, gingiva, implants (one or more), dental restorations (one or more), dental prostheses, alveolar ridges (one or more), and / or combinations thereof. Alternatively, a dental object can be a plaster or plastic model representing the subject's teeth. As an example, a dental object may include the subject's teeth and / or gingiva. A dental object may only be a part of the subject's teeth and / or oral cavity, as it is not necessary to scan the entire set of teeth of the subject during a scanning procedure. A scanning procedure in this document can be understood as a period of time during which data (such as 2D images) of a dental object are acquired / obtained.

[0092] Scanning equipment

[0093] The scanning device disclosed herein can be an intraoral scanning device for acquiring images of the oral cavity of a subject. The intraoral scanning device is a handheld scanning device, i.e., a device configured to be held in a human hand. Alternatively, the scanning device can be a fixed dental scanner (also known as a laboratory scanner) for acquiring images of plaster or plastic models representing the teeth of a subject. The scanning device can employ any suitable scanning principle, such as triangulation-based scanning, stereo vision, structure from motion, confocal scanning, or other scanning principles.

[0094] In some embodiments, the scanning device employs a triangulation-based scanning principle. As an example, a projector unit and one or more camera units can be used to determine points in 3D space based on triangulation. In other embodiments, the scanning device employs a focus-based scanning principle.

[0095] Projector unit

[0096] The projector unit described herein can be understood as a device configured to project light onto a surface, such as the surface of a dental object. In a preferred embodiment, the projector unit is configured to project a light pattern onto the surface of the dental object. The projector unit can be configured to project a light pattern such that the light pattern is focused at a predetermined focal length measured along the projector's optical axis.

[0097] Each projector unit may include one or more light sources. The projector unit may be configured to project a light pattern defined by multiple projector rays when the light sources(s) ...

[0098] In some embodiments, each projector unit includes a light source for generating white light. Alternatively, the projector unit may include multiple light sources (such as LEDs) that individually generate light of different wavelengths (such as red, green, and blue), which can be combined to form light comprising different wavelengths. Thus, the light generated by one or more light sources can be defined by wavelengths that define a particular color, or by different wavelength ranges that define combinations of colors (such as white light). In some embodiments, the light source is a diode, such as a white light diode or a laser diode.

[0099] In some embodiments, the scanning device includes a light source configured to excite fluorescent material to obtain fluorescence data from a dental object, such as a tooth. This light source may be configured to produce a narrow range of wavelengths. In other embodiments, the scanning device includes an infrared light source configured to generate wavelengths in the infrared range, such as between 700 nm and 1.5 µm. In some embodiments, the scanning device includes one or more light sources selected from the group consisting of infrared (IR) light sources, near-infrared (NIR) light sources, blue light sources, violet light sources, ultraviolet (UV) light sources, and / or combinations thereof. In some embodiments, the scanning device includes a first light source forming part of a projector unit, and one or more second light sources (e.g., one or more infrared LEDs and / or one or more ultraviolet LEDs) located at a distal portion of the scanning device, such as the tip of the scanning device.

[0100] The projector unit may include a digital light processing (DLP) projector, or a diffractive optical element (DOF), or a front-illuminated reflective mask projector, or a micro-LED projector, a liquid crystal on silicon (LCoS) projector, or a back-illuminated mask projector, wherein the light source is placed behind a mask having a spatial pattern, thereby patterning the light projected onto the surface of the dental object. This pattern may be dynamic, i.e., the pattern changes over time; or the pattern may be time-static, i.e., the pattern remains unchanged over time. The projector unit may include one or more collimating lenses for collimating the light from the light source. One or more collimating lenses may be positioned between the light source and the mask. The projector unit may further include one or more focusing lenses or lens elements configured to focus the light at a predetermined focal length.

[0101] In some embodiments, the projector unit of the scanning device includes at least one light source and a pattern generating element for defining a light pattern. The pattern generating element is preferably configured to generate a light pattern to be projected onto the surface of a dental object. As an example, the pattern generating element may be a mask with a spatial pattern. Therefore, the projector unit may include a mask configured to define a light pattern. The mask may be placed between the light source of the projector unit and one or more focusing lenses such that light passing through the mask is patterned into a light pattern. As an example, the mask may define a polygonal pattern comprising multiple polygons, such as a checkerboard pattern. The projector unit may further include one or more lenses, such as collimating lenses or projection lenses. In other embodiments, the pattern generating element generates the light pattern based on diffraction and / or refraction, such as a pattern comprising a discrete, unconnected array of dots.

[0102] The pattern generating element can be a mask, such as a transparent mask or a transmissive mask with a spatial pattern. In other embodiments, the pattern generating element is configured to generate light patterns using diffraction and / or refraction.

[0103] A spatial pattern can be a polygonal pattern comprising multiple polygons. Polygons can be selected from the group consisting of triangles, rectangles, squares, pentagons, hexagons, and / or combinations thereof. Other polygons are also conceivable. Generally, polygons consist of sides and angles. In some embodiments, polygons are repeated in a predefined manner within the pattern. As an example, the pattern may include multiple repeating units, each comprising a predetermined number of polygons, and the repeating units are repeated throughout the pattern. Alternatively, the pattern may include a predefined arrangement comprising stripes, squares, dots, triangles, rectangles, and / or combinations thereof. In some embodiments, the pattern is non-coded, and therefore no part of the pattern is unique.

[0104] In some embodiments, the generated light pattern is a polygonal pattern, such as a checkerboard pattern comprising multiple squares. Similar to a common checkerboard pattern, the squares in the pattern may have alternating areas of light and dark, corresponding to areas of low (dark) light intensity and areas of high (higher) light intensity (bright). In some embodiments, the light pattern is a checkerboard pattern comprising alternating light and dark squares. In some embodiments, the light pattern comprises a discrete, unconnected distribution of light spots.

[0105] The pattern preferably includes multiple pattern features. The pattern can be a high-density pattern, understood as including more than 3000 pattern features. However, the currently disclosed systems and methods are not limited to high-density patterns, as they are equally effective for low-density patterns. In some embodiments, the pattern includes at least 1000 pattern features, or at least 3000 pattern features, or at least 10000 pattern features. When a pattern including pattern features is projected onto the surface of a 3D object, the resulting object image will similarly include multiple image features corresponding to the pattern features. A pattern feature / image feature can be understood as a single, well-defined location within a pattern / image. Examples of image features / pattern features include corners, edges, vertices, points, transition points, dots, stripes, etc. In some embodiments, image features / pattern features include the corners of squares in a checkerboard pattern. In other embodiments, image features / pattern features include corners in polygonal patterns (such as triangular patterns).

[0106] The projector unit can be configured to generate a predefined static pattern that can be projected onto the surface of an object. The advantage of using a static pattern is that it allows all image data to be captured simultaneously, thus preventing distortion caused by motion. Another advantage is that static patterns allow each camera unit to acquire only one image from a set of images, thereby reducing power consumption, such as that of the light source.

[0107] Alternatively, the projector unit can be configured to generate dynamic patterns that vary over time. The projector unit can be associated with its own projector plane, which is defined by the projector optics. As an example, if the projector unit is a back-illuminated mask projector, the projector plane can be understood as a plane containing the mask. The projector plane includes multiple pattern features of the projected pattern. Preferably, the camera unit and the projector unit are arranged such that the image sensor and the projector plane (e.g., the plane defined by the mask) are in the same plane.

[0108] A projector unit can define a projector optical axis. An optical axis can be understood as a line in an optical system (such as a camera lens or projector unit) along which it has a degree of rotational symmetry. In some embodiments, the projector optical axis of the projector unit is substantially parallel to the longitudinal axis of the scanning device. In other embodiments, the projector optical axis of the scanning unit defines an angle with the longitudinal axis of the scanning device, such as at least 45° or at least 75°. In still other embodiments, the projector optical axis of the projector unit is substantially orthogonal to the longitudinal axis of the scanning device.

[0109] Camera unit

[0110] The camera unit in this document can be understood as a device for capturing 2D images of a dental object. Each camera unit may include an image sensor for generating an image based on incident light (e.g., incident light received from an illuminated dental object). As an example, the image sensor may be an electronic image sensor, such as a charge-coupled device (CCD) or an active pixel sensor (CMOS sensor). In some embodiments, the image sensor is a global shutter sensor configured to simultaneously expose the entire image area (all pixels) and generate an image at a single point in time. The image sensor may have an image frame rate of at least 30 frames per second, such as at least 60 frames per second, or even at least 90 frames per second.

[0111] One or more image sensors may include a pixel array, where each pixel is associated with a corresponding camera ray. Similarly, each image in an image set may consist of a pixel array, where each pixel includes a pixel color defined by one or more color channels. The pixel array may be a two-dimensional (2D) array. In some embodiments, the image sensor is a CMOS sensor, including an analog-to-digital converter (ADC) for each column of pixels, thereby significantly accelerating the conversion speed and enabling each camera unit to benefit from the higher speed. Each image sensor may define an image plane, which can be understood as a plane containing a projected image of an object. Each image acquired by one or more image sensors may include multiple image features, where each image feature is derived from pattern features of the projected pattern. In some embodiments, one or more camera units include a light field camera. Preferably, each camera unit defines a camera optical axis. The camera unit may also include one or more focusing lenses for focusing light.

[0112] In some embodiments, the image sensor is a monochrome image sensor, where each pixel is associated with a single color channel (e.g., a grayscale color channel), and the value of each pixel represents only the amount of light. In other embodiments, the image sensor is a color image sensor, or an image sensor that includes an array of color filters on a pixel array. As an example, the color filter array may be a Bayer filter, using an arrangement of four color filters: red (R), green (G), blue (B). Bayer filters may also be referred to as RGGB filters. When utilizing image sensor data, color pixels can be combined into 2x2 monochrome pixels for 3D depth reconstruction. In this case, the resolution of the 3D depth reconstruction is only half the resolution of the image sensor in each direction. When obtaining a textured (color) image, it is preferable to utilize the full original resolution (pixels with color filters).

[0113] According to some embodiments, the projector optical axis and camera optical axis of at least one camera unit define a camera-projector angle of approximately 5 to 15 degrees, preferably 5 to 10 degrees, and even more preferably 8 to 10 degrees. In some embodiments, the camera unit is out of focus at the probe opening of the scanning device and / or on the optical window surface of the probe. In some embodiments of the scanning device, the camera unit and projector unit of a given scanning unit are focused at the same distance. In some embodiments, the field of view of each camera unit is 50 to 115 degrees, such as 65 to 100 degrees or 80 to 90 degrees.

[0114] Processing equipment

[0115] According to some embodiments, the scanning system includes a processing device having one or more processors selected from the group consisting of: a central processing unit (CPU), an accelerator (offload engine), a general-purpose microprocessor, a graphics processing unit (GPU), a neural processing unit (NPU), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a dedicated logic circuit, a dedicated artificial intelligence processor unit, or a combination thereof. The processor(s) may be located on the scanning device. Alternatively, the processor(s) or a subset thereof may be located on a computer system and / or a remote computing device, as described herein.

[0116] The scanning system may also include computer memory for storing instructions that, when executed, cause one or more processors to perform steps of determining image features in an image set. The computer memory may also store instructions that, when executed, cause one or more processors to perform steps of generating a digital representation of a three-dimensional (3D) object. Generally, one or more processors may be configured to perform any computer-implemented method disclosed herein, whether in whole or in part; for example, when executed, some processors perform some method steps while other processors perform other method steps.

[0117] Computer System

[0118] A computer system can be understood as an electronic processing device used to perform a series of arithmetic or logical operations. In the context of this article, a computer system refers to one or more devices, including at least one processor (such as a central processing unit (CPU)) and some type of computer memory. Examples of computer systems falling under this definition include desktop computers, portable computers, computer clusters, servers, cloud computers, quantum computers, mobile devices (such as smartphones and tablets), and / or combinations thereof.

[0119] A computer system may include hardware such as one or more central processing units (CPUs), graphics processing units (GPUs), and computer memory such as random access memory (RAM) or read-only memory (ROM). The computer system may include a CPU configured to read and execute instructions stored in the computer memory (e.g., in the form of random access memory). The computer memory is configured to store instructions for execution by the CPU and data used by those instructions. As an example, the memory may store instructions that, when executed by the CPU, cause the computer system to perform, in whole or in part, any computer-implemented methods disclosed herein. The computer system may also include a graphics processing unit (GPU). The GPU may be configured to perform various tasks such as video decoding and encoding, rendering of digital representations, and other image processing tasks.

[0120] The computer system may also include non-volatile storage in the form of hard disk drives. Preferably, the computer system also includes I / O interfaces configured to connect peripheral devices used with the computer system. More specifically, a display may be connected and configured to display output from the computer system. For example, the display may show a 2D rendering of a generated digital 3D mesh. Input devices may also be connected to the I / O interface. Examples of such input devices include a keyboard and mouse, which allow users to interact with the computer system. A network interface may also be included as part of the computer system to allow it to connect to a suitable computer network, thereby receiving data (e.g., scan data, subscans, and / or images) from other computing devices and transferring data (e.g., scan data, subscans, and / or images) to other computing devices. The CPU, volatile memory, hard disk drives, I / O interfaces, and network interfaces may be connected together via a bus.

[0121] The computer system is preferably configured to receive data from a scanning device, either directly from the scanning device or via a computer network, such as a wireless network. The data may include images, processed images, subscans, point clouds, data point sets, or other types of data. Wireless connections, wired connections, and / or combinations thereof may be used to transmit / receive the data. In some embodiments, the computer system is configured to generate a digital representation of the dental object described herein. In some embodiments, the computer system is configured to receive data (such as subscans or point clouds) from a scanning device, then perform a reconstruction step and render the digital representation of the dental object. Rendering can be understood as the process of generating one or more images from three-dimensional data. The computer system may include a computer memory for storing a computer program including computer-executable instructions that, when executed, cause the computer system to perform a method for generating a digital representation of the dental object.

[0122] Obtain 2D images

[0123] The methods disclosed herein include the steps of acquiring or obtaining one or more 2D images. The 2D images can be digital images, such as digital color images. 2D images can be acquired using scanning devices disclosed herein, such as intraoral scanning devices. Each camera unit may include an image sensor having an array of pixels. Each pixel may be associated with a corresponding camera ray originating from the pixel in three-dimensional (3D) space. Typically, a point located along a given camera ray corresponding to a given pixel is imaged onto the pixel of the image sensor, and the pixel is read to generate an image pixel of the 2D image.

[0124] Each 2D image can consist of an array of image pixels, such as a two-dimensional image pixel array, corresponding to a pixel array on one or more image sensors. Each image pixel can include a pixel color defined by one or more color channels. An example of a single color channel is a grayscale color channel, where the value of each pixel represents only the amount of light, thus carrying intensity information. Therefore, images in an image set may be grayscale images. Another example of a single color channel is red, green, or blue. Therefore, the currently disclosed methods are not limited to using pixel color information; instead, pixel intensity information can also be utilized.

[0125] The scanning system can be configured to continuously acquire / obtain 2D images at a predetermined frame rate during a scanning procedure, wherein one or more objects (such as a patient's teeth) are scanned using an intraoral scanning device.

[0126] Generate point cloud data

[0127] A point cloud is a collection of discrete data points in space. These points can represent 3D shapes or objects, in this case, dental objects. Each point location has its own set of Cartesian coordinates (X, Y, Z). Points can contain data other than location, such as RGB color, normals, timestamps, etc. As mentioned above, several principles for extracting depth information from one or more 2D images are known, such as triangulation or focusing principles. Once one or more 2D images have been obtained, depth information is extracted to generate a depth image, where image pixels are assigned depth Z values, thus effectively converting one or more 2D images or processed versions of one or more 2D images from 2D information to 3D.

[0128] When scanning dental objects using a scanning device, the captured point clouds generated from multiple 2D images each contain scene fragments acquired at different times and / or locations. These need to be aligned to generate a complete distribution map of the scanned dental object. This process, known as point set registration, or simply registration or stitching, aligns the point clouds obtained from different 2D images to each other, placing them in a common coordinate system to generate the final point cloud data. The Iterative Closest Point (ICP) algorithm can be used to align two point clouds that overlap and are separated by rigid body transformations.

[0129] Surface part

[0130] Generally, the more polygons used to represent a surface feature being scanned (i.e., the higher the resolution), the finer the resolution of that feature. However, finer details only translate into higher accuracy when the object being scanned itself has fine features. Therefore, dental objects with sparse or rough surface features (e.g., flat surfaces and / or generally uniform color) can be adequately represented with fewer polygons without sacrificing the accuracy of the final 3D mesh. On the other hand, dental objects with fine surface features (e.g., sharp edges, textured / rough surfaces, grooves, etc.) and / or diverse colors may benefit from being represented with more polygons (i.e., at a higher resolution), as this allows the mesh to display finer details.

[0131] However, dental objects are rarely composed entirely of either coarse or fine features, but rather a combination of both. Therefore, according to the invention, mesh generation can advantageously involve dividing the scan data into multiple distinct surface portions, allowing each individual surface portion to be processed according to the requirements of the features presented by the dental object in that portion. As an example, a dental object can be a patient's mandibular teeth, including restorative sites. This can be divided into multiple surface portions, where more detailed surface portions (e.g., those including restorative sites or occlusal surfaces) can be processed at a higher resolution, while less detailed surface portions (e.g., the sides of teeth and gums) can be processed at a lower resolution, resulting in a variable-resolution mesh that uses high resolution for surface portions requiring high resolution and low resolution for surface portions where fewer polygons may suffice.

[0132] Geometric information / data

[0133] To determine the required resolution for a specific surface portion, geometric information needs to be extracted from one or more 2D images and / or point cloud data. Geometric information indicates the rate of change of one or more surface properties; in other words, it provides information about the amount of variation of a dental object in a given area, and thus also information about the amount of detail present in the surface portion corresponding to that area.

[0134] One or more surface properties may include a profile, in which case geometric information will indicate how much curvature or how small the curvature is in the various surface portions. The profile can be estimated based on the point distribution of point cloud data for a given surface portion. Wide point diffusion may indicate that the surface portion has high curvature and / or a large number of sharp features, i.e., a high profile rate of change; while low point diffusion may indicate that the surface portion has low curvature and / or fewer sharp features, i.e., a low profile rate of change.

[0135] One or more surface properties may include color, in which case geometric information will indicate the degree of color variation in the various surface portions. Color can be estimated based on 2D images and / or point cloud data of a given surface portion. Surface portions with relatively high color variation will have a high rate of color variation, while surface portions with relatively low color variation will have a low rate of color variation.

[0136] Determine the grid resolution

[0137] After obtaining the geometric information of each surface portion, the mesh resolution of each surface portion can be determined based on the rate of change of its surface properties (also known as surface features). As mentioned above, a high rate of change of a surface portion will result in a higher determined mesh resolution; conversely, a lower rate of change will result in a lower determined mesh resolution. Therefore, each surface portion will be rendered at a mesh resolution suitable for the surface features in that surface portion of the dental object, without unnecessarily consuming processing power to render the surface portion at an excessively high mesh resolution. Besides unnecessary processing power, this would also lead to excessive storage / data transfer requirements.

[0138] The mesh resolution can be selected from two or more fixed resolution levels, each associated with a range of rates of change for one or more surface properties. That is, if the rate of change for one or more surface properties falls within the range associated with that resolution level, then that resolution level is selected for that surface portion. The two or more fixed resolution levels can include a first resolution level (e.g., approximately 45 polygons per square millimeter) and a second mesh resolution level (e.g., approximately 180 polygons per square millimeter). When the rate of change for one or more surface properties falls within the first range, the first resolution level can be determined to be used for the surface portion; when the rate of change for one or more surface properties falls within the second range, the second resolution level can be determined to be used for the surface portion, where the midpoint of the first range is lower than the midpoint of the second range. Alternatively, the mesh resolution can be selected from a continuous range of resolutions based on geometric information.

[0139] Generate digital 3D mesh

[0140] Points from point cloud data are used to form a digital 3D mesh, which comprises multiple polygons representing the surfaces of the scanned dental object. These polygons are connected by their common edges (i.e., sides) and vertices (i.e., corners). In a preferred embodiment, the polygons are triangles. Each triangle, and indeed all other polygons, has a face, defined by its edges. For a flat face (such as the face of a triangle), this face has a normal, also called a normal vector, which indicates the orientation of the face in space. Polygons with more than three edges may have non-flat faces, which do not have unique normal vectors, i.e., the normal vectors depend on the position on the face. Methods for generating meshes from point cloud data (such as Delaunay triangulation) are known and used for this purpose. Another method could be to construct a directed distance field of volume from one or more 2D images and / or point clouds using Poisson reconstruction, and then extract an initial mesh as an isosurface / contour of the directed distance field at zero (using a marching cubes method).

[0141] The number of vertices used in a digital 3D mesh varies and depends on the mesh resolution determined for each surface portion. Therefore, in surface portions where the geometric information indicates a high rate of change in one or more surface properties, more vertices will be inserted into the digital 3D mesh so that the mesh can accurately represent the surface features in these portions. Similarly, in surface portions where the geometric information indicates a low rate of change in one or more surface properties, fewer vertices will be inserted into the digital 3D mesh so that processing power is not unnecessarily consumed to render these surface portions at excessively high resolutions.

[0142] Determine the circumferential mesh boundary in a digital 3D mesh

[0143] To determine the circumferential mesh boundaries in a digital 3D mesh, it may be necessary to identify the region of interest (ROI) within the dental object that corresponds to the current dental procedure. In the case of restorative dentistry, the ROI may be the edge line of the scanned preparation site.

[0144] A digital 3D mesh can be input into a trained model (e.g., running on the NPU of a processing device) that is trained to identify regions of interest (ROIs) (e.g., the edges of a pre-defined area). The trained model can output a sub-region of the digital 3D mesh defined by a circumferential mesh boundary that includes the ROI but excludes the circumferential mesh boundary. The trained model can perform a coarse identification of the ROI and set the circumferential mesh boundary at a buffer distance to the identified ROI (e.g., the edge line), ensuring that the entire ROI is included within the determined circumferential mesh boundary. Alternatively or additionally, the circumferential mesh boundary in the digital 3D mesh can be determined based on user input, for example, by identifying the circumferential mesh boundary in a graphical user interface (GUI) displaying the rendering of the 3D mesh, where the user can input the circumferential mesh boundary or correct the coarse identification of the ROI by the processing device. The sub-region defined by the circumferential mesh boundary can be extracted from the digital 3D mesh for subsequent processing steps.

[0145] Determine the target shape using the circumferential target boundary.

[0146] Generally, the target shape can be arbitrary, but for a mesh flattened into a 2D mesh, the target shape must be planar. The target shape can often be chosen as a circle because the continuous rotational symmetry of a circle allows for less deformation within the mesh after flattening.

[0147] Perform consistent mapping

[0148] Mesh flattening methods involve marking so-called anchor points. Anchor points mark the fixed positions of vertices on the mesh boundary, which can be original vertex positions or entirely new vertex positions. Before flattening a 3D mesh into a 2D mesh, the boundary vertices should be anchored / translated to the circumferential target boundary of the planar target shape. To achieve this, each vertex of the boundary vertex undergoes a consistent mapping to find the corresponding anchor point on the target boundary.

[0149] Consistent mapping should include several key attributes, such as:

[0150] Uniqueness: Each boundary vertex is unique All should be mapped to a single anchor point Here R n The dimension number, when flattened to a 2D mesh, would be... .

[0151] Order Preservation: Each boundary vertex should be mapped in a transitive manner for placement along the target boundary, i.e., if the order of the boundary vertices is... Then the mapping It should be ensured that the order remains unchanged after mapping. .

[0152] These properties are designed to eliminate the possibility of introducing self-intersections into the mesh during the mapping process.

[0153] The goal of the method and system of this invention is to obtain an ordered set of boundary vertices in a manner that satisfies the aforementioned uniqueness and order preservation properties. A set of anchor points that are consistently mapped onto the boundary of the target shape. First, a general description of the method involved in this invention will be given. Given an ordered set of boundary vertices... :

[0154] 1. Calculate the total boundary length / perimeter of the circumferential mesh boundary in a digital 3D mesh, i.e., the sum of the lengths of all edges. For the boundary of the loop, .

[0155] 2. Calculate the total length of the target boundary of the target shape. This is usually determined by the choice of the target shape.

[0156] 3. Construct the initial boundary vertices of the circumferential mesh boundary in the digital 3D mesh. and its corresponding initial anchor points mapped onto the target shape boundary. initial pair Initial boundary vertices The initial anchor point can be chosen arbitrarily from the boundary vertices. It can be set to any point on the circumferential target boundary.

[0157] 4. Map the remaining boundary vertices to the target shape using the following rules. For Each boundary vertex .

[0158] a. Calculate from arrive The distance of the covered boundary,

[0159] .

[0160] b. Calculate from arrive The relative distance covered .

[0161] c. Calculation For the circumferential target boundary, it is equivalent to the circumferential target boundary from arrive Coverage relative distance Anchor point.

[0162] Now, a method implementation is disclosed for a target shape of a unit disk in the XY plane (with its normal along the Z-axis). Given an ordered set of boundary vertices... :

[0163] 1. Calculate the total boundary length of the circumferential mesh boundary in a digital 3D mesh, i.e., the sum of the lengths of all edges. For the boundary of the loop, .

[0164] 2. Taking a unit disk as the target shape, the boundary length of the target shape's perimeter is known to be... .

[0165] 3. By setting To construct the initial pair ,in It is an arbitrary radian offset.

[0166] 4. Map the remaining boundary vertices to their corresponding anchor points on the circumferential target boundary using the following rules. For Each boundary vertex .

[0167] a. Calculate from arrive The distance of the covered boundary,

[0168] .

[0169] b. Calculate from arrive The relative distance covered .

[0170] c. Calculation Because rotating around a unit circle for the entire distance is equivalent to rotating... radian.

[0171] The obtained set of anchor points (also known as the boundary vertices of the mapping). It can be used as an anchor point for subsequent steps of flattening vertices in a digital 3D mesh that are surrounded by circumferential mesh boundaries to the target shape.

[0172] Grid flattening

[0173] According to the present invention, after mapping and translating the boundary vertices to the target boundary, the interior vertices (i.e., vertices surrounded by the circumferential mesh boundary) can be flattened to the target shape. Flattening the interior vertices can be performed in accordance with the disclosure in US11321918B2 (the entire contents of which are incorporated herein by reference).

[0174] Figure 1aAn intraoral scanner 100 used in the scanning system 700 or method of the present invention is illustrated. The intraoral scanner 100 includes a distal end 102 configured to be at least partially inserted into a patient's oral cavity. The distal end 102 includes a scanner window 150 through which light can be projected onto a dental object 202 to be scanned, and an image of the illuminated dental object 202 can be captured through the scanner window 150. The distal end 102 is also configured to releasably attach a sleeve 110 thereto. The sleeve 110 is provided as a hygienic barrier and is replaceable between patients to prevent patient-to-patient contamination. The sleeve 110 includes a sleeve window that at least partially overlaps with the scanner window 150 when the sleeve 110 is attached to the distal end 102.

[0175] The intraoral scanner 100 also includes a proximal end 104 opposite the distal end 102. The proximal end 104 may include an opening configured to receive and accommodate a battery 120 (preferably a rechargeable battery) to power the intraoral scanner 100. Alternatively, the battery 120 may be disposed internally within the intraoral scanner 100, in which case it can be charged via a power interface of the intraoral scanner 100. Alternatively, the intraoral scanner 100 may employ a wired connection, in which case power may be supplied by a wire instead of the battery 120. Between the distal end 102 and the proximal end 104, the intraoral scanner 100 includes a scanner body 106 configured for holding by a user (e.g., a dentist or dental technician). To enable the user to control the intraoral scanner 100 during a scanning procedure, the intraoral scanner 100 includes a scanner interface 108, which may include one or more buttons and / or one or more touch interfaces.

[0176] Figure 1b The illustration shows a cross-sectional view of the distal end 102 of an intraoral scanner 100 used in the scanning system 700 or method of the present invention. The illustrated intraoral scanner 100 includes a plurality of camera units 140, such as four or eight camera units 140, each camera unit 140 configured to acquire a 2D image of the object 202. By having multiple camera units 140, the intraoral scanner 100 can acquire multiple 2D images substantially simultaneously. The illustrated intraoral scanner 100 includes a plurality of projector units 130 configured to project light (e.g., patterned light) onto the dental object 202. Both the camera units 140 and the projector units 150 are arranged such that their optical axes are parallel to or slightly inclined to the longitudinal axis of the intraoral scanner 100. To allow the camera units 140 and the projector units 150 to observe / illuminate the object 202, mirrors 180 are arranged in their optical paths to guide their light towards the scanner window 170.

[0177] Figure 2 The illustration shows a laboratory scanner 200 used in the scanning system 700 or method of the present invention. Intraoral scanners 100 are typically used in dental clinics, while laboratory scanners 200 are primarily used in dental laboratories that manufacture or customize various products for dental clinics. The laboratory scanner 200 is configured to scan a plastic or plaster model 202 molded from a subject's teeth, the model being arranged within the scanning volume of the laboratory scanner 200.

[0178] Figure 3 A block diagram of the scanning system 700 of the present invention is shown. The scanning system 700 includes scanning devices 100 and 200 (which may be an intraoral scanner 100 or a laboratory scanner 200), as shown in Figure 1 and... Figure 2 The scanner shown is a scanning device 100, 200. The scanning devices 100, 200 include one or more projector units 130 and one or more camera units 140, the one or more projector units 130 being configured to illuminate a dental object 202, and the one or more camera units 140 being configured to capture a 2D image of the illuminated dental object 202.

[0179] Scanning devices 100 and 200 are operable and connected to scanning station 300 (such as a laptop or desktop computer) at least during the scanning procedure. Scanning station 300 includes input device 306 (such as a mouse, touchpad, and / or keyboard) that a user can use to control portions of the scanning process and / or subsequently process 2D images to generate a 3D mesh. Scanning station 300 includes a display that can be used to display the 3D mesh and / or provide a viewfinder to help users manipulate and guide the intraoral scanner 100 so that they can position the intraoral scanner 100 relative to the dental object 402.

[0180] To process one or more 2D images captured during a scanning procedure, the scanning system includes a processing device comprising one or more processors 150, 350. The one or more processors 150, 350 may be located in scanning devices 100, 200 and / or scanning station 300. The scanning system 700 also includes computer-readable media 160, 360, comprising instructions that, when executed by the scanning system 700, cause scanning devices 100, 200 and / or scanning station 300 to perform the methods of the present invention. Computer-readable media 160, 360 may be located in scanning devices 100, 200 and / or scanning station 300.

[0181] Figure 4 A block diagram of another scanning system 700 of the present invention is shown. Figure 4 The scanning system 700 and Figure 3The difference in the scanning system 700 is that it also includes a server 400 that operates connected to the scanning station 300. The server 700 may be located locally, i.e., close to the scanning station 300, and operate connected to the scanning station 300 via a local area network (LAN). The server 700 may also be located remotely, i.e., far from the scanning station 300, and operate connected to the scanning station 300 via a wide area network (WAN) (e.g., the Internet). The server 400 may be configured to perform at least a portion of the methods of the present invention, such as some or all of the steps related to data processing. For this purpose, the server 400 includes at least some (if not all) of one or more processors 150, 350, 850, and optionally also includes some or all of computer-readable media 160, 360, 860. Generally, processing devices, namely one or more processors 150, 350, 850 and / or computer-readable media 160, 360, 860, may be present in one of scanning devices 100, 200, scanning station 300 or server 400, or they may be distributed among two or three of these devices in any combination.

[0182] The scanning system 700 may also include a 3D printer and / or a 3D milling machine 302, configured to manufacture objects for dental treatment based on data generated by the scanning system 700. If the scanning system 700 is used for a patient requiring restorative dentistry, the 3D printer and / or the 3D milling machine 302 can be used to manufacture dental restorations, such as bridges, crowns, or veneers, based on digital 3D mesh and edge line data (i.e., edge lines identified in the 3D mesh by the method of the present invention) generated by the scanning system.

[0183] Figure 5 A visualized point cloud data 500 is shown, comprising multiple points 502 representing surface points of the dental object 202. The point cloud data 500 is generated by extracting depth information from one or more 2D images acquired during the scanning procedure. The depth information is extracted based on the operating principles of the scanning devices 100 and 200. A triangulation-based scanning system 700 determines a point in 3D space by projecting it onto two or more camera pixels and / or one or more projector pixels, while a focus-based scanning system 700 determines depth by linearly moving a lens and determining the focusing distance in 3D space when the given point is in focus based on the lens's position.

[0184] Point cloud data 500 collected from multiple starting points (i.e., the positions and / or orientations of scanners 100, 200 relative to dental object 202) are registered into a common coordinate system. Registration can be performed using the Iterative Closest Point (ICP) algorithm, which fixes one point cloud (reference or target) while transforming another point cloud (source) to best match the reference. The transformations (i.e., translations and / or rotations) are iteratively estimated to minimize the error value, typically the sum of the squared differences between the paired coordinates. As shown in the figure, the final point cloud data represents dental object 202, i.e., the entire dental arch of patient 400's mandible.

[0185] Figure 6 A 3D mesh 600 of a dental object 202, which is the entire dental arch of the mandible of a patient 400, is shown. The 3D mesh 600 comprises a digital 3D mesh composed of a plurality of polygons 608, which represent the surface of the scanned dental object 202. Triangular polygons 608 are advantageous for the 3D mesh for a number of reasons.

[0186] Triangles are always planar, meaning they lie flat on their respective planes. This simplifies many mathematical calculations, such as rendering, shading, and collision detection. Linear interpolation on triangles is straightforward, which is crucial for shading techniques like Gouraud shading and texture mapping.

[0187] Any surface, regardless of its complexity or curvature, can be approximated by a sufficient number of triangles. This makes triangles highly versatile in modeling a wide variety of shapes. Triangular meshes allow for different levels of detail. Complex surfaces can be represented by a large number of small triangles for areas with rich detail, and by fewer, larger triangles for areas with less detail. Therefore, triangular meshes offer a high degree of geometric flexibility.

[0188] Graphics hardware (GPUs) is optimized for processing triangles. Operations such as transformations, lighting calculations, and rasterization are highly efficient with triangles. Algorithms for mesh operations (such as subdivision, simplification, and deformation) for triangular meshes are well-developed and efficient. Furthermore, many 3D graphics standards and file formats (such as STL, OBJ, and 3DS) natively support triangular meshes, ensuring broad compatibility across different software and platforms. Most 3D modeling and animation software is designed to handle triangular meshes, providing robust tools for editing and manipulating these meshes. Therefore, triangular meshes offer high compatibility with a wide range of hardware / software platforms that users might use to process 3D meshes.

[0189] Triangles ensure consistent behavior across various operations. For example, a polygon with more sides (like 608) might be non-planar or concave, complicating computation. Triangular meshes can be easily subdivided to create finer details. Subdivision algorithms like Loop Subdivision handle triangles flawlessly. Reducing the number of triangles in a mesh while preserving the overall shape and appearance is a well-studied problem, with many efficient algorithms available. In summary, triangular meshes offer a balance of mathematical simplicity, geometric flexibility, and computational efficiency, making them ideal for representing 3D objects.

[0190] As will be described in more detail below, the number of polygons 608 for each surface portion of the digital 3D mesh will depend on the regional level of detail of the respective surface portion, as indicated by the geometric information associated with that surface portion.

[0191] Figures 7a to 7e Examples of meshes according to various embodiments are shown. As described above, a digital 3D mesh / 2D mesh comprises multiple polygons 608. A polygon 608 is defined by its corners (i.e., vertices 602) and its edges 604. The region enclosed by a number of edges (depending on the type of polygon 608 applied (e.g., a triangle has three edges)) is called a face 606. In various embodiments, the mesh may differ in several aspects. Faces 606 may have a consistent number of vertices or a different number of vertices. Edges 604 may have a consistent length or different lengths. Non-boundary vertices may have a different number of adjacent vertices or a consistent number of adjacent vertices.

[0192] Figure 7a It is a mesh according to one embodiment, the mesh having faces with different numbers of vertices 602, edges 604 of different lengths, and non-boundary vertices 602 with different numbers of adjacent vertices 602. Figure 7b It is a mesh according to one embodiment, the mesh having faces 606 with a consistent number of vertices 602, edges 604 of varying lengths, and non-boundary vertices 602 with varying numbers of adjacent vertices 602.

[0193] Figure 7c It is a mesh according to one embodiment, the mesh having faces 606 with a consistent number of vertices 602, edges 604 of consistent length, and non-boundary vertices 602 with a consistent number of adjacent vertices 602. Figure 7d This is a mesh according to one embodiment, which has faces 606 with a consistent number of vertices 602, edges 602 of consistent length, and non-boundary vertices 602 with a consistent number of adjacent vertices 602. Different meshes serve different functions, but when representing 3D objects, meshes with triangular faces 606 (such as...) Figure 7b (As shown) is often used for the reasons mentioned above.

[0194] Each polygon 608 can be assigned a color value so that the digital 3D mesh represents not only the shape of the scanned dental object 202, but also its color. This is advantageous when dental restorations (such as bridges, crowns, and / or veneers) are obtained from the 3D mesh 600, because the restoration(s) can be color-matched to the patient's other teeth (such as adjacent teeth to the restoration site where the restoration will be placed).

[0195] Figure 7e It is a triangular grid used to explain the terminology in this article. Figure 7e A sub-section 642 of a 3D mesh 600 is shown. Sub-section 642 consists of a circumferential mesh boundary 640 ( Figure 7e The circumferential mesh boundary 640 is defined by the boundary vertex 644 (shown by the dashed line in the image). Figure 7e The solid circle shown in the diagram) and the edges connecting the boundary vertices 644 (these edges may also be called boundary edges) are defined. In some embodiments, the total distance around the circumferential mesh boundary 640 is used to map the boundary vertices 644 to their corresponding anchor points. The total distance around the circumferential mesh boundary 640 is the sum of the lengths of the boundary edges. Vertices surrounded by the boundary vertices 644 are called inner vertices 646 (e.g., solid circles). Figure 7e (as shown by the solid circle in the image).

[0196] Figures 8a to 8c An example of a 3D mesh 600 generated according to the present invention is shown. The 3D mesh 600 represents a dental object 202, i.e., a collection of teeth and gums. The digital 3D mesh used for the 3D mesh comprises multiple triangles, wherein the resolution (i.e., the number of polygons 608) of the various surface portions of the mesh depends on the level of detail.

[0197] Figure 8b A first region 610 of a digital 3D mesh is shown. The first region 610 contains the lingual surface of an incisor, a relatively flat surface with a relatively low level of detail / variation in its surface characteristics. Accordingly, the resolution for determining the surface portion 612 applicable to this region 610 is also low. Therefore, the mesh generated in the surface portion 612 of this region 610 comprises fewer polygons 608 per region, as this is sufficient to resolve the relatively coarse features of the first region 610 without expending excessive processing power on generating redundant polygons 608. Figure 8cA second region 620 of the digital 3D mesh is shown. The second region 620 contains the incisal surface of the incisor, a highly curved surface with a relatively high level of detail in its surface features. Accordingly, the resolution of the surface portion applicable to this region 620 is also high. Therefore, the mesh generated in the surface portion 614 of this region 620 includes more polygons 608 per region, enabling the mesh to accurately represent fine features. For ease of comparison, Figure 8c and Figure 8b Both were magnified by the same factor. By comparing the two, it is easy to see that the relatively flat lingual surface of the tooth located in the first region 610 was resolved at a lower resolution, resulting in a lower density of its polygon 608 than that of the incisal surface located in the second region 620.

[0198] Figure 9 Another 3D mesh 600 of a dental object 202, which is a molar of a patient 400, is shown. As can be seen from the enlarged view on the right, the relatively flat sides have been identified as having a low rate of change in surface properties (in this case, a low rate of change in profile), and therefore, the mesh resolution determined for these surface portions has also been determined to be low. On the other hand, the occlusal surface has been identified as having a high rate of change in surface properties (in this case, a high rate of change in profile), and therefore, the mesh resolution determined for these surface portions has also been determined to be high.

[0199] To determine the required mesh resolution for a surface portion, its geometric information must first be obtained. This geometric information indicates the rate of change of one or more surface properties (such as the rate of change of profile and / or color), thus providing the information needed to determine the required mesh resolution for the 3D mesh 600 to accurately represent the surface of the dental object 202 within this surface portion.

[0200] Geometric information is obtained from one or more 2D images and / or point cloud data 500, depending on the surface characteristics required. If color information is needed, it can be obtained from color information in the 2D images and / or from color information extracted from point cloud data (one or more) of the 2D images. If contour information is needed, it can be obtained from the point cloud data 500. Contour information can be obtained by comparing surface gradients at adjacent points, for example, by performing ray tracing to determine how light from one or more projector units 130 is reflected at points in the point cloud data and comparing it to the reflection of light at its adjacent points; or by determining how densely or sparsely distributed the points are in a surface portion, i.e., sparsely distributed points indicate a high degree of surface contour variation, while densely distributed points indicate a low degree of surface contour variation.

[0201] Figure 10 The illustration shows an initial mesh generated from point cloud data 500, comprising multiple initial polygons. The initial mesh is generated at a preset initial resolution prior to the 3D mesh. The initial mesh can be used to obtain at least some geometric information. This can be done by deriving the initial mesh into a data fit error metric, which indicates how well the initial polygons fit the point cloud data 500 used to generate them. If points are highly distributed around the initial polygons, the geometric information will indicate a poor fit error; if points are distributed close to the initial polygons, the geometric information will indicate a good fit error. A poor fit error will mean that a higher mesh resolution would be preferable; while a good fit error will mean that a lower mesh resolution would be sufficient.

[0202] Figure 11a and Figure 11b Two examples of boundary vertex mapping according to existing technology are illustrated. Figure 11a In this context, the circumferential mesh boundary 640, represented by triangles, is monotonic, meaning that as one moves around the circumferential mesh boundary 640, the radial angle 660 only increases until a starting point on the circumferential mesh boundary 640 is reached. For such a circumferential mesh boundary 640, known mapping rules in the prior art will produce an effective mapping, because the mapping rules in the prior art use the radial angle 660 of the boundary vertices x1, x2, x3 to determine the anchor points x'1, x'2, x'3 on the circumferential target boundary 650.

[0203] However, not all circumferential mesh boundaries 640 can be assumed to be monotonic, such as... Figure 11b As shown. In Figure 11b In this process, the circumferential mesh boundary 640 bends towards itself, causing the radial angle 660 to decrease when moving from boundary vertex x1 to boundary vertex x2. When the mapping rules of the prior art are used on such a circumferential mesh boundary 640, the requirements of uniqueness and order preservation are no longer met, resulting in self-intersection along the target boundary 650. Therefore, the obtained anchor points x'1, x'2, and x'3 will be invalid.

[0204] Figure 12 An example of a 3D mesh 600 with a circumferential mesh boundary 640 is shown, which cannot be guaranteed to be monotonic in the form of edge lines 630. Figure 12 The advantages of grid flattening are further illustrated because the edge line 630 cannot be fully identified because its portion is hidden (i.e. occluded) from the view of the tooth being restored 632 (also known as the preparation site).

[0205] Figure 13 This is a flowchart illustrating an embodiment of the method of the present invention, which can be used to flatten a 3D mesh 600, for example, for (e.g., from) Figure 12 (Edge lines 630 are identified / extracted in the 3D mesh 600 shown). The method begins by obtaining a digital 3D mesh 800 representing the dental object 202. This can be done by importing a 3D mesh 600 acquired elsewhere, or by scanning the patient with an intraoral scanner 100 or by scanning impressions of the patient's teeth with a laboratory scanner 200. If the 3D mesh 600 is acquired by scanning, the step of obtaining the digital 3D mesh 800 representing the dental object 202 may also require image processing to extract depth information from 2D images acquired by the intraoral / laboratory scanners 100, 200 to generate the 3D mesh 600. This may also involve obtaining geometric information from the acquired 2D images or scan data generated by processing the acquired 2D images to determine the resolution of the surface portions of the 3D mesh 600.

[0206] Having obtained an 800-dimensional 3D mesh 600, the method includes determining a circumferential mesh boundary 640 within the 810-dimensional digital 3D mesh 600. This may involve detecting user input, such as a user drawing a closed boundary around a region of interest in a GUI 900 displaying the 3D mesh 600. Alternatively, the circumferential mesh boundary 640 can be automatically set by an algorithm (e.g., a trained model) configured to coarsely identify a region of interest 644 (e.g., an edge line 630). After coarse identification, a buffer distance can be added to ensure that the region of interest 644 falls within the determined circumferential mesh boundary 640. The sub-portion of the 3D mesh 600 defined by the circumferential mesh boundary 640 (i.e., the portion of the 3D mesh 600 falling within the circumferential mesh boundary 640) can be extracted from the rest of the 3D mesh. In this paper, vertices forming the circumferential mesh boundary 640 are referred to as boundary vertices, and vertices falling within the circumferential mesh boundary 640 are referred to as interior vertices.

[0207] Next, the method includes determining the target shape 820 having a circumferential target boundary 650. The disclosed method can be applied to any arbitrary 2D shape; however, a circle will result in less edge distortion due to its continuous rotational symmetry.

[0208] To flatten a sub-part of the 3D mesh 642 falling within the circumferential mesh boundary 640, the method includes performing a consistent mapping of each of the 830 boundary vertices to an anchor point on the target boundary 650. This step is performed in such a manner that the mapping pair of each boundary vertex and anchor point satisfies the aforementioned uniqueness and order preservation criteria.

[0209] The above discloses a method to ensure that these criteria are met, which includes: selecting an initial vertex of the boundary vertices and an initial anchor point on the target boundary, and mapping the initial vertex to the initial anchor point; then, for each vertex in the boundary vertices other than the initial vertex, determining the relative distance (i.e., relative to the total perimeter of the mesh boundary 640) between the corresponding vertex and the initial vertex; and, for each vertex in the boundary vertices other than the initial vertex, mapping the corresponding vertex to an anchor point whose relative distance (i.e., relative to the total perimeter of the target boundary 650) from the initial anchor point along the target boundary 650 is the same as the relative distance between the corresponding vertex and the initial vertex.

[0210] The method may then include translating each boundary vertex 840 to its corresponding anchor point on the target boundary 650, and flattening the interior vertices 850 to the target shape based on the anchor points to generate a flattened mesh 652. The method may then include outputting the resulting flattened mesh 652 at 890 for further processing as required by the application, or for display on a monitor for user inspection.

[0211] Figure 14 It shows the relationship with Figure 13 The flowchart is similar to the flowchart. Figure 14 The method shown begins with the step of flattening the internal vertex 646 850 to the target shape to generate a flattened mesh 652, and continues by further including the following steps: identifying the edge line 630 of the dental object in the flattened mesh 652 860, and identifying the vertex or surface point in the flattened mesh 652 that represents the edge line 630. It is worth noting that, unlike vertices, the surface points of the flattened mesh 652 can be located on vertices 602, edges 604, or faces 606 of the flattened mesh 652. Therefore, identifying the surface point representing the edge line 630 will provide a more accurate identification of the edge line than using the vertices of the flattened mesh. This step can be based on user input, for example, the user identifying the edge line 630 by inputting in a GUI displaying the flattened mesh 652. This step can be performed wholly or partially by a second trained model configured to identify edge lines in a 2D mesh.

[0212] This method can proceed by identifying vertices or surface points representing edge lines 630 in the 864 3D meshes 600 based on the identified vertices or surface points representing edge lines 630 in the flattened mesh 652. This can be done by determining which vertices of the 3D mesh 600 correspond to those vertices in the flattened mesh 652 identified as representing edge lines 630. Alternatively, it can be done by tracing back the surface points in the flattened mesh 652 identified as representing edge lines to the 3D mesh 600, thereby identifying the edge lines in the 3D mesh 600 and the surface points representing edge lines 630 in the 3D mesh 600. It is worth noting that, unlike vertices, surface points in the 3D mesh 600 can be located on vertices 602, edges 604, or faces 606 of the 3D mesh 600. Therefore, identifying surface points representing edge lines 630 will provide a more accurate identification of the edge lines than using vertices.

[0213] This method can output a 3D mesh 600 with identified edge lines 630, for example, for user inspection on a display. Optionally, it can continue by generating edge line data 866 to identify the edge lines 630 in the 3D mesh 600. This method may be desirable if used in a restorative dental workflow that requires designing a restoration suitable for placement on a tooth with edge lines 630. Therefore, the method optionally includes: generating 870 3D dental restoration data representing the shape of a dental restoration configured for placement on the identified edge lines 630, based on the 3D mesh 600 and the edge line data. The 3D dental restoration data can then be output, for example, to a 3D printer or 3D milling machine, for the step 872 of manufacturing a dental restoration shaped to fit the preparation site and edge lines 630 based on the 3D dental restoration data.

[0214] Figures 15a to 15c The method according to the present invention is shown for flattening a 3D mesh 600. First, in Figure 15a In this process, a region of interest 644 is determined within a 3D mesh 600. In the example provided, the region of interest 644 is identified by a first trained model configured to coarsely identify edge lines 630 in the 3D mesh. The identified region of interest 644 is provided with a buffer distance, i.e., a circumferential buffer, thereby determining the circumferential mesh boundary 640 and ensuring that the edge lines 630 fall within the circumferential mesh boundary 640.

[0215] In Figure 15B, a sub-part 642 defined by the circumferential mesh boundary 640 in the 3D mesh is extracted from the rest of the 3D mesh. As described above, the boundary vertices constituting the circumferential mesh boundary 640 are consistently mapped onto a planar circle to determine their corresponding anchor points on the target boundary 650. The boundary vertices are then translated to their anchor points on the target boundary 650, and the internal vertices are flattened accordingly onto the planar circle. The resulting flattened mesh 652 is as follows: Figure 15c As shown.

[0216] Figures 16a to 16c It shows the relationship with Figures 15a to 15c Similar 3D mesh processing. Figure 16a A sub-part 642 of the 3D mesh 600 is shown after the step of determining the circumferential mesh boundary 640 in the 810 digital 3D mesh 600. The resulting sub-part 642 has boundary vertices defining the circumferential mesh boundary 640, and interior vertices representing the dental object (which is a preparatory site with edge lines 630). As described above, the determination of the 810 circumferential mesh boundary 640 can be performed using a first trained model configured to target the region of interest 642 in the 3D mesh. Figures 16a to 16c The edge lines in the image are roughly identified.

[0217] Figure 16b The flattened shape according to the method of the present invention is shown. Figure 16a The sub-part 642 generates a flattened mesh 652. Since the edge line 630 is fully visible in the flattened mesh 652, the edge line and the vertices representing it can be identified in the flattened mesh 652. This can be done by feeding the flattened mesh 652 into a second trained model configured to identify edge lines in 2D. Figure 16c The flattened mesh 652 and the identified edge line 654 (shown as dashed lines) are shown. The edge line 630 can then be identified in the 3D mesh 600 by identifying the vertices in the 3D mesh 600 that correspond to the vertices representing the edge line 630 in the flattened mesh 652, i.e., by identifying the edge lines in 2D.

[0218] Figure 17 The illustration shows the method of the present invention for edge line 630 detection. Figure 17The upper portion illustrates the graphical user interface (GUI) 900 of the scanning system 700 of the present invention, which can be displayed on one or more displays, such as a computer screen. The GUI 900 shows a rendering of a 3D mesh 600, in which sub-parts 642 identified by a first model are highlighted. In the illustrated embodiment, the circumferential mesh boundary 640 around the edge lines is determined by roughly identifying the edge lines 630 using a first model configured to identify pre-defined areas in 3D, and by adding a buffer around the identified pre-defined areas.

[0219] In the right half of the GUI 900, a flattened grid 652 of the sub-part 642 identified by the first model is shown, which is flattened by the method of the present invention. It is worth noting that the left and right halves of the GUI 900 can be shown sequentially on a single screen, or simultaneously on one or more screens (e.g., dual desktops). Figure 17 The lower half of the diagram illustrates the underlying processing of GUI 900. An image rendering of the flattened mesh 652 can be input into the second model, i.e. Figure 17 The step marked AI involves a model trained to recognize edge lines in 2D. The neural network (i.e., the second trained model) essentially operates on the rendering of images of a flattened mesh. These images are rendered using orthographic projection. In the image, each pixel can be traced back to a surface point (if it hits any surface) via orthographic projection, and vice versa. Thus, the second model, along with optional post-processing, can output the identified edge lines 654 in the flattened mesh 652. The identified edge lines 654 can be used to identify vertices or preferred surface points in the flattened mesh 652 that represent edge lines 630. These, in turn, can be traced back to the flattened mesh 652 shown in GUI 900 and optionally highlighted in GUI 900 for user inspection, and / or, traced back to the 3D mesh 600 shown in GUI 900 and optionally highlighted in GUI 900 for user inspection.

[0220] The identified flattened mesh and / or (one or more) 3D mesh edge lines 654 can be included in the edge line data indicating the identified edge lines 654, allowing the identified edge lines 654 to be used in the GUI 900 for purposes other than mere inspection. For example, the 3D mesh 600 (i.e., data including the 3D mesh) and edge line data can be used to design a dental prosthesis for placement on the preparation site 632. This may involve designing the inner surface (also known as the concave surface) of the dental prosthesis based on the outer surface of the preparation site as shown in the 3D mesh 600 and the identified edge lines 654 as shown in the edge line data, and designing the outer surface of the dental prosthesis based on the surface of the teeth adjacent to the preparation site 632 (e.g., to obtain a prosthesis that matches the height of the adjacent teeth) (as indicated by the identified edge lines 654 shown in the 3D mesh 600 and edge line data). The 3D mesh preferably also includes color / texture data that can be used to design dental restorations so that they can be designed with optical properties (e.g., color, translucency, hue, etc.) that match the teeth adjacent to the preparation site 632. The designed dental restoration can be included in 3D dental restoration data, which represents the 3D surface of the designed restoration and its preferred optical properties. The 3D dental restoration data can be input into a 3D printer and / or 3D milling machine, such as a local 3D printer / milling machine located within a dental clinic, or a remote 3D printer / milling machine located outside a dental clinic (e.g., in a dental laboratory).

[0221] Figure 18 The diagram illustrates the relationship with Figure 17 Similar methods and GUIs. Figure 17 The 3D mesh 600 in the image was obtained using an intraoral scanner 100, while Figure 18 The 3D mesh 600 is obtained on a plaster model of the dental object using a laboratory scanner 200. Regardless of how the 3D mesh 600 representing the dental object is obtained, the method of the present invention can be used for flattening due to the consistent mapping from boundary vertices to anchor points. As can be seen from the flattened mesh 652 or its image rendering, the preparation site 642 of the dental object at hand is prepared for a partial crown on the occlusal surface of one of the patient's first molars, i.e., a so-called high-inlay dental restoration. The disclosed method and system are particularly useful for such procedures because the boundary orientation of the edge lines 630 in such procedures is generally not monotonous, as can also be seen from the identified edge lines 654 in the lower right corner. Performing such procedures using prior art methods may result in problems when mapping boundary vertices to the target boundary.

[0222] Further details

[0223] Embodiments of the present invention are disclosed in the following list of items:

[0224] 1. A computer-implemented method for manipulating at least a portion of a digital 3D mesh representing a dental object, the method comprising:

[0225] • Obtain a digital 3D mesh;

[0226] • Determine the circumferential mesh boundary of the digital 3D mesh, wherein the mesh boundary includes multiple boundary vertices and one or more internal vertices surrounded by the boundary vertices;

[0227] • Determine the shape of the target having a circumferential target boundary, wherein the target boundary is planar; and

[0228] • For each vertex on the boundary, perform a consistent mapping to the anchor point corresponding to the vertex on the target boundary.

[0229] 2. According to the method of Clause 1, wherein the consistency mapping satisfies the order preservation criterion such that the order of the anchor point around the target boundary is the same as the order of its corresponding boundary vertex around the mesh boundary, and / or, wherein the consistency mapping satisfies the uniqueness criterion such that each unique boundary vertex maps to a unique anchor point.

[0230] 3. The method according to any of the foregoing clauses, wherein the consistency mapping includes:

[0231] o Select the initial vertex of the boundary vertex and the initial anchor point on the target boundary, and

[0232] o maps the initial vertex to the initial anchor point.

[0233] 4. According to the method in Clause 3, the consistency mapping includes:

[0234] For each vertex in the boundary vertices except the initial vertex, determine the relative distance between the corresponding vertex and the initial vertex along the mesh boundary, and

[0235] For each vertex in the boundary vertices other than the initial vertex, map the corresponding vertex to an anchor point along the target boundary at a distance equal to the relative distance between the initial anchor point and the corresponding vertex and the initial vertex.

[0236] 5. The method according to any of the foregoing clauses, wherein the method further comprises:

[0237] • For each boundary vertex, translate the corresponding boundary vertex to its corresponding anchor point on the target boundary.

[0238] 6. The method according to Clause 5, wherein the target shape is planar.

[0239] 7. The method according to Clauses 5 and 6, wherein the method further comprises:

[0240] Flatten the internal vertices to the target shape to generate a flattened mesh.

[0241] 8. The method according to any of the preceding clauses, wherein determining the circumferential mesh boundary in the digital 3D mesh includes:

[0242] To roughly identify the edge lines of dental objects in a digital 3D mesh, and

[0243] o Set a circumferential grid boundary around the identified edge lines.

[0244] 9. The method according to Clause 8, wherein a buffer distance is provided between the grid boundary and the roughly identified edge line.

[0245] 10. The method pursuant to Clause 7 and Clause 8 or 9, wherein the method further comprises:

[0246] • Identify edge lines in a flattened grid.

[0247] 11. The method according to Clause 10, wherein identifying edge lines in the flattened grid includes:

[0248] o determines the vertices representing the edge lines in the flattened mesh.

[0249] 12. The method according to Clause 11, wherein the method further comprises:

[0250] • Identify vertices in the digital 3D mesh that are identified as vertices representing edge lines in the flattened mesh.

[0251] 13. The method according to Clause 12, wherein the method further comprises:

[0252] • Determine the edge lines in the digital 3D mesh based on the vertices in the identified digital 3D mesh that correspond to the vertices of the determined flattened mesh.

[0253] 14. The method according to any of the preceding clauses, wherein the target shape is circular.

[0254] 15. The method according to any of the preceding clauses and clause 8, wherein a machine learning model trained for identifying edge lines is used to identify the edge lines of a dental object in a digital 3D mesh.

[0255] 16. The method according to any of the preceding clauses, wherein the step of obtaining the digital 3D mesh includes:

[0256] o Obtain one or more 2D images of a dental object;

[0257] o Generates point cloud data based on one or more 2D images, wherein the point cloud data includes multiple points representing surface points of a dental object; and

[0258] o Generates a digital 3D mesh based on point cloud data, which includes multiple polygons representing the surface of a dental object.

[0259] 17. The method according to Clause 16, wherein generating a digital 3D mesh comprises:

[0260] Geometric information is obtained based on one or more 2D images, wherein the geometric information indicates the rate of change of one or more surface properties in multiple surface portions of a dental object.

[0261] 18. The method according to Clause 17, wherein generating a digital 3D mesh comprises:

[0262] The mesh resolution for each surface portion is determined based on geometric information.

[0263] 19. The method according to Clause 18, wherein generating a digital 3D mesh comprises:

[0264] A mesh is generated for each surface portion, with its resolution determined based on the corresponding surface portion.

[0265] 20. The method according to any one of clauses 17 to 19, wherein at least a portion of the geometric information associated with the corresponding surface portion is based on the point distribution of the point cloud data corresponding to the surface portion.

[0266] 21. The method according to any one of clauses 17 to 20, wherein at least a portion of the geometric information associated with the corresponding surface portion is based on ray tracing of light reflected at points in the point cloud data corresponding to the surface portion.

[0267] 22. The method according to Clause 20 or 21, wherein the geometric information includes information indicating the rate of change of the profile of the surface portion.

[0268] 23. The method according to any one of clauses 17 to 22, wherein at least a portion of the geometric information associated with the corresponding surface portion is based on color data corresponding to that surface portion.

[0269] 24. The method according to Clause 23, wherein the geometric information includes information indicating the rate of color change of the surface portion.

[0270] 25. The method according to any one of clauses 16 to 24, wherein obtaining one or more 2D images of the object comprises: obtaining a plurality of 2D images of the object.

[0271] 26. The method according to Clause 25, wherein generating point cloud data includes: resolving a corresponding problem to determine which parts of a first image among a plurality of 2D images correspond to which parts of a second image among a plurality of 2D images.

[0272] 27. The method according to Clause 25 or 26, wherein generating point cloud data includes: registering points from a plurality of points obtained from each corresponding 2D image to a common coordinate system.

[0273] 28. The method according to any one of clauses 16 to 27, wherein generating a digital 3D mesh comprises: assigning a color value to each polygon based on one or more 2D images.

[0274] 29. The method according to any one of clauses 16 to 28, wherein obtaining one or more 2D images includes projecting light onto a dental object.

[0275] 30. The method according to Clause 29, wherein light is projected in the form of a pattern.

[0276] 31. The method according to Clause 30, wherein the projected pattern is static over time.

[0277] 32. The method according to Clause 30, wherein the projected pattern changes over time.

[0278] 33. The method according to any one of clauses 16 to 32, wherein obtaining one or more 3D images of a dental object is performed using an intraoral scanner.

[0279] 34. The method according to any one of clauses 16 to 32, wherein obtaining one or more 3D images of a dental object is performed using a laboratory scanner.

[0280] 35. The method according to clause 33 or 34, wherein generating point cloud data includes: performing a first transformation to generate a coordinate dataset of at least a portion of the object in the scanning device coordinate system.

[0281] 36. The method according to Clause 35, wherein generating point cloud data comprises: performing a second transformation on a coordinate dataset of at least a portion of an object in a scanning device coordinate system to generate at least a portion of point cloud data in a real-world coordinate system.

[0282] 37. The method according to any one of clauses 16 to 36, wherein generating the digital 3D mesh further includes:

[0283] An initial mesh is generated based on point cloud data, wherein the initial mesh includes multiple initial polygons.

[0284] 38. The method according to Clause 37, wherein the generation of the digital 3D mesh is based on an initial mesh.

[0285] 39. The method according to clause 37 or 38, wherein obtaining the geometric information of the surface portion includes:

[0286] Compare the normal vectors of adjacent initial polygons within the surface portion to obtain information about the rate of change of the surface portion's profile.

[0287] 40. The method pursuant to the foregoing provisions and any one of provisions 11 and / or 12, wherein the method further comprises:

[0288] • Generate 3D dental restoration data representing a dental restoration, which is configured to be placed on a tooth with an edge line, based on one or both of the vertices identified as representing edge lines in a digital 3D mesh and in a flattened mesh or in a digital 3D mesh.

[0289] 41. A computer program product comprising instructions that, when executed by a computer, cause the computer to perform the method described in any of the preceding clauses.

[0290] 42. A non-transient computer-readable medium comprising instructions that, when executed by a computer, cause the computer to perform any one of the methods described in clauses 1 to 40.

[0291] 43. A dental scanning system, comprising:

[0292] - Processing equipment, which is configured for:

[0293] • Obtain a digital 3D mesh;

[0294] • Determine the circumferential mesh boundary in the digital 3D mesh, wherein the mesh boundary includes multiple boundary vertices and one or more internal vertices surrounded by the boundary vertices;

[0295] • Determine the shape of the target having a circumferential target boundary, wherein the target boundary is planar; and

[0296] • For each vertex on the boundary, perform a consistent mapping to the anchor point corresponding to the vertex on the target boundary.

[0297] 44. A dental scanning system according to Clause 43, wherein the consistency mapping satisfies the order preservation criterion such that the order in which anchor points have around the target boundary is the same as the order in which their corresponding boundary vertices have around the mesh boundary.

[0298] 45. A dental scanning system according to clause 43 or 44, wherein the consistency mapping includes:

[0299] o Select the initial vertex of the boundary vertex and the initial anchor point on the target boundary, and

[0300] o maps the initial vertex to the initial anchor point.

[0301] 46. ​​A dental scanning system according to Clause 45, wherein the consistency mapping includes:

[0302] For each vertex in the boundary vertices except the initial vertex, determine the relative distance between the corresponding vertex and the initial vertex along the mesh boundary, and

[0303] For each vertex in the boundary vertices other than the initial vertex, map the corresponding vertex to an anchor point along the target boundary at a distance equal to the relative distance between the initial anchor point and the corresponding vertex and the initial vertex.

[0304] 47. A dental scanning system according to any one of clauses 43 to 46, wherein the processing device is further configured to:

[0305] • For each boundary vertex, translate the corresponding boundary vertex to its corresponding anchor point on the target boundary.

[0306] 48. A dental scanning system according to Clause 47, wherein the target shape is planar.

[0307] 49. A dental scanning system according to clause 47 or 48, wherein the processing device is further configured to:

[0308] Flatten the internal vertices to the target shape to generate a flattened mesh.

[0309] 50. A dental scanning system according to any one of clauses 43 to 49, wherein determining the circumferential mesh boundary in a digital 3D mesh includes:

[0310] o Identify the edge lines of dental objects in a digital 3D mesh, and

[0311] o Set a circumferential grid boundary around the identified edge lines.

[0312] 51. A dental scanning system according to Clause 50, wherein a buffer distance is provided between the grid boundary and the identified edge line.

[0313] 52. Dental scanning systems pursuant to Clause 49 and Clause 50 or 51, wherein the processing device is further configured to:

[0314] • Re-identify the edge lines in the flattened grid.

[0315] 53. A dental scanning system according to Clause 52, wherein re-identifying the edge lines in a flattened grid includes:

[0316] o determines the vertices representing the edge lines in the flattened mesh.

[0317] 54. A dental scanning system pursuant to Clause 53, wherein the processing device is further configured to:

[0318] • Identify vertices in the digital 3D mesh that are identified as vertices representing edge lines in the flattened mesh.

[0319] 55. A dental scanning system pursuant to Clause 54, wherein the processing device is further configured to:

[0320] • Determine the edge lines in the digital 3D mesh based on the vertices in the identified digital 3D mesh that correspond to the vertices of the determined flattened mesh.

[0321] 56. A dental scanning system according to any one of clauses 43 to 55, wherein the target shape is circular.

[0322] 57. A dental scanning system pursuant to any of Clauses 43 to 56 and Clause 50, using a machine learning model trained to identify edge lines to identify the edge lines of a dental object in a digital 3D mesh.

[0323] 58. A dental scanning system according to any one of clauses 43 to 57, wherein the system comprises:

[0324] - An intraoral scanner, which is configured to acquire one or more 2D images of a dental object;

[0325] The processing equipment is also configured to:

[0326] o Generate point cloud data based on one or more 2D images, wherein the point cloud data includes multiple points representing surface points of a dental object;

[0327] o Generates a digital 3D mesh based on point cloud data, which includes multiple polygons representing the surface of a dental object.

[0328] 59. A dental scanning system according to Clause 58, wherein generating a digital 3D mesh includes:

[0329] Geometric information is obtained based on one or more 2D images, wherein the geometric information indicates the rate of change of one or more surface properties in multiple surface portions of a dental object.

[0330] 60. A dental scanning system according to Clause 59, wherein generating a digital 3D mesh includes:

[0331] The mesh resolution for each surface portion is determined based on geometric information.

[0332] 61. A dental scanning system according to Clause 60, wherein generating a digital 3D mesh includes:

[0333] A mesh is generated for each surface portion, with the resolution determined based on the corresponding surface portion.

[0334] 62. A dental scanning system according to any one of clauses 59 to 61, wherein at least a portion of the geometric information associated with the corresponding surface portion is based on the point distribution of point cloud data corresponding to the surface portion.

[0335] 63. A dental scanning system according to any one of clauses 59 to 62, wherein at least a portion of the geometric information associated with the corresponding surface portion is based on ray tracing of light reflected at points in point cloud data corresponding to the surface portion.

[0336] 64. A dental scanning system pursuant to Clause 62 or 63, wherein the geometric information includes information indicating the rate of change of the contour of a surface portion.

[0337] 65. A dental scanning system according to any one of clauses 43 to 64, wherein at least a portion of the geometric information relating to a corresponding surface portion is based on color data corresponding to that surface portion.

[0338] 66. A dental scanning system according to Clause 65, wherein the geometric information includes information indicating the rate of color change of a surface portion.

[0339] 67. A dental scanning system according to any one of clauses 58 to 66, wherein obtaining one or more 2D images of a dental object includes obtaining multiple 2D images of the object.

[0340] 68. A dental scanning system according to Clause 67, wherein generating point cloud data includes: resolving a corresponding problem to determine which parts of a first image among a plurality of 2D images correspond to which parts of a second image among a plurality of 2D images.

[0341] 69. A dental scanning system according to clause 67 or 68, wherein generating point cloud data includes: registering points from a plurality of points obtained from each corresponding 2D image to a common coordinate system.

[0342] 70. A dental scanning system according to any one of clauses 43 to 69, wherein generating a digital 3D mesh comprises: assigning color values ​​to each polygon based on one or more 2D images.

[0343] 71. A dental scanning system according to any one of clauses 43 to 70, wherein obtaining one or more 2D images includes projecting light onto a dental object.

[0344] 72. A dental scanning system according to Clause 71, wherein light is projected in the form of a pattern.

[0345] 73. A dental scanning system according to Clause 72, wherein the projected pattern is static over time.

[0346] 74. A dental scanning system according to Clause 72, wherein the projected pattern changes over time.

[0347] 75. A dental scanning system according to any one of clauses 43 to 74, wherein generating point cloud data includes: performing a first transformation to generate a coordinate dataset of at least a portion of an object in the scanning device coordinate system.

[0348] 76. A dental scanning system according to Clause 75, wherein generating point cloud data comprises: performing a second transformation on a coordinate dataset of at least a portion of an object in the scanning device coordinate system to generate at least a portion of point cloud data in a real-world coordinate system.

[0349] 77. A dental scanning system according to any one of clauses 43 to 76, wherein generating a digital 3D mesh further includes:

[0350] An initial mesh is generated based on point cloud data, wherein the initial mesh includes multiple initial polygons.

[0351] 78. A dental scanning system pursuant to Clause 77, wherein the generation of a digital 3D mesh is based on an initial mesh.

[0352] 79. A dental scanning system according to clause 77 or 78, wherein obtaining geometric information of the surface portion includes:

[0353] Compare the normal vectors of adjacent initial polygons within the surface portion to obtain information about the rate of change of the surface portion's profile.

[0354] 80. A dental scanning system according to any one of clauses 43 to 76, wherein the processing device is further configured to:

[0355] • Generate 3D dental restoration data representing a dental restoration, which is configured to be placed on a tooth with an edge line, based on one or both of the vertices identified as representing edge lines in a digital 3D mesh and in a flattened mesh or in a digital 3D mesh.

[0356] 81. A dental scanning system according to Clause 80, wherein the system further comprises:

[0357] - A 3D printer or 3D milling machine, configured to manufacture dental prostheses based on 3D dental prosthesis data, which are configured to be placed on teeth with marginal lines.

[0358] The use of terms such as "first," "second," "third," and "fourth," "primary," "secondary," and "tertiary" does not imply any particular order, but is included to identify individual elements. Furthermore, the use of terms such as "first," "second," "third," and "fourth," "primary," "secondary," and "tertiary" does not indicate any order or importance, but is used to distinguish elements from each other. Please note that the terms "first," "second," "third," and "fourth," "primary," "secondary," and "tertiary" are used here and elsewhere for labelling purposes only and are not intended to indicate any specific spatial or temporal order. Additionally, the labeling of a first element does not imply the existence of a second element, and vice versa.

[0359] It is important to note that the word "including" does not necessarily exclude the existence of other elements or steps besides those listed. It is also important to note that the use of "a" or "one" before an element does not exclude the existence of multiple such elements.

[0360] Furthermore, it should be noted that any reference numerals do not limit the scope of the claims, exemplary embodiments may be implemented at least in part by hardware and software, and multiple “means,” “units,” or “devices” may be represented by the same hardware item.

[0361] It should be understood that references to "an embodiment," "embodiment," "aspect," or "may" in this specification refer to a specific feature, structure, or characteristic associated with that embodiment that is included in at least one embodiment of this disclosure. Furthermore, specific features, structures, or characteristics may be suitably combined in one or more embodiments of this disclosure. The foregoing description is intended to enable those skilled in the art to practice the various aspects described herein. Various modifications to these aspects will be apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects.

[0362] While embodiments and features have been shown and described, it should be understood that they are not intended to limit the claimed invention, and those skilled in the art will understand that various changes and modifications can be made without departing from the spirit and scope of the claimed invention. Therefore, the specification and drawings should be considered illustrative rather than restrictive. The claimed invention is intended to cover all alternatives, modifications, and equivalents.

[0363] Figure Labels

[0364]

Claims

1. A computer-implemented method for flattening a digital 3D mesh representing a dental object, the method comprising: • Obtain the digital 3D mesh; • Determine the circumferential mesh boundary in the digital 3D mesh, wherein the mesh boundary includes a plurality of boundary vertices surrounding a plurality of internal vertices; • Determine the target shape having a circumferential target boundary, wherein the target shape is planar; • For each of the boundary vertices, perform a consistent mapping to the anchor point corresponding to the corresponding vertex on the target boundary; • For each of the boundary vertices, translate the corresponding boundary vertex to its corresponding anchor point on the target boundary; and • Flatten the internal vertices to the target shape based on the anchor points to generate a flattened mesh.

2. The method according to claim 1, wherein, The consistency mapping satisfies the order preservation criterion, such that the order of the anchor points around the target boundary is the same as the order of the corresponding boundary vertices around the mesh boundary.

3. The method according to claim 1 or 2, wherein, The consistency mapping satisfies the uniqueness criterion, such that each unique boundary vertex is mapped to a unique anchor point.

4. The method according to any one of the preceding claims, wherein, The consistency mapping includes: o Select the initial vertex of the boundary vertex and the initial anchor point on the target boundary, and o maps the initial vertex to the initial anchor point.

5. The method according to claim 4, wherein, The consistency mapping includes: For each vertex in the boundary vertices other than the initial vertex, determine the relative distance between the corresponding vertex and the initial vertex along the mesh boundary, and For each vertex in the boundary vertices other than the initial vertex, map the corresponding vertex to an anchor point at a distance along the target boundary from the initial anchor point that is the same as the relative distance between the corresponding vertex and the initial vertex.

6. The method according to any one of the preceding claims, wherein, Determining the circumferential mesh boundary in the digital 3D mesh includes: o Roughly identify the edge lines of the dental object in the digital 3D mesh, and o Set the circumferential grid boundary around the roughly identified edge lines.

7. The method according to claim 6, wherein, A buffer distance is provided between the grid boundary and the roughly identified edge line.

8. The method according to any one of the preceding claims, wherein, The method further includes: • Identify the edge line of the dental object in the flattened grid / the edge line.

9. The method according to claim 8, wherein, The method further includes: • Based on the identified edge lines in the flattened grid, identify the vertices or surface points in the flattened grid that represent the edge lines.

10. The method according to claim 8 or 9, wherein, The method further includes: • Identify the edge lines in the 3D mesh based on the edge lines identified in the flattened mesh.

11. The method according to claim 10, wherein, The method further includes: • Based on the edge lines identified in the 3D mesh, identify the vertices or surface points in the digital 3D mesh that represent the edge lines.

12. The method according to the preceding claims and any one of claims 9 and / or 11, wherein, The method further includes: • Based on the digital 3D mesh and one or both of the vertices or surface points in the flattened mesh that are identified as representing the edge line, or the vertices or surface points in the digital 3D mesh that are identified as representing the edge line, 3D dental restoration data representing the dental restoration is generated, the dental restoration being configured to be placed on the edge line.

13. A dental system comprising: - A processing device configured to perform the method according to any one of claims 1 to 12.

14. The dental system of claim 13, further comprising: - An intraoral scanner, configured to acquire one or more 2D images of the dental object; The processing device is further configured to generate the digital 3D mesh based on the one or more 2D images, the digital 3D mesh comprising a plurality of polygons representing the surface of the dental object.

15. The dental system of claim 13 or 14, further comprising: - A 3D printer or 3D milling machine configured to manufacture a dental prosthesis based on the 3D dental prosthesis data, the dental prosthesis being configured to be placed on a tooth having the marginal line.

Images