A method for fully automatic processing of prefabricated pressed film retainer based on intelligent design

Abstract

The present invention relates to the field of orthodontic retainer manufacturing, and specifically to a method for prefabricated pressed film retainers based on intelligent design and full-automatic processing, comprising the following steps: step S1: intraoral scanning before bracket removal to obtain dentition three-dimensional data; step S2: based on the dentition three-dimensional data, protrusions on the tooth surface are identified, the protrusions are automatically removed and optimized, and virtual dentition model data with tooth surface feature objects removed is generated; step S3: crown neck edge information is automatically identified, the order of crown positions is determined, and undercuts are filled in the gaps, a continuous retainer edge line is automatically generated, and a cutting direction, depth and path are calculated; step S4: a dentition entity model is manufactured by three-dimensional printing technology, a vacuum pressed film retainer is automatically manufactured, and the retainer is automatically cut according to the calculated cutting path to complete the manufacturing of the retainer, and the whole process of intelligent digital design and manufacturing is adopted, which can greatly reduce the overall cost and reduce the economic burden on patients.

Description

Technical Field

[0001] The invention relates to the field of orthodontic retainer manufacturing, and in particular to a method for intelligently designing and fully automatically processing a prefabricated pressed film retainer. Background Art

[0002] Orthodontic retention is an indispensable and important part of the correction process. The pressed film retainer has a beautiful appearance and a small size. It is currently the most widely used. After the doctor and the patient confirm that the orthodontic treatment can be completed, the brackets or accessories must be removed first to fully expose the tooth surface; then the impression is taken and transferred to the denture processing center, where the retainer is made; finally, it is worn in the patient's mouth in the clinic to maintain the tooth position. Since the patient's dentition begins to relapse after the archwire is removed, and the early relapse is obvious, in order not to affect the correction effect, the retainer must be made as soon as possible. However, the retainer production process is long and cumbersome, including mold taking, model pouring, filling concave, pressed film production, cutting and shaping, plus the round-trip transportation time, ideally it takes at least half a day. In most cases, the clinic needs to remove the intraoral brackets or accessories before ten in the morning, and the patient can only wear the retainer after four in the afternoon. The waiting time in the middle is too long, which greatly affects the treatment experience, delays the patient's daily work and life, and consumes a lot of time and energy. Sometimes the impression may be deformed and the retainer cannot be worn, which not only requires the patient to visit the doctor again, but also causes relapse of dentition due to the inability to wear the retainer in time for a long time. In addition, the traditional manual process of making retainers is greatly affected by human factors, and may result in manufacturing defects, and it is impossible to ensure that all retainers are made to the same level. In addition, the traditional process is inefficient and cumbersome, resulting in high production costs. Orthodontic patients are mostly students or working young people, and their consultation time is overly concentrated on holidays. The traditional process puts great pressure on outpatient clinics and medical care.

[0003] In view of this, the present invention provides a method for preparing a prefabricated pressed film retainer based on intelligent design and full-automatic processing. Summary of the invention

[0004] The object of the present invention is to provide a method for fully automatically processing a prefabricated pressed film retainer based on intelligent design, so as to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] In a first aspect, the present invention provides a method for fully automatically processing a prefabricated pressed film retainer based on intelligent design, comprising the following steps:

[0007] Step S1: intraoral scanning before bracket removal to obtain 3D data of the dentition;

[0008] Step S2: based on the three-dimensional data of the dentition, protrusions on the tooth surface are identified, the protrusions are automatically removed and optimized, and virtual dentition model data with the protrusions on the tooth surface removed is generated;

[0009] Step S3: automatically identifying crown neck margin information, determining the order of crown positions, filling the gaps with undercuts, automatically generating a continuous retainer edge line, and calculating cutting direction, depth, and path;

[0010] Step S4: a dentition solid model is produced by three-dimensional printing technology, a retainer is produced by automated vacuum lamination, and the retainer is produced by automatic cutting according to the calculated cutting path.

[0011] As a preferred embodiment of the present invention, in step S1, the logic for acquiring the dentition three-dimensional data is:

[0012] Collect tooth surface conditions based on oral scanning instruments, and identify scan data defects based on tooth surface conditions;

[0013] The defects of the scanned data are repaired, and the shell data of the tooth surface condition is closed to generate three-dimensional data of the dentition, which includes three-dimensional image information and three-dimensional point cloud data.

[0014] As a preferred embodiment of the present invention, in step S2, the implementation logic of the virtual dentition model data is:

[0015] Step S21: firstly, obtaining three-dimensional point cloud data of the oral cavity from an oral scanning instrument, and constructing three-dimensional surface data by Poisson reconstruction based on the point cloud data, wherein the three-dimensional surface data includes tooth surface data, gum surface data and protrusion surface data;

[0016] Step S22: preliminarily label the three-dimensional face data through a deep learning neural network and assign a probability of belonging to a category to each labeled face;

[0017] Step S23: determining adjacent faces between different faces based on the three-dimensional face data divided in step S22, assigning correlations to virtual edges between adjacent faces, and then segmenting the three-dimensional image information with an image segmentation algorithm to obtain segmentation lines;

[0018] Step S24: generating independent and separated tooth and gum surface models according to the segmentation lines, and superimposing the tooth and gum surface models to generate virtual dentition model data.

[0019] As a preferred embodiment of the present invention, the image segmentation algorithm realizes stereo matching and point denoising to obtain segmentation lines through minimum cut;

[0020] Divide the image into several non-overlapping regions through minimum cut;

[0021] Then use the minimized energy function to find the optimal solution, where: the energy function usually includes data terms and smoothing terms;

[0022] After determining the energy function, the three-dimensional model is segmented by maximum flow and minimum cut to extract the tooth surface and gum surface.

[0023] As a preferred embodiment of the present invention, in step S3, the logic for automatically identifying the crown neck margin information is:

[0024] Step S31: obtaining the intersection line of the tooth surface and the gum surface based on the virtual dentition model, determining the intersection of triangles by using the Devillers algorithm, and obtaining the intersection of each independent tooth surface and the surrounding gum surface;

[0025] Step S32: Based on the intersection of the tooth patch and the gum patch corresponding to each independent tooth in step S31, the intersection of two triangles is calculated by a spatial geometric point projection algorithm;

[0026] Step S33: Based on the intersection point in step S32, the centroid method is used to determine whether the point is located in the triangle to obtain the intersecting gear circle point, and based on the integration of the intersecting gear circle points, the intersecting gear circle point set of each tooth and gum is obtained;

[0027] Step S34: Based on the geometric features of the tooth ring line, the intersecting tooth ring point set in step S33 is projected onto the base plane, and then the crown neck margin line is generated by continuing to sort through the plane rolling method.

[0028] As a preferred embodiment of the present invention, the logic for obtaining the crown neck margin line is:

[0029] Step A1: Determine the initial point K based on the concave hull;

[0030] Step A2: Starting from the initial point K, find the first chord;

[0031] Step A3: Loop to find the next chord and gradually build a complete curve;

[0032] Step A4: Find all chords in sequence until no new chord is found or a point that has been used as a chord is encountered;

[0033] Step A5: sorting the tooth ring lines of each tooth in the order according to a unified coordinate system, and dividing each tooth ring line into inner and outer circles according to the adjacent points between the tooth ring lines;

[0034] Step A6: Provide the user with smoothness settings to adjust the transition smoothness of adjacent tooth ring lines on the same side. Using the median of the angles of adjacent points and the distance from the tooth ring segmentation point as indicators, generate a closed curve from the inside to the outside that fits the crown neck edge based on the sorted intersecting tooth ring point set, and mark the closed curve as the crown neck edge line.

[0035] As a preferred embodiment of the present invention, in step S3, the processing method of filling the undercut at the gap includes:

[0036] Based on the tooth ring line obtained, the three-dimensional points at the gap between two adjacent teeth are intercepted, and these three-dimensional points are offset by a preset distance along the center direction of their respective tooth rings, and then projected onto the tooth surface along the planned direction through a spatial point projection algorithm;

[0037] The obtained projection points are retracted inward by a preset distance along the normal direction, and a patch is constructed based on the positions of the projection points.

[0038] As a preferred embodiment of the present invention, in step S3, the processing logic of the cutting direction, depth and path includes:

[0039] Based on the spatial partitioning data structure, a KD tree is constructed for the three-dimensional point cloud data. The data points in the multi-dimensional space are obtained based on the KD tree. The space is divided into different regions through recursion. The average normal of the surrounding face of each neighbor point is obtained through the K nearest neighbor algorithm to calculate the cutting direction.

[0040] Use one-dimensional Gaussian blur function to perform weighted averaging of neighboring points to determine the cutting line position;

[0041] The outer product of the cutting direction and the direction of the line connecting the adjacent points is calculated to obtain the point offset direction, and the point offset is gradually offset along the point offset direction through the smaller step length of the adjacent points to determine the offset distance, and the offset distance is marked as the tool radius compensation distance;

[0042] Use the tool model for CNC simulation, fine-tune the cutting direction, obtain the cutting path based on the cutting direction, cutting line point and tool radius compensation distance, and automatically cut to complete the production of the retainer.

[0043] In a second aspect, the present invention provides an electronic device, comprising:

[0044] at least one processor; and,

[0045] A memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor so that the at least one processor can execute the method described in the first aspect.

[0046] In a third aspect, the present invention provides a computer-readable storage medium storing instructions, which, when executed on a computer, causes the computer to execute the method described in the first aspect.

[0047] Compared with the prior art, the present invention has the following beneficial effects:

[0048] The oral correction treatment of the present invention is completed directly in the outpatient clinic. By changing the timing of retainer production, no additional chairside equipment and operators are required, which improves design efficiency, reduces human errors, ensures good accuracy, reduces manpower allocation, and saves labor costs; at the same time, the retainer can be removed and put on immediately, reducing patient waiting time, reducing outpatient pressure, and improving retention effect. Since this method adopts full-process intelligent digital design and production, it can greatly reduce the overall cost and reduce the economic burden on patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 A flow chart of the method for the present invention based on intelligent design and fully automatic processing of a prefabricated pressed film retainer;

[0050] Figure 2 A three-dimensional image of the dentition scanned in the mouth before the bracket is removed according to the present invention;

[0051] Figure 3 It is a defect map of the scan data for identifying the tooth surface condition of the present invention;

[0052] Figure 4 Generating a three-dimensional data map of dentition for the closed case data of the present invention;

[0053] Figure 5 Identify and remove protrusions on the tooth surface for the present invention;

[0054] Figure 6 The invention automatically identifies the crown and neck margin information map;

[0055] Figure 7 Automatically generate edge line diagram for filling undercuts of the present invention;

[0056] Figure 8 This is a schematic diagram of the simulation of the intersection of two triangle corresponding planes of the present invention;

[0057] Fig. 9 It is a diagram of a retainer pressed by a dentition physical model of the present invention;

[0058] Fig.10 This is a completed diagram of the retainer of the present invention. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] In the description of the present invention, it should be noted that the terms "vertical", "up", "down", "horizontal", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0061] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0062] Example 1: Please refer to Figure 1 The present invention provides a technical solution: a method for prefabricated film retainer based on intelligent design and full-automatic processing, comprising the following steps:

[0063] Step S1: intraoral scanning before bracket removal to obtain 3D data of the dentition;

[0064] It should be noted that during orthodontic treatment, bracket removal includes the steps of removing tooth surface brackets, archwires and accessories. During use, patients apply force through tooth surface brackets, archwires and accessories to gradually move teeth to the correct position. Before removal, the position of the teeth is fixed, and after removal, the tooth surface is fully exposed. After removal, the retainer needs to maintain and shape the tooth surface to avoid early recurrence and affect the correction effect. Based on this situation, intraoral scanning is performed before bracket removal to obtain a more accurate tooth surface condition, and the tooth surface condition is characterized by 3D data of the dentition. The tooth surface condition includes information such as the dentition, surrounding oral soft tissue, orthodontic brackets, archwires, accessories and overflow adhesive.

[0065] The logic of obtaining the three-dimensional data of the dentition:

[0066] Based on the oral scanning instrument to collect the tooth surface conditions, Figure 2 As shown, the defects of the scan data are identified based on the tooth surface condition. Figure 3As shown, the scan data defects include tooth surfaces having non-popular edges, self-intersecting areas, highly refractive edges, spikes, small components, small channels and small holes;

[0067] Repair the scan data defects and seal the case data of the tooth surface. Figure 4 As shown, three-dimensional dentition data is generated, and the three-dimensional dentition data includes three-dimensional image information and three-dimensional point cloud data.

[0068] It should be noted that the tooth surface conditions collected by oral scanning instruments are generally image data. During the process of automatic recognition and analysis by the system, repair can be performed from the direction of image processing. Various scanning data defects may appear in the tooth surface conditions during the scanning process, and these scanning data defects will affect the authenticity judgment of the tooth surface conditions. Therefore, along the physical position data corresponding to the scanning data defects, analysis is performed from the direction of position data to achieve repair; the scanned case data is reconstructed through Poisson's equation and Marching Cubes and other algorithms to generate three-dimensional data of the dentition; and according to user needs, a base of appropriate height is generated to facilitate subsequent printing, lamination and other process operations.

[0069] Step S2: Based on the three-dimensional data of the dentition, the protrusions on the tooth surface are identified, the protrusions are automatically removed and optimized, and virtual dentition model data with the tooth surface features removed is generated. Figure 5 As shown;

[0070] It should be noted that protrusions are attachments, brackets and other tooth surface attachments required for orthodontic treatment. According to the curvature and morphological characteristics of the tooth surface, protrusions on the tooth surface are automatically identified and removed through definition or feature learning, thereby optimizing the tooth surface and generating a smoother and more accurate virtual dentition model data.

[0071] Specifically, the virtual dentition model data includes N tooth patches and 1 gum patch; the implementation logic of the virtual dentition model data is:

[0072] Step S21: firstly, obtaining three-dimensional point cloud data of the oral cavity from an oral scanning instrument, and constructing three-dimensional surface data by Poisson reconstruction based on the point cloud data, wherein the three-dimensional surface data includes tooth surface data, gum surface data and protrusion surface data;

[0073] It should be noted that the oral scanning instrument obtains the tooth surface condition of the oral cavity, and the three-dimensional point cloud data is the tooth surface condition including protrusions. Poisson reconstruction uses the Poisson surface reconstruction algorithm to convert point cloud data into three-dimensional patch data. The Poisson surface reconstruction algorithm contains an indicator function to process point cloud data with noise and missing data; three-dimensional patch data is to construct the tooth surface condition into three-dimensional patch data, so that the corresponding features of teeth, gums and protrusions can be respectively entered into the deep learning neural network, and the three-dimensional patch data can be analyzed based on the deep learning neural network.

[0074] Step S22: preliminarily label the three-dimensional face data through a deep learning neural network and assign a probability of belonging to a category to each labeled face;

[0075] It should be noted that: based on the deep learning neural network, the three-dimensional surface data is divided into tooth surface data, gum surface data and protrusion surface data, and the difference points of the tooth surface conditions are expanded based on the assignment method.

[0076] Step S23: based on the three-dimensional patch data divided in step S22, adjacent faces between different patches are determined, and correlation is assigned to virtual edges between adjacent faces. Then, the three-dimensional image information is segmented to obtain segmentation lines in conjunction with an image segmentation algorithm; stereo matching and point denoising are implemented through minimum cut to form an accurate segmentation line;

[0077] It should be noted that the assignment information of the three-dimensional patch data is combined with the three-dimensional image information, and the patches are mapped to the nodes on the three-dimensional image information, indicating the adjacency relationship between the patches. The virtual edge assignment indicates the correlation between whether two patches are segmented in the same area.

[0078] Image segmentation algorithm is a problem-solving method for image segmentation, energy minimization and graph optimization. It uses graph structure to express the problem and forms an accurate segmentation line through stereo matching and point denoising through minimum cut.

[0079] The image is divided into several non-overlapping areas through the minimum cut, and then the optimal solution is found by minimizing the energy function. The energy function usually includes a data term (data fitting term) and a smoothing term (regularization term). The data term measures the degree of fit of the model to the data, and the smoothing term encourages the smoothness of the solution. That is, the problem is expressed using the graph structure and solved by the minimum cut or maximum flow method.

[0080] The energy function is specifically: ;

[0081] Where: Represents the energy function data items, Representation patch Tags, Representation patch Assigned tags The cost function of Represents the relationship between labels and encourages adjacent faces to have the same labels, thereby achieving smooth boundaries. and represents all adjacent face pairs, Representation patch and The cost function for assigning different labels between them.

[0082] After determining the energy function, the three-dimensional model is segmented by the maximum flow minimum cut, and its expression is: ; Where: Indicates the maximum flow, from the source point To the meeting point The total flow is: . Define the cut ,in: and yes A division of , Cutting capacity is from arrive The sum of the capacities of all edges of ,The graph cut algorithm can be used to more finely divide the tooth patch and gum patch, compensating for the defect of messy edges of deep learning patch marking.

[0083] Step S24: generating independently separated tooth patches and gum patches according to the segmentation line, and superimposing the tooth patches and gum patches to generate virtual dentition model data;

[0084] Tooth and gingival segmentation through deep learning network and graph cut algorithm can reduce the complexity of the deep model and the processing time of the actual segmentation. It also has good stability and adaptability. It can automatically identify protrusions on the tooth surface through definition or feature learning according to the curvature and morphological characteristics of the tooth surface, automatically remove and optimize them, and generate virtual dentition model data with the tooth surface features removed.

[0085] Step S3: Automatically identify the crown edge information and determine the positioning method, such as Figure 6 As shown in the figure, the gap is filled with undercuts, a continuous retainer edge line is automatically generated, and the cutting direction, depth and path are calculated, as shown in the figure. Figure 7 shown.

[0086] It should be noted that the tooth patch and gum patch in the virtual dentition model data in step S2 are both composed of a group of triangular patches. In order to find the intersection line between each tooth patch and the gum patch, it is necessary to find the intersection point of each tooth patch and the gum patch respectively, sort all the intersection points on each tooth patch into tooth ring lines, and characterize the crown neck margin information based on the tooth ring lines.

[0087] The logic for automatically identifying the crown edge information is as follows:

[0088] Step S31: obtaining the intersection line of the tooth surface and the gum surface based on the virtual dentition model, determining the intersection of triangles by using the Devillers algorithm, and obtaining the intersection of each independent tooth surface and the surrounding gum surface;

[0089] Specifically:

[0090] The devillers algorithm uses the positive and negative geometric meanings of the determinant formed by the vertices of the triangle to determine the relative position relationship between the midpoints, lines, and planes of the triangle, thereby determining whether the two triangles intersect. If they intersect, continue, and if they do not intersect, skip directly. The basic principle is as follows:

[0091] Get four points in space marked as , , and ; Define the determinant as follows:

[0092] ;

[0093] The right-hand screw rule is used to define the positional relationship of the four spatial points. Indicates the direction of the positive normal of the triangle formed by d in the counterclockwise order (i.e. upward); Indicates that d is in the negative normal direction of △abc (i.e., downward); Indicates that the four points are coplanar.

[0094] Let two triangles and ,triangle The vertices of are: , and ;triangle Vertex: , and ;triangle The plane is and triangle The plane is , The plane normal vector is , The plane normal vector is ,like Figure 8 shown.

[0095] The triangle Vertices and triangles Any vertex of a triangle constitutes four points in space. Based on the above Devillers algorithm, the relative position relationship between the triangle and the plane where another triangle is located is determined, and non-intersecting situations are eliminated in advance.

[0096] Specifically, by calculating To determine the triangle and Relationship in the plane: If all determinants are nonzero and have the same sign, then the triangle and triangle Do not intersect; otherwise the triangle and The planes intersect. Similarly, find the triangle and Intersection of planes.

[0097] It should be noted that: based on the above algorithm, the intersection of each triangular facet of the tooth with the surrounding gum facets is calculated in turn. If they intersect, the process continues; if they do not intersect, the process is skipped.

[0098] Step S32: Based on the intersection of the tooth patch and the gum patch corresponding to each independent tooth in step S31, the intersection of two triangles is calculated by a spatial geometric point projection algorithm;

[0099] Specifically, the spatial geometric point projection algorithm includes: assuming that the spatial plane equation is , where A, B, C and D are the parameters corresponding to the current space plane equation;

[0100] Assume that the coordinates of the three-dimensional space point not on the plane are , the coordinates of its projection point on the plane are . Because the projection point to the current point is perpendicular to the plane, according to the vertical constraint, it is easy to know and The following conditions:

[0101] ;

[0102] ;

[0103] Among them, the coordinates of the projection point can be obtained according to the above two formulas:

[0104] ;

[0105] ;

[0106] Step S33: Based on the intersection point in step S32, the centroid method is used to determine whether the point is located in the triangle to obtain the intersecting gear circle point, and based on the integration of the intersecting gear circle points, the intersecting gear circle point set of each tooth and gum is obtained;

[0107] It should be noted that the intersection of the two triangles obtained in step S32 is not necessarily within the scope of the two triangles. Therefore, the centroid method is used to determine whether the intersection obtained in step S32 is within the scope of the triangle. Through the above three steps, all the intersecting tooth circle points of a group of triangular facets of teeth and a group of triangular facets of gums can be obtained, and the intersecting tooth circle point set of each tooth and gum can be obtained based on the integration of the intersecting tooth circle points.

[0108] Specifically, the steps of the centroid method include: For any point P in the plane, it can be expressed by the following equation: (1)

[0109] in, and represents a vector, and represents a scalar;

[0110] remember , , , then substitute into equation (1) and solve:

[0111] ;

[0112] ;

[0113] when , and When , it indicates that the point is inside the triangle, and the intersection point is marked as the intersection gear circle point; if it is not inside the two triangle face patches, the intersection point is discarded.

[0114] Step S34: projecting the intersecting tooth ring point set in step S33 onto the base plane based on the geometric features of the tooth ring line, and then continuing to sort by plane rolling method to generate the crown neck margin line;

[0115] It should be noted that the intersection points in the intersecting tooth ring point set obtained in step S31 to step S33 are in disorder. If such a crown neck margin line needs to be obtained, the intersection points are sorted by plane rolling method. The plane rolling method simplifies the calculation and is more stable than directly sorting the three-dimensional points.

[0116] Specifically, the algorithm steps of the plane rolling method include:

[0117] Step A1: Determine the initial point K based on the concave hull;

[0118] Specifically, the initial point K is the smallest point corresponding to the y-axis coordinate or the largest point corresponding to the x-axis coordinate in the three-dimensional patch data;

[0119] Step A2: Starting from the initial point K, find the first chord;

[0120] Specifically, take (1,0) as the reference vector and find an edge smaller than R as the initial edge. The other endpoint of the initial edge is point L. At this time, a circle with a radius of R / 2 is stuck on the initial edge KL. The initial edge is marked as the first chord, that is, point L is sorted after point K.

[0121] Step A3: Loop to find the next chord.

[0122] Specifically, if the previous chord is KL, then the next chord must start from L and connect to a point E in the R region of L. The following principles can be used to find point E:

[0123] First, sort the points in the R domain of L in polar coordinates based on the LE vector centered at L, and then sort the points E0 to E0 in the R domain. N Build with LE i The circle of the chord, and then check whether it contains other E0~E N If the point does not exist, then LE i If it is a new string, the loop will be exited.

[0124] Step A4: Find all chords in sequence until no new chord is found or a point that has previously been a chord is encountered.

[0125] Step A5: sorting the tooth ring lines of each tooth in the order according to a unified coordinate system (such as clockwise or counterclockwise), and dividing each tooth ring line into inner and outer circles according to the adjacent points between the tooth ring lines;

[0126] Step A6: Provide the user with smoothness settings to adjust the transition smoothness of adjacent tooth ring lines on the same side, using the median of the angle between adjacent points and the distance between the tooth ring segmentation points as indicators. Based on this, a closed curve from the inside to the outside that fits the crown neck edge is generated based on the sorted tooth ring line point set, and the closed curve is marked as the crown neck edge line.

[0127] On the other hand, in step S3, the processing method of filling the undercut at the gap includes:

[0128] Based on the tooth ring line obtained, three-dimensional points within a certain range of the gap between two adjacent teeth are intercepted;

[0129] After these points are offset along the center direction of their respective tooth rings for a certain distance, they are then projected onto the tooth surface along the planned direction using a spatial point projection algorithm;

[0130] The obtained projection points are retracted a certain distance inward along the normal direction, and finally the surface is constructed based on these points.

[0131] It should be noted that: by constructing undercuts through the above method, relying on the tooth cervical margin line, the gaps between teeth can be accurately filled; and by adjusting the intercepted point range and the planned projection direction, undercuts of different shapes and fullness can be easily defined to meet the needs of filling various types of tooth undercuts.

[0132] After the crown neck margin line is downsampled, the NURBS algorithm is used to generate a coherent cutting line based on the retained feature points. NURBS (Non-Uniform Rational B-Splines) is a mathematical model used to represent and generate curves and surfaces. The calculation formula is as follows.

[0133] ;

[0134] Where: refers to the interpolation point of the Nurbs curve, is a parameter variable, It is control points, It is The weight of the control point, It is B-spline basis functions of order p.

[0135] On the other hand, in step S3, the calculation logic of the cutting direction, depth and path is:

[0136] Based on the spatial partitioning data structure, a KD tree is constructed for the three-dimensional point cloud data. The data points in the multidimensional space are obtained based on the KD tree. The space is divided into different areas through recursion. The average normal of the surrounding patches of each neighbor point is obtained through the K nearest neighbor algorithm, and the cutting direction is calculated.

[0137] The KD tree construction process is as follows:

[0138] 1) Select the split dimension: alternately select by coordinate axis (x, y, z);

[0139] 2) Select the split point: select the median of the current dimension as the split point;

[0140] 3) Recursively construct subtrees: Divide the data points into two parts, the left subtree contains points less than the median, and the right subtree contains points greater than or equal to the median, and then recursively process each subtree.

[0141] After constructing the KD tree, you can perform a fast K-nearest neighbor search. The steps include:

[0142] 1) Select the split axis: Select the split axis according to the current depth.

[0143] 2) Calculate the distance between the current point and the splitting surface: Determine whether the point to be found is on one side of the splitting surface and decide whether to recurse to the other side.

[0144] 3) Recursive search for nearest neighbors: Recursively search for K nearest neighbors in the subtree until the entire tree is traversed or enough nearest neighbors are found;

[0145] All neighboring points are determined as cutting line points through Gaussian blur smoothing, and then the cutting line points are arranged in sequence and connected.

[0146] Using one-dimensional Gaussian blur can avoid excessive local changes in the cutting direction. The calculation formula is:

[0147] ;

[0148] in, yes The mean of yes The variance of is the cutting line point coordinate;

[0149] The outer product of the cutting direction and the direction of the line connecting the adjacent points is calculated to obtain the point offset direction. The offset distance is determined by gradually offsetting the adjacent points with smaller steps along the point offset direction, and the offset distance is marked as the tool radius compensation distance.

[0150] After obtaining the initial cutting point and cutting direction, the cutting direction is fine-tuned through CNC simulation using the tool model to avoid physical interference during the actual cutting process. The retainer is automatically cut according to the calculated cutting path to complete the production.

[0151] Step S4: Make a dentition physical model by 3D printing technology, such as Fig. 9 As shown, the retainer is automatically made by vacuum lamination, and the retainer is automatically cut according to the calculated cutting path, such as Fig.10 shown.

[0152] After the above steps are completed, through highly automated processes and technical means, the process from data acquisition to product processing is fast, accurate and efficient. The retainer is made in advance. After all processing steps are completed, the patient will be scheduled to go to the clinic to remove the brackets, archwires or accessories, and immediately try on the retainer. This "remove and wear" method greatly improves the convenience and comfort of patients, improves the patient's treatment experience, and reduces the workload of medical staff.

[0153] Embodiment 2: The parts not described in this embodiment are the same as those in Embodiment 1. This embodiment shows an electronic device, including:

[0154] at least one processor; and,

[0155] A memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor so that the at least one processor can execute a method based on intelligent design and fully automatic processing of a prefabricated die-casting retainer as described in the first aspect.

[0156] The electronic device may have relatively large differences due to different configurations or performances, and may include one or more processors (Central Processing Units, CPU) and one or more memories, wherein the memories store at least one computer program, and the at least one computer program is loaded and executed by the processor to implement a method based on intelligent design and fully automatic processing of a prefabricated film retainer provided by the above-mentioned various method embodiments.

[0157] The electronic device may also include other components for realizing the functions of the device, for example, the electronic device may also have components such as a wired or wireless network interface and an input / output interface for input and output. The embodiments of the present application will not be described in detail here.

[0158] Example 3: The parts not described in this example are as in Example 1. This example also provides a computer-readable storage medium storing instructions, which, when executed on a computer, enable the computer to execute the method of intelligent design and fully automatic processing of prefabricated die-casting retainers.

[0159] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0160] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0161] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for fully automatic processing of prefabricated film retainers based on intelligent design, characterized in that: The following steps are involved: Step S1: intraoral scanning before bracket removal to obtain 3D data of the dentition; Step S2: based on the three-dimensional data of the dentition, protrusions on the tooth surface are identified, the protrusions are automatically removed and optimized, and virtual dentition model data with the protrusions on the tooth surface removed is generated; In step S2, the implementation logic of the virtual dentition model data is: Step S21: firstly, obtaining three-dimensional point cloud data of the oral cavity from an oral scanning instrument, and constructing three-dimensional surface data by Poisson reconstruction based on the point cloud data, wherein the three-dimensional surface data includes tooth surface data, gum surface data and protrusion surface data; Step S22: preliminarily label the three-dimensional face data through a deep learning neural network and assign a probability of belonging to a category to each labeled face; Step S23: determining adjacent faces between different faces based on the three-dimensional face data divided in step S22, assigning correlations to virtual edges between adjacent faces, and then segmenting the three-dimensional image information with an image segmentation algorithm to obtain segmentation lines; Step S24: generating independent and separated tooth and gum surface models according to the segmentation line, and superimposing the tooth and gum surface models to generate virtual dentition model data; Step S3: automatically identifying crown neck margin information, the order of crown positions, filling the gaps with undercuts, automatically generating a continuous retainer edge line, and calculating cutting direction, depth, and path; In step S3, the logic for automatically identifying the crown neck margin information is as follows: Step S31: obtaining the intersection line of the tooth surface and the gum surface based on the virtual dentition model, determining the intersection of triangles by using the Devillers algorithm, and obtaining the intersection of each independent tooth surface and the surrounding gum surface; Step S32: Based on the intersection of the tooth patch and the gum patch corresponding to each independent tooth in step S31, the intersection of two triangles is calculated by a spatial geometric point projection algorithm; Step S33: Based on the intersection point in step S32, the centroid method is used to determine whether the point is located in the triangle to obtain the intersecting gear circle point, and based on the integration of the intersecting gear circle points, the intersecting gear circle point set of each tooth and gum is obtained; Step S34: projecting the intersecting tooth ring point set in step S33 onto the base plane based on the geometric features of the tooth ring line, and then continuing to sort by plane rolling method to generate the crown neck margin line; In step S3, the processing logic for cutting direction, depth and path includes: Based on the spatial partitioning data structure, a KD tree is constructed for the three-dimensional point cloud data. The data points in the multi-dimensional space are obtained based on the KD tree. The space is divided into different regions through recursion. The average normal of the surrounding face of each neighbor point is obtained through the K nearest neighbor algorithm to calculate the cutting direction. Use one-dimensional Gaussian fuzzy function to perform weighted averaging of neighboring points to determine the cutting line point position; The outer product of the cutting direction and the direction of the line connecting the adjacent points is calculated to obtain the point offset direction, and the point offset is gradually offset along the point offset direction through the smaller step length of the adjacent points to determine the offset distance, and the offset distance is marked as the tool radius compensation distance; Use the tool model to perform CNC simulation, fine-tune the cutting direction, obtain the cutting path based on the cutting direction, cutting line point and tool radius compensation distance, and automatically cut to complete the production of the retainer; Step S4: a dentition solid model is produced by three-dimensional printing technology, a retainer is produced by automated vacuum lamination, and the retainer is produced by automatic cutting according to the calculated cutting path.

2. A method for intelligently designing and fully automatically processing a prefabricated pressed film retainer according to claim 1, characterized in that: In step S1, the logic for acquiring the three-dimensional data of the dentition is: Collect tooth surface conditions based on oral scanning instruments, and identify scan data defects based on tooth surface conditions; The defects of the scanned data are repaired, and the shell data of the tooth surface condition is closed to generate three-dimensional data of the dentition, which includes three-dimensional image information and three-dimensional point cloud data.

3. A method for intelligently designing and fully automatically processing a prefabricated pressed film retainer according to claim 1, characterized in that: The image segmentation algorithm realizes stereo matching and point denoising to obtain segmentation lines through minimum cut; Divide the image into several non-overlapping regions through minimum cut; Then use the minimized energy function to find the optimal solution, where: the energy function includes data terms and smoothing terms; After determining the energy function, the three-dimensional model is segmented by maximum flow and minimum cut to extract the tooth surface and gum surface.

4. A method for intelligently designing and fully automatically processing a prefabricated pressed film retainer according to claim 1, characterized in that: The logic of obtaining the crown neck margin line: Step A1: Determine the initial point K based on the concave hull; Step A2: Starting from the initial point K, find the first chord; Step A3: Loop to find the next chord and gradually build a complete curve; Step A4: Find all chords in sequence until no new chord is found or a point that has been used as a chord is encountered; Step A5: sorting the tooth ring lines of each tooth in the order according to a unified coordinate system, and dividing each tooth ring line into inner and outer circles according to the adjacent points between the tooth ring lines; Step A6: Provide the user with smoothness settings to adjust the transition smoothness of adjacent tooth ring lines on the same side. Using the median of the angles of adjacent points and the distance from the tooth ring segmentation point as indicators, generate a closed curve from the inside to the outside that fits the crown neck edge based on the sorted intersecting tooth ring point set, and mark the closed curve as the crown neck edge line.

5. A method for intelligently designing and fully automatically processing a prefabricated pressed film retainer according to claim 1, characterized in that: In step S3, the processing method of filling the undercut at the gap includes: Based on the tooth ring line obtained, the three-dimensional points at the gap between two adjacent teeth are intercepted, and these three-dimensional points are offset by a preset distance along the center direction of their respective tooth rings, and then projected onto the tooth surface along the planned direction through a spatial point projection algorithm; The obtained projection points are retracted inward by a preset distance along the normal direction, and a patch is constructed based on the positions of the projection points.

Images