Patent Application
20230089649
The present disclosure relates to the technical field of orthodontics, and in particular, to a method for identifying a gum line of a tooth model, a device, and a storage medium.
Invisible Orthodontic: refers to invisible orthodontics without brackets, and belongs to one type of orthodontics.
With the development of scientific and technical information, technologies such as computer technology, manufacturing technology, digital modeling technology, material science, numerical control technology, etc. have rapidly developed and grew stronger, and these subjects merge with each other. In particular, the computer technology has been increasingly infiltrated into various aspects such as teaching, scientific research and clinical applications in various medical fields, and can achieve more progressive mutual cooperation. Moreover, with the development and popularity of measurement technology, people can conveniently acquire digitized tooth models, which plays an important role in the process of oral clinical diagnosis and treatment. The emergence and development of 3D printing technology has become one of the current hot topics, and applying the 3D printing technology to the medical field is also very common. 3D printing has been applied in the medical field for more than two decades, and is widely applied to operations such as oral planting, orthopaedics, and neurosurgery.
Dental treatment is an unavoidable problem for most people. With the improvement of computer software and hardware technologies and the appearance of increasingly precise tooth data, people also turn their attention from traditional pure manual dental treatment to digitized dental treatment such as tooth diagnosis and treatment in the invisible orthodontic, so as to improve the safety and success rate of dental treatment with the help of priori knowledge provided by advanced digital technology.
In the current application scenarios of tooth diagnosis and treatment in the invisible orthodontic, performing data processing on tooth models are all needed; however, data pre-processing of the tooth models occupies most of the time of 3D printing of tooth model. In particular, in application scenarios of orthodontics, in addition to requiring to perform data pre-processing including placement, hollowing and Boolean operations on a tooth model, after printing is completed, it is also necessary to perform film pressing, and after the film pressing, manual cutting is required, or a gum line is manually drawn to supply to a CNC machine for cutting. This process needs to consume significant labor costs and significant workload. In order to increase the efficiency of the whole diagnosis and treatment process and enhance the user experience, it is necessary to shorten the data processing time of tooth model data.
In addition, currently, an automatic processing flow of orthodontics is still in a blank stage in China, and therefore there is an urgent need to achieve automatic generation and processing flow of orthodontics according to technologies such as artificial intelligence algorithm, 3D printing and numerical control machining, so as to improve the working efficiency and competitiveness of the whole industry.
Regarding the current domestic technical problems of low degree of automation, poor efficiency and poor user experience during data processing of tooth models in an orthodontic automatic process, the technical solutions of some embodiments of the present disclosure take efficiency improvement as a basis, innovatively explore a combination of the dental field and the 3D printing technology, and integrate the digital 3D printing technology into a dental restoration process. By introducing artificial intelligence technology and a series of algorithms to automatically identify data of a gum line, manpower is reduced, digital production is deepened, and the vacancy of technology of low degree of automation, poor efficiency and poor user experience during data processing of tooth models in an orthodontic process is filled up, laying a firm foundation for subsequent application scenarios of combining dentistry with 3D technology.
In order to solve the described technical problems, the object of the present disclosure is to provide a method and system for identifying a gum line of a tooth model, a device, and a storage medium.
A first technical solution used in some embodiments of the present disclosure is:
A second technical solution used in some embodiments of the present disclosure is:
a device, including a memory and a processor, wherein the memory is used for storing at least one program, and the processor is used for loading the at least one program to execute the described method.
A third technical solution used in some embodiments of the present disclosure is:
a storage medium, in which a program executable by a processor is stored, wherein the program executable by the processor is used for executing the described method when executed by the processor.
As shown in
In the present embodiment, the tooth model is a tooth model which is prepared by processing and repairing based on a 3D printing technology and a computer program and has a flat bottom surface. The geometric algorithm of curvatureis a method to calculate a rotation rate of a tangential direction angle of a surface on the tooth model to anarc length, which indicates the recessed or raised degree of the surface. A direction of the target reference line is the same as a direction of the contour point group (a direction of the tooth model,that is, a direction of an opening of the tooth model), so as to ensure that the range of the target reference line can cover a region of the contour point group. Specifically, after the shape parameter and a flat bottom surface of the imported tooth model are identified, the tooth model is adjusted to a target position; then the plurality of feature points are extracted from the tooth model at the target position according to the geometric algorithm of curvature; a reference line having the minimum deviation from a pre-stored reference line pool is obtained, as the target reference line, wherein the target reference line is matched with the identified shape parameter of the tooth model; the iterative operation on the target reference line is performed according to the contour point group, to generate the fitting reference line; and the smoothing processing on the fitting reference line is performed by using the dimensionality reduction algorithm, to finally output the gum line, so as to complete preprocessing of data of the tooth model. Regarding the problems of large workload, high labor costs, low working efficiency and poor user experience in data processing of a tooth model, such as in the current application scenarios of invisible orthodontic, the technical solutions of some embodiments of the present disclosure provide a method for identifying a gum line of a tooth model, which can greatly simplify an orthodontic automatic generation and processing flow, reduce the working costs, shorten the data processing time of tooth model data, improve the working efficiency, and enhance the user experience, and also establish application scenarios combining dental diagnosis and treatment with a 3D technology, and supplement the blank of domestic orthodontic automatic processing flow. In addition, the method for identifying a gum line of a tooth model provided in the present embodiment is also applicable to tooth models having arbitrary flat bottom surfaces and arbitrary directions, and the implementation of some embodiments of the present disclosure is not affected by the other tooth model types having arbitrary flat bottom surfaces.
As at least one alternative embodiment, the pre-processing in the step S1 includes filtering or denoising; the step S1 includes the following steps:
In the present embodiment, the initially obtained feature points of the tooth model are denoised or filtered, to eliminate the effect of feature points in non-gum line regions on the fitting process subsequently performed on contour lines; and the geometric algorithm of curvatureis a method to calculate a rotation rate of a tangential direction angle of a surface on the tooth model to aarc length, which indicates the recessed or raised degree of the surface, and saidrecessed or raised degree is referred to as feature in the solutions of some embodiments of the present disclosure, and the feature points can be extracted from recessed and raised regions of the tooth model by means of a curvature calculationmethod.
As at least one alternative embodiment, the step S11 includes the following steps:
In the present embodiment, the first objective normal vector is a corresponding normal vector of the tooth model when the tooth model is rotated to the target plane, that is, the normal vector of the target plane; and the second objective normal vector is a corresponding normal vector of the tooth model when the tooth model is rotated to the target position, that is, the normal vector corresponding to the target position. Specifically, first, an imported tooth model is acquired, and by acquiring a tooth model having a flat bottom surface in any direction, a normal vector of the bottom surface is acquired, and the tooth model is rotated to the target plane on based on a cross product method and the first objective normal vector; second, when the tooth model is located in the target plane, the tooth model is projected to obtain a contour line; skeleton extraction on the contour line is performed, and the extracted skeleton is performed noise processing, to acquire a smooth and continuous skeleton line; a skeleton curvilinear equation is solved, to obtain a direction vector of thetooth model, that is, the direction vector of the opening of the tooth model ; and based on the obtained direction vector of thetooth model and the second objective normal vector of the target position, a rotation angle and a rotation axis are worked out according to a cross product method, and then the tooth model is rotated to the target position, thereby a tooth model in any direction is automatically adjusted to the target position.
As at least one alternative embodiment, the shape parameter includes the shape of the tooth model, and the step S2 includes the following steps:
In the present embodiment, the coordinate system is a rectangular coordinate system with X, Y and Z axes, and other coordinate systems may also be selected according to application scenarios, which are not described herein again. The preset threshold refers to a value that achieves a minimum deviation between the obtained initial reference lines and the tooth model in the established coordinate system. Each initial reference line refers to a gum line file of tooth model, and the file is composed of point coordinates; and the fitting reference pool actually includes historical gum lines that have been subjected to smoothing processing, these gum lines being generated based on other tooth models.
As at least one alternative embodiment, the step S3 includes the following steps:
In the present embodiment, the approximate iterative algorithm means that after the target reference line is projected, it iterates circularly until the projection error is less than asecond threshold;Specifically, the contour point group and the target reference line are acquired, the centroid of the target reference line and a centroid of a contour point group are geometrically superposed in the pre built coordinate system, and after the target reference line is circularly projected, the final projected points are connected to form the fitting reference line, so as to ensure that the fitting reference line can cover the region of the contour point group.
As at least one alternative embodiment, the dimensionality reduction algorithm uses a principal component analysis method, and the step S4 includes the following steps:
In the present embodiment, the generated fitting reference line is not a smooth line, and may have situations of local folding and deviation. With regard to such situations, the principal component analysis (PCA) method is used to acquire a main contour shape of the fitting reference line to determine main direction points of the contour, and the main direction pointsof the fitting reference line are extracted based on the main contour shape, and the main direction points are connected smoothly after interpolation using interpolation splines such as a Kochanek-Bartels pattern, thereby avoiding unsmooth regions of line segments, and a final smooth gum line required is obtained.
As shown in
As at least one alternative embodiment, the extraction module includes:
As at least one alternative embodiment, the identification unit includes:
As at least one alternative embodiment, the matching module includes:
As at least one alternative embodiment, the iteration module includes:
As at least one alternative embodiment, the output module includes:
A device, including a memory and a processor, wherein the memory is used for storing at least one program, and the processor is used for loading the at least one program to execute the method according to the method embodiment.
The device of the present embodiment can execute the method for identifying a gum line of a tooth model provided by the method embodiment of the present disclosure, can execute any combined implementation steps of the method embodiment, and has the corresponding functions and beneficial effects of the method.
A storage medium, in which a program executable by a processor is stored, wherein the program executable by the processor is used for executing the method of the method embodiment when executed by the processor.
S51, The tooth model is acquired, and the tooth model is automatically adjusted to a target position.
As at least one alternative embodiment:
S52, Feature points of the tooth model are extracted based on a geometric algorithm of curvature.
Specifically: after the toothmodel is placed to thetarget position, the extracted feature points of the tooth model are filtered and de-noised according to thegeometric algorithm of curvature, so as to obtain a feature contour of the tooth model. The geometric algorithm of curvatureis a method to calculate a rotation rate of a tangential direction angle of a certain surface on a tooth model with respect to a corresponding arc length, which indicates the recessed or raised degree of the surface, and said degree is also referred to as feature in the present solution.That is, the feature points of recessed and raised regions of the tooth model can be acquired by the geometric algorithm of curvature (the true gum line on a tooth model is also reflected by recessed and raised regions).
S53, The feature points are optimized, to obtain the contour point group of the tooth model.
Specifically: for feature points of the tooth model obtained in Step S52, it is necessary to remove the noise. In addition tothe gum line region, other parts of teeth also have recessed or raised regions which are also mistaken as features, which will affect subsequent fitting with contour lines; thereforeneed to be filtered or removed, so as to finally obtain optimal feature points; wherein the contour point group includes a plurality of feature points remaining after optimization, and the distribution shape of the feature points in the contour point group (or contour point group shape for short) is similar to the gum line contour of the tooth model. It should be noted that the centroid of the contour point group may be understood as the centroid of the contour point group shape.
S54, According to the shape and contour of the tooth model, an optimal fitting reference line is automatically selected.
Specifically: the optimal fitting reference line is automatically selected according to the shape and contour of the tooth model after being imported and straightened, wherein the fitting reference line refers tothegum line file of tooth model, and the file is composed of point coordinates The main method for searching for the optimal fitting reference line is: first, a batch of initial reference lines in the same direction are matched according to the direction of the tooth model; and second, a initial reference line with the minimum deviation, i.e. the optimal target reference line is selected from the screened same-direction initial reference lines, according to the matching degree between the fitting reference lines and the tooth model determined based on three directions X, Y and Z..
S55, Feature points and the target reference line are fitted, to obtain a fitting reference line.
After obtaining the contour point group and target reference line, the target reference line needs to be overlapped or drawn as close as possible to the feature points in the contour point group.
Specifically: (1) After obtaining the contour point group and the target reference line, and the target reference line needs to be overlapped with or drawn close to the feature points of contour point group as far as possible, i.e. the centroid of the target reference line is translated in three directions, i.e. X, Y and Z, to coincide with the centroid of the tooth model; the centroid of the target reference line can made to coincide with the centroid of the contour point group shape, and also it is ensured that the target reference line can cover the region of the feature points in the contour point group. (2) An approximate iterative algorithm is used to fit the target reference line and the feature points of the contour point group, to form a fitting reference line. The fitting is to project points of the target reference line (substantially include a series of coordinate points) to the feature points closest thereto; and the approximate iteration means that after the target reference line is projected, iteration is continued until the projection error is less than a set threshold. (3) After the iteration is completed, the final projected points are connected to form a fitting reference line.
S56, Smoothing on the fitting reference line is performed, to obtain a final gum line.
Specifically: the fitting reference line is not a smooth line, and may have situations of local folding and deviation. With regard to such situations, in the present solution, a principal component analysis method is used to acquire a main contour shape of the fitting reference line, main direction points of the contour are determined, and the main direction points are smoothly connected by using a Kochanek-Bartels pattern, therebyunsmooth regions of line segments are avoided, and the final smooth gum line required is obtained.
Ending, the final gum line of the tooth model is output.
The present embodiment provides a method for producting of tooth instruments, including the following steps: The gum line of the tooth model is identified by the method of identifying a gum line of a tooth model described above. A cutting line based on the gum line is generated, and the cutting line into a data file recognizable is converted by a numerical control cutting machine.The membrane is pressed on the tooth modelto obtain a shell membrane.The shell membrane is cut based on the cutting line to obtain a shell shaped dental instrument.
In the present embodiment, a post-processing of the gum line can be carried out according to different application needs in the production of tooth instruments, for example, it mainly includes automatic acute angle removal and overall offset of gum line, andthe post-processing gum line is the cutting line.Specifically, after the gum line is generated, thegeneratedgum line is closely fitted with a real gum line of teeth. Since the intersection between two adjacent teeth is sharp, which does not meet the production process requirements of tooth instruments,thegenerated gum line needs to be smoothed, that is, the generatedgum line is removed the acute angle of the gum line, and a certain margin should be left between the gum line and the tooth, the gum line generated needs to offset as a whole downward (or upward), that is, the gum line is offset.Optionally, the angle of acute angle is limited to less than 180 °, and the overall offset is between 0 and 2 mm. Further,after the cutting line is obtained,the cutting lineis converted into the data (which can be called NC file) recognizable by the CNC cutting machine through a cutter path algorithm (NC program), and the data is performed simulation verification. For the data that fails to pass, the data of the gum line is readjusted, and the NC file is circularly generated until the verification is passed; Finally, NC cutting files that meet the requirements are output. The cutter path algorithm determines the final output NC file based on the machine parameters (such as rotation axis, coordinate positioning, speed, etc.)
The content above makes specific explanation on the preferred embodiments of the present disclosure, but the present innovative creation is not limited thereto. A person skilled in the art could also make various equivalent modifications or replacements without departing from the spirit of the present disclosure, and these equivalent modifications or replacements are all included in the scope defined by the claims of the present disclosure.
| Number | Date | Country | Kind |
|---|---|---|---|
| 202010634000.2 | Jul 2020 | CN | national |
The present disclosure is proposed based on Chinese Patent Application No. 202010634000.2 and filed on 02 Jul. 2020, and claims priority to the Chinese Patent Application, the disclosure of which is hereby incorporated into the present disclosure for reference in its entirety.
| Number | Date | Country | |
|---|---|---|---|
| Parent | PCT/CN2020/134573 | Dec 2020 | US |
| Child | 18072732 | US |
