Patent for “Aligner damage prediction, mitigation”

3D Printing – Yuxiang Wang, Rohit Tanugula, Reza Shirazi AGHJARI, Andrew JANG, Chunhua Li, Jun Sato, Luyao Cai, Viktoria Medvinskaya, Arno Kukk, Andrey Cherkas, Anna Akopova, Kangning Su, Align Technology Inc

Search Patent for “Aligner damage prediction, mitigation”

Abstract for “Aligner damage prediction, mitigation”

“Embodiments are a solution to aligner breakage. This method involves the creation of a digital model of a polymeric aligner that can be used to fit a patient’s dental arch. The shape of the polymeric aligner is to apply forces on teeth in the dental arch. This method includes analyzing the digital design of the aligner with at least one of the following: a) A trained machine learning model; b) A numerical simulation; c) A geometry evaluator, or d) A rules engine. Based on the analysis, the method might also include determining whether the digital design for the polymeric alignmenter contains probable points of damage. A probable point of damage is defined as a threshold probability of breakage, deformation or warpage. This method could also include corrective actions, based on probable points.

Background for “Aligner damage prediction, mitigation”

Orthodontics is a process that moves a patient’s tooth to a position where their function and aesthetics can be optimized. Braces, which are appliances like braces, are traditionally applied to the teeth of a patient by an orthodontist. The braces exert constant force on the teeth and push them towards their final destination. The braces are adjusted over time by the orthodontist through a series clinical visits and adjustments. This allows the teeth to reach their final destination.

Alternatives to traditional orthodontic treatment using fixed appliances (e.g. braces), include systems that use a variety of preformed aligners. These systems allow patients to design and/or fabricate multiple aligners, sometimes all of them, before they are given to the patient. They also permit the patient to reposition their teeth at any time (e.g., during treatment). Computer-based, three-dimensional (3D), planning/design tools may be used to design or plan a personalized treatment for a patient. Computer modeling can be used to design the aligners. The individual aligners are worn over the teeth to reposition the teeth according to the plan.

A series of preformed aligners can be made from a material that imparts forces to patient’s smiles. One or more polymeric materials are examples. Thermoforming aligners can be made using a variety of molds, such as 3D-printed molds or directly fabricating them. To achieve negatives in some thermoforming techniques, shells are molded around the molds. Shells are then removed and re-used for different purposes. Corrective dentistry and orthodontic treatment are two examples of situations where a shell may be formed around a mold. The mold could be a positive mold for a dental arch and the shell might be an aligner for one or more of the patient’s teeth. The mold can also contain features that are associated with attachments, such as planned orthodontic attachments.

Molds can be made using many techniques such as casting or rapid prototyping equipment. 3D printers can make molds of aligners by using either additive manufacturing techniques (e.g. stereolithography) and subtractive manufacturing techniques like milling. The molds can then be covered with aligners using thermoforming techniques. After an aligner has been formed, it can be manually or automatically trimmed. Sometimes, an automated 4-axis or 5-axis trimming device (e.g. a laser trimmer or mill) can be used to trim the alignment along a cutline. The cutting machine uses electronic data to identify the cutline for trimming the aligner. The aligner can then be taken out of the mold and shipped to the patient. Another option is to directly fabricate aligners using digital light processing (DLP), stereolithography (SLA), or other 3D printing methods.

It may be beneficial to identify certain portions of aligners that are prone to warpage, deformation or breakage during fabrication (e.g. in response to removal form a mold) or use (e.g. in response to the removal from a patient?s dentition). However, existing techniques make this difficult.”

“Aligners are also referred to as?orthodontic alignmenters? One type of dental appliance (also known as an ‘appliance?) may be used. It is used to treat malocclusions and to apply to the patient’s teeth. FIGS. shows examples of aligners and aligner system. 15A and 15B. FIG. 15C. Aligners can be made from polymeric materials by using either indirect or direct fabrication techniques. Examples of these may be found in the discussion of FIGS. 15A, 15B and 15C. Many aligners can experience strains/stresses when they are removed from their molds, as noted above. Many aligners, whether directly or indirectly, may also experience strains/stresses due to prolonged use in an intra-oral environment (e.g. up to twenty-three hour per day for several months) or repeated removals (e.g. several times per day for several years) from the patient’s dentition. Strains and stress can cause permanent or temporary damage such as warpage, cracks, fractures, and damage. Aligners There are serious issues with physical damage in manufacturing processes, such as materials waste, supply chain issues and inability meet consumer demand. Also, physical damage from use can cause serious problems such as adversely impacting staging situations and/or the effectiveness of treatment plans.

“The embodiments are systems, methods, or computer-readable media that can predict deformation, warpage and/or breakage in custom manufactured products (e.g. of aligners) before and/or during fabrication (e.g. in response to removal of a mold) as well as use (e.g. in response to the removal of a patient’s tooth). There are also embodiments that address resolving or mitigating predicted points of damage. These embodiments also cover methods to optimize properties of aligners, such as thicknesses. This includes predicting portion of aligners that are susceptible to deformation, warpage or breakage and identifying how these properties correspond with clinical goals. Many embodiments can also use the optimized properties of aligners to create efficient manufacturing processes and customized or optimized treatment plans. These features can be used together or alone to provide solutions for aligner damage, such as solutions that allow for alignment damage during manufacturing, use, and so on.

“Alterer damage solutions systems and methods can be used in certain embodiments of the design of aligners before or during manufacturing. It can be difficult to design custom-made products, especially when orthodontic aligners are made for each patient. Many orthodontic treatment plans recommend a series aligners to be used for treatment. Each aligner may be used for a specific stage in a treatment plan and/or have different properties (e.g. shape) than other aligners in a series. Many orthodontic treatment plans will provide patients with two aligners, one for the upper and one for the lower. In some cases, one treatment may include 50-60 stages to treat a complex case. This means that 100-120 aligners can be made for a single patient.

For aligners made by indirect fabrication techniques (e.g. thermoforming), the force and/or moment required to remove an aligner from a mould may be applied to it. Sometimes, the force exerted on the aligners can cause the polymeric material to break, warp, or deform. Additionally, aligners may be damaged by force or torque when removed from patients’ teeth. Aligners may also become brittle, warpy, or deformed. If the aligner is damaged, patients may request a replacement. A new aligner is made and shipped to the patient. As you can see, the manufacturing cost increases as more replacement aligners are manufactured. Sometimes, an aligner replacement may need to be modified manually after manufacturing in order to account for breakage. To strengthen the alignmenter, it may be possible to add filler material to the aligner if the area where the aligner has broken is located. It can be difficult and slow to modify the aligners manually after manufacturing, especially if you have hundreds, thousands, or many replacement aligners. Modifying the replacement aligners after an alignment problem has occurred is reactive. The present disclosure could provide an automated, more scalable and proactive method to detect potential damage in an aligner design and take corrective action before the manufacturer of the aligner is made. Embodiments can reduce the likelihood of alignment damage and may decrease the number of replacement aligners manufactured. This reduction in damage may lower the cost of manufacturing aligners, and it may also reduce the time required to resolve aligner problems.

An aligner can be made from a polymeric shell designed to hold a patient’s upper or lower dental arch at a specific stage of treatment. An aligner can be designed to apply force to the patient’s tooth at a particular stage of orthodontic treatment. Each aligner has a tooth-receiving cavity that receives and resiliently positions the teeth according to a specific treatment stage. A?cap’ may be used to refer to each tooth-receiving cavities. The aligners can be used to reposition teeth by moving one or more of them vertically (e.g. extruding or inserting teeth), rotating one (or more) teeth (e.g. through moments applied on the teeth, second/third order rotations etc. The aligners can be used to move one or more of the teeth in a transverse orientation relative to the arch and/or one or more in an anterior-posterior relative to that arch. An aligner can also include features that attach to the patient’s dentition and allow for tooth repositioning or rotation.

“Embodiments can identify individual aligners with probable points of failure, and/or sets of aligners (e.g. for a patient or a specific dental arch of a person) that contain one more aligner that has a likely point of failure. Aligners may have manufacturing flows based on the likelihood of them becoming damaged (e.g., a point damage). Manufacturing flows can also be used to determine sets of aligners. This could include all aligners that are part of a treatment plan or all aligners that are part of a treatment plan.

“As we have already mentioned, different points of damage may be determined by the embodiments for a particular set of digitally designed aligners. Breakage, warpage or deformation are all possible points of damage. The possibility of detecting the points of potential damage can be used to modify or fix the aligner’s digital design to eliminate them before manufacturing. This will increase the likelihood of aligners being manufactured successfully, reduce patient complaints about aligners not working, lower manufacturing costs for replacement aligners, and/or prevent the manufacture of an aligner with a probable point. Note that “probable” is not always the best term. Points of damage are sometimes interchangeable with ‘likely? Points of damage (also known as?likely?) may be used interchangeably with?likely. Not all probable points of damage must indicate damage caused by, e.g., an overwhelming amount. Probable points may indicate the likelihood of damage exceeding any threshold, even a preponderance.

“In some cases, the alignment’s digital design may help determine the likely points of damage. A digital model of an aligner, including its geometry, may be called the digital design of the alignmenter. The digital model of each aligner can be included in a file that is associated with the aligner. The digital model of an aligner can be created using scanning of the aligner, such as with an intraoral scanner, or any other 3D scanner, and then generating the digital version of the aligner based on the scan. Another method is to use a digital model that represents the dental arch of the patient to create the digital model. To create the digital model, the digital mold model may be offset (e.g. enlarged). The?digital design for the aligner? The?digital design of the aligner? These terms may be interchangeable. Analyses of the digital design of an aligner may be done using at least one of the following: a) A trained machine learning model that can identify aligners with potential points of damage; b) A numerical simulation that simulates the removal of the alignmenter from a dental arch mold of the patient; c) A numerical simulation that simulates the loading around weak areas (e.g. interproximal regions) within the aligner; e) A geometry evaluator that calculates or evaluates geometry-related parameters. The analysis may determine whether the digital design for the aligner contains one or more points of damage. A probable point of potential damage must be present if there is a minimum threshold probability of breakage, deformation or warpage, failure, or any combination thereof. will occur. If the alignment’s digital design contains the potential points of damage, you can take one or more corrective measures (e.g. altering the aligner’s digital design, changing attachments, notifying a dentist, etc.). Based on one or more points of damage, the corrective actions may be taken. The disclosed embodiments have the potential to automatically detect and correct various points of damage in aligners. They also allow for automated selection of manufacturing flows for aligners that are based on the presence or absence of possible points of failure/damage. These embodiments can reduce the number and cost of replacement aligners, which in turn will lower manufacturing costs and improve customer satisfaction. Some implementations may require that the probable points of damage be determined before designing aligners with variable thicknesses to support those points. These flexible thicknesses can be made using direct fabrication techniques or other methods to accommodate damage points. The embodiments can also increase the number of aligners manufactured without deformations, warpages, breakages or warpages.

“The disclosed embodiments may be implemented using a variety of software and/or hardware components, as illustrated in FIG. 14. Software components can include instructions stored on a tangible, nontransitory media. These instructions are executed by one to several processing devices in order to create aligner damage solutions for digital aligners. Hardware components can include a memory device, network device and processing device.

The shape of an aligner (e.g. the shape of each tooth receiving a cavity (cap), the shapes between the tooth receiving cavities and the outlines of any additional cavities created to accommodate attachments on patients’ teeth) will all impact whether or not an aligner will fracture, warp, or become damaged when it is removed from a dental arch-like structure. The shape of an alignment can be altered to address potential points of damage. This may lead to an aligner with variable thicknesses. Modified aligners can be made using a variety of techniques, including direct fabrication techniques.

“In accordance with the inventions, each aligner may have a digital design. This can be used to analyze the alignments at specific stages. A digital design for an aligner can be linked to a digital design for a dental arch at each stage of treatment. Corrective actions can be taken if there are any probable points of harm in a digital aligner design.

“Embodiments” are described using aligners and orthodontic aligners (e.g. polymeric aligners or polymeric orthodontic aligners). Aligners can be described as a type of dental appliance. Aligners, which are described in detail below and above, may be used to correct malocclusions. The embodiments discussed herein with regard to aligners can also be applied to other types and types of dental appliances and dental shells, in particular other types and types of polymeric dental appliance, such as night guards and sleep apnea treatment device, and so forth.

Each aligner can be made by molding polymeric material to implement one or several stages of a treatment plan for a patient’s teeth. This could be done either through indirect fabrication techniques, or direct fabrication techniques. With respect to FIGS., we will also discuss indirect and direct fabrication methods. 15A, 15B and 15C. FIG. FIG. 1A shows a flow diagram of a method 100 for performing a corrective assessment on a digitally designed polymeric aligner. This is in accordance to one embodiment. Processing logic on a computing device performs one or more of the operations in method 100. Processing logic can include hardware (e.g. circuitry, dedicated logic or microcode), as well as software. Software (e.g. instructions executed by a processor device), firmware, and/or a combination thereof. A processing device that executes the aligner design analysis module 1450 in FIG. may perform one or more of the operations described above. 14. The method 100 can be used for any unique aligner in a patient’s treatment plan or at any stage (e.g. key stages) of that treatment plan.

“At block102, processing logic might obtain a digital model of an aligner to fit a dental arch. In the digital design, the aligner (e.g. a polymeric one) is designed to apply forces to one of the arch’s teeth. The processing logic may be provided with a file that includes the digital model of the mold used in creating the aligner. To dynamically create the digital design for the aligner, the processing logic can manipulate the geometry of the digital mold model (e.g. enlarge it). The processing logic can receive the digital design for the aligner using another system or scan a pre-made aligner. The digital design of an aligner may be a three-dimensional (3D), virtual model that was created based on a virtual model of the dental arch at treatment stage.

“At block104, processing logic can perform an analysis of the digital design for the polymeric alignmenter using at most one of a) A trained machine learning model that is trained to identify possible points of damage in polymeric aligners, b) A numerical simulation associated to the removal of the aligner from the mold of the patient’s dental arch, c) A numerical simulation that simulates loading around weak areas (e.g. interproximal regions of the aligner), d) A geometry evaluator that a rule engine that contains one or more rules that are indicative of polymeric alignments of their parameters. Further details regarding the analysis of the digital design for the polymeric alignmenter using the trained model are provided in FIGS. 2A-2D. Further details regarding the analysis of the digital design of aligner using numerical simulation in conjunction with the removal of aligner from mold are discussed below. Refer to FIGS. 3A-5B. FIGS. 8A-8C provide additional details on how to analyze the digital design of an aligner using the numerical simulator associated with progressive damage. 8A-8C More details about performing an analysis of the digital alignment using the numerical simulation, or a geometrical evaluation that simulates loading around weak points of the aligner, are discussed in FIGS. 6A-7G. FIG. 6A-7G provides additional details on how to perform analysis of the digital design using the rules engine. 10.”

“At block106, processing logic can determine, based upon the analysis, if the digital design for the aligner contains one or more likely points of damage. A probable point may be defined as an area with a minimum probability of breakage, deformation or warpage due to the removal of the aligner form the mold, the removal of the aligner from teeth, or the use of the aligner. Processing logic determines whether there are any possible points of damage at block 107. The method will continue to block 108 if at least one point of potential damage has been identified. The method could end if it is not.

“At block108, processing logic can perform one or more corrective action and/or choose a manufacturing flow based upon the one or multiple points of damage. One or more corrective actions may include modifying the digital design for the aligner in order to create a modified version of the aligner. One or more corrective actions may include modifying the digital design associated with the aligner’s digital design. The digital design for the aligner can also be modified to reflect the new digital design.

Modifying the alignment’s digital design may be necessary if a point of potential damage is found to be near an outline. In the digital design of an aligner, for example, the outline can be reduced to be more straightened, rather than more pointed. This reduces the strength and durability of the aligner at the point. The alignment’s strength may be increased if the outline is straightened. This may also reduce the likelihood of any damage due to the digital design. Modifying the digital design may be necessary if there is a possibility of damage to the interproximal area between the two teeth. An example of this is to make the outer surface of an aligner’s digital design flatter by increasing its thickness. The portion of the digital layout that is thinner may help strengthen the alignmenter and remove any potential damage. Some embodiments allow for control of the thickness of the aligner for aligners made directly using 3D printing techniques, but not for aligners made by thermoforming processes.

“In certain embodiments, the modification of the digital design may involve inserting an indicator into the aligner. The indicator is a place where it is recommended to start removing the aligner. The digital design analysis may determine the best place to place the indicator. The analysis might show that applying force to a specific location on the alignment to remove the digitally designed aligner may cause less damage than other locations on the aligner’s digital design. The indicator could be placed in that location.

“In certain embodiments, if there is a possibility of damage at or near a place in the digital design for the aligner that is associated to an attachment (to teeth), then the corrective action could include changing one or more attachments that are associated with the possible point of damage on one of the virtual 3D models of the dental arch. Modifying the dental arch’s 3D model may result in a modified 3D model that is generated from the changes made to the attachments. A cavity that houses the attachment can be moved, increased, decreased, or modified in the modified 3D model. This is based on changes to the attachment in 3D model.

“In certain embodiments, if there is a possibility of damage at or near the location of two teeth, then corrective action could include adding a virtual filler or expanding an existing virtual filler to the area on the virtual 3D dental arch that corresponds to the possible points of damage. A virtual filler can be a digital element or addition to a virtual 3D model (such as a model of a dental arch) that adds an object between two or more adjacent tooth. The virtual filler in the virtual model alters the geometry of the respective mold and reduces the likelihood of fabrication problems. Based on the modified 3D model, the aligner can be created using the modified virtual model. To accommodate virtual fillers, the aligner may have a flatter surface between two teeth. An aligner with a flatter surface between teeth can increase its strength and reduce the risk of injury from digitally designed aligners.

A modified digital model of the alignmenter can be created based on any modifications made to it. Processing logic can determine whether the altered digital design of an aligner contains the one or more likely points of damage. Processing logic can respond to the determination that the altered digital design of the alignmenter contains one or more likely points of damage and may take one or more additional corrective actions based upon the probable points. This may be repeated until all the possible points of damage have been removed from the aligner’s digital design.

“In some cases, the digital design for the aligner may be received during the treatment planning phase. If there are any probable points of damage to the digital design of an aligner, some embodiments may recommend that one or more attachments be modified on one or several teeth to decrease the probability of the probable point failing to reach the threshold probability. The corrective action could include the recommendation to modify the digital design for the aligner to move one of the teeth with another aligner at a different stage in the treatment plan. This will reduce the likelihood that the probable point will fail to below the threshold probability. A particular attachment may be used to achieve a specific tooth rotation. You can modify the treatment plan to move the specific tooth rotation to a later stage of treatment. This will allow you to use the attachment at the later stage.

“In some cases, the corrective action may be to recommend one or more procedures to remove the aligner from a patient’s dental arch to lower the chance that it will fail below the threshold probability. In some cases, corrective action could include notifying a dentist during the treatment planning phase about the possibility of damage to the aligner’s digital design. If the likely point of damage cannot easily be fixed by altering the digital design of an aligner, then processing logic might notify the dentist.

“In certain embodiments, the corrective action may be performed based on one or more points of damage. This could include attaching a flag to the aligner to indicate quality inspection should take place on the aligner following manufacturing. The flag could cause quality inspection to focus on the likely points of damage. In certain embodiments, the corrective actions may include recommending that a targeted inspection be conducted by notifying a system of inspectors.

“In certain embodiments, corrective action may include setting a flag to avoid using a mold that is too fragile or one that has been damaged during manufacturing. Breakable molds may be defined as molds that have been broken to remove the aligner. The likelihood of the aligner failing during removal can be decreased by using less force on the broken mold.

“In some cases, corrective action may involve changing the geometry of the 3D model of a mold. A portion of the mold’s virtual 3D model may be thickened or bubbled. Based on the modified model, a modified virtual model of an aligner can be created. The shape of this modified model could be different from the original 3D model. The amount of force required to remove the aligner (or arch) from the mold or dental arch at any given location can be reduced by bubbling out, thickening, or expanding the digital model (and so the aligner). Thus, breakage, warpage, etc. This location could be mitigated.

“In some embodiments, a manufacturing process may be chosen for an aligner located at block 108 based either on a prediction about a probable cause of damage or loss for the aligner, or on the absence of such a probable cause of damage or loss for the aligner.”

“FIG. “FIG. Processing logic on a computing device performs one or more of the operations in method 110. Processing logic can include hardware (e.g. circuitry, dedicated logic or microcode), as well as software. Software (e.g. instructions executed by a processor device), firmware, and/or a combination thereof. A processing device that executes the aligner design analysis module 1450 in FIG. may perform one or more of the operations of method 110. 14. The method 110 can be used to align aligners for patients, as well as aligners associated with treatment plans for particular arches. In embodiments, method 110 can be used for each individual aligner in a patient’s treatment plan or at key stages of the plan. Method 110 can be performed at block 100 of method 100 in embodiments.

“At block 110 of method 110 processing logic receives data about the possible points of failure of a plurality aligners. These data could be associated with aligners that are part of one or more treatment plans for one patient. Data on the likelihood of points failure could have been generated by: a) a trained model that can identify aligners with probable points of harm, b) an numerical simulation associated to the removal of the aligner form a dental arch of the patient and c) a simulation associated to loading around weak spots in aligner, as well as e) a geo evaluator that evaluates parameters related to the geometry of the polymeric alignmenter or f) rules engines that include one or more rules that are associated with aligner parameters that indicate damage. In some cases, data about the likelihood of points of failure can be generated by multiple of the above simulators, rule engines, and/or machine-learning models.

“At block114, processing logic aggregates data for aligners that are associated with the same treatment plans into one or more sets. One embodiment aggregates the failure probability data of all aligners that are associated with a treatment program into one set. Alternately, failure probability data for aligners that are part of the same treatment plan can be combined into multiple sets. Data for a patient’s lower dental arch (e.g. data for each stage of treatment) can be combined into one data set. The data for a patient’s upper dental arch may be combined with the first data set.

If probabilities of damage to aligners are given by multiple methods (e.g., a machine learning model and simulation outputs or two simulation outputs or geometry evaluation), then it is possible to combine the predictions from the different techniques to increase accuracy. Data may be received for one aligner. This data could include a first probability that the aligner will fail as an output from a machine learning model, and a second probability that the aligner will fail as an output from a numerical simulation.

“Block 116 is where each data set is evaluated to determine if all aligners within any data set have a probability that they will cause damage or failure below a lower threshold. For example, the lower threshold could have a value of 2%, 5% or 10% and a 15% chance of a point causing damage/failure. Low risk aligner sets may include aligners that do not contain aligners that have points with a higher probability of damage/failure than the lower threshold. These aligner sets with low manufacturing risk may be manufactured quickly and require fewer manufacturing steps. This may help to reduce manufacturing costs and speed up manufacturing. If the probability of any one aligner failing in a set is below the lower threshold, then the method blocks 118 and determines a first manufacturing flow for that aligner set. For example, a fast-track manufacturing flow could be the first manufacturing flow. Fast track manufacturing flows may assume that there will be no exceptions, that no aligners will need rework and that the manufacturing process can be completed in minimal time. If any of the aligners in an alignment set have points that are at or above the lower threshold for failure/damage, the method could continue to block 120.

“In some embodiments processing logic selects between two manufacturing flows. The operations of block 112 are skipped. The method proceeds from block 114 to block 120.”

“At block 120 each data set is evaluated to determine if any aligners in any data sets have a probability or failure that is higher than an upper threshold. For example, the upper threshold could be a value of 45%, 50% or 55% and a 60% chance of a point damage/failure. High-risk aligner sets may include aligners that contain at least one alignment with a probability for damage/failure exceeding the upper threshold. These high-risk aligner sets could be subjected to more scrutiny, slower manufacturing, additional quality control steps, etc., which may decrease the likelihood of aligners being damaged or increase detection of any damage. If there are no points in an aligner set that have a probability of failure or damage at the upper threshold, then the method will continue to block 122. A second manufacturing flow can be chosen. If there is a higher probability of a set of aligners failing than the upper threshold, then the method blocks 124. A third manufacturing flow for the aligner set is created. The second manufacturing flow could be a standard manufacturing process for aligners. The third manufacturing flow could be a quality control manufacturing process (e.g. that inspects all or part of an aligner set using an image-based quality control inspection station). In some embodiments, the third manufacturing flow can be done by the most skilled technicians or operators. One embodiment increases the cycle time for the third manufacturing process to allow the operator more time to handle aligners (e.g. to remove aligners form molds). Block 118’s first manufacturing flow may have a low level of complexity. The second manufacturing flow at block 122 may be a workflow with a higher level of complexity. Block 124’s third manufacturing flow may be the most complex.

“FIG. 2A shows a flow diagram of a 200-step process of training a machinelearning model to analyze a digitally designed aligner. This is in accordance with one embodiment. According to one embodiment, the machine learning model can be used to predict whether an aligner will become damaged during manufacture.

Processing logic on a computing device performs one or more of the operations in method 200. Processing logic can include hardware (e.g. circuitry, dedicated logic or microcode), as well as software. Software (e.g. instructions executed by a processor device), firmware, and/or a combination thereof. A processing device that executes the aligner design analysis module 1450 in FIG. may perform one or more of the operations of method 200. 14.”

“At block 200 of method 200 processing logic may preprocess digital design for a plurality orthodontic aligners in order to use the digital designs as training data for a machine-learning model. Some digital designs may be used in conjunction with already manufactured orthodontic aligners. A clinical data store might store information about the manufacturing damage of each associated orthodontic aligner. Some digital designs of orthodontic aligners could be linked to other digital designs. It is possible that such digital designs are not associated with orthodontic aligners that were damaged in manufacturing.

Blocks 204 to 208 are used in one embodiment for creating digital designs for orthodontic aligners that haven’t been manufactured or for which damage information is not available. These can be used in training a machine-learning model. Block 204 is where processing logic processes digital designs of one or more aligners. This may be done using one or several numerical simulations that determine the likelihood of damage. To determine the likely points of failure or damage, any of the numerical simulations discussed herein can be used. The digital design of an alignmenter can be used to determine the probable points of damage. 3A-8C are examples.

“As mentioned above, and further discussed below with regard to FIGS. 3A-5B, a numerical simulation can be done on the digital model of the aligner in order to simulate one or more forces or displacements on it. The forces may be used to simulate the removal of the aligner from a dental-arch-like structure (e.g. teeth or mold). The numerical simulation can calculate the amount of force needed to remove the aligner form a dental arch-like structure. It also determines the stress/stress/deformation energy or deformation level at any point on an aligner that exceeds a threshold, which could indicate that the point is likely to fail. The displacement, motion, or geometry changes at the points can be used to determine the strain. Force applied to the aligner may also be used to determine the stress. A strain threshold or stress threshold can be used in some embodiments to predict when an aligner point will fail. The numerical simulation can be used to predict potential points of damage on the aligner’s digital design. These simulations can be performed multiple times on different digital designs of aligners. Labels may also be included to indicate whether the digital designs contain one or more points of potential damage.

“At block206, processing logic might determine for each of these digital designs whether there are likely points of damage for the respective aligners. Processing logic may also add information to block 208 about the likely points of damage to each respective orthodontic aligner’s digital design. This may include information about the likely points of failure, and/or the likelihood of damage/failure at each point of probable failure. Processing logic can also add information to indicate that there are no probable points for failure in digital designs of orthodontic aligners. In some cases, the probable points of failure are those points on an aligner that have a higher probability of damage than a threshold value, such as 50%, 60% or another value. Digital designs of aligners may use the probable points of failing and their absence as labels. Digital designs with one or more likely points of failure may have a label of 1. This indicates that it is probable that the aligner will fail during manufacturing. Digital designs without probable points of failure may have a label of 0. This indicates that it is likely that the aligner will not be damaged during manufacturing.

Blocks 210 to 216 are used in one embodiment for creating digital designs for orthodontic aligners that have already been manufactured and for which damage information is available. This data can be used in training a machine-learning model. Processing logic may be able to receive digital designs of one or more aligners at block 210. Processing logic might receive information at block 212 indicating the damage to one or more orthodontic aligners during manufacturing. Processing logic might also receive information about the location of damage to manufactured aligners, and/or the type of damage (e.g. cracking, warping, deformation). In some cases, patients may also be able to report actual damage to aligners.

Historical patient feedback may provide information about whether or not the aligners sustained damage. Patients may submit a report detailing the failure of an aligner and/or the location of damage. This report can be scanned, emailed, or printed. The patient can also specify which aligner was used at what stage of the treatment plan. Sometimes, the patient can return the damaged aligner to the site. The broken aligner could be scanned at that site to get an image of its digital design and indicate the exact location of the damage. Images of broken aligners can be taken to create image corpora, which is a collection of images that may include large numbers of images, and then used in training data. The patient’s information about the alignment may be used to identify the individual aligner. This can be done by using scanned images or the information provided by the patient. An indication label may be added to the aligner that indicates there is damage.

“At block 216, processing logic might add information about damage (e.g. about points of damage), to the digital designs for each aligner. This may include information about the location of any damage or failures. Processing logic can also add information to digital designs of orthodontic aligners about whether there was any damage during manufacture. Digital designs for orthodontic aligners may be labeled with the probable damage and absence of damage. Digital designs that have suffered damage may have a label of 1. This indicates that the aligner was damaged during manufacturing. Digital designs that have not suffered any damage may be given a label of 0. This indicates that the aligner was uninjured during manufacturing. As such, the actual damage to physical aligners can be added as metadata or labels to the digital designs. Digital aligners may be labeled with information to indicate whether the physical aligners have one or more damaged points. However, the exact locations of the damaged points will not be indicated.

“At block 218, processing logic can extract at least one of the geometrical characteristics, treatment-related characteristics or clinical characteristics of the digital designs for the orthodontic aligners. One embodiment extracts the characteristics using a software module that analyses three-dimensional virtual models of aligners and/or dental arches. Based on this analysis, it determines the characteristics of associated aligners and/or arches. There are many characteristics that can be extracted, such as those that do not affect whether or not an aligner will break. Geometrical characteristics can include the individual tooth shape of one or more teeth, the location of teeth in relation to each other, jaw shape, as well as the relationship between teeth and dental arch. Treatment-related characteristics include the number of stages, location and number of attachments to teeth and whether or not aligners are active aligners. Some examples of clinical characteristics are tooth crowding, deep bit, malocclusion, etc. The characteristics extracted from processing logic in embodiments may be presented as structured or tabular data. As such, the attributes about an aligner that is associated with a digital device can be represented as structured data or tabular data.

“A block 220 may allow you to select a subset from the characteristics of each digital design. One embodiment includes the same characteristics in the subsets of each digital design. A subset of characteristics could be characteristics that are related to manufacturing defects or damage in aligners.

“Table 1 below lists many characteristics that can be extracted from a digital digital model of a tooth arch or an aligner. This is according to one embodiment. For one embodiment, Table 1 also indicates whether each characteristic was part of the subset at block 223. The table 1 only shows a sample of the many types of characteristics that can be extracted from a digital digital model of a dentist arch or an aligner. Although most of the characteristics shown in Table 1 are included in this subset, some embodiments may include less than half (e.g. just a fraction) of the total extracted characteristics.

“In\nCharacteristics Description of characteristics subset?\nActive aligner count Number of active aligners (integer) Yes\nLeft molar shift Left molar’s shift from ideal Class1 position divided by distance Yes\nbetween ideal BiteClass2 and ideal BiteClass1 (%)\nLeft canine shift Left canine’s shift from ideal Class1 position divided by distance No\nbetween ideal BiteClass2 and ideal BiteClass1 (%)\nRight molar shift Right molar’s shift from ideal Class1 position divided by distance Yes\nbetween ideal BiteClass2 and ideal BiteClass1 (%)\nRight canine shift Right canine’s shift from ideal Class1 position divided by distance No\nbetween ideal BiteClass2 and ideal BiteClass1 (%)\nCanine average tooth width Average width of canine teeth (mm) Yes\nCanine average tooth height Average height of canine teeth (mm) Yes\nCanine ridge count Total number of ridges on canines (integer) No\nCanine depth delta Delta between initial depth and planned depth for a canine (mm) Yes\nCanine maximum angulation Maximum tooth angulation of canines in one or more axes Yes\n(degrees)\nCanine maximum inclination Maximum tooth inclination of canines (degrees) Yes\nIncisor attachment count Total number of attachments on incisors (integer) Yes\nIncisor average crown height Average height of crown height of incisors (mm) Yes\nIncisor maximum angulation Maximum tooth angulation of incisors in one or more axes Yes\n(degrees)\nIncisor maximum inclination Maximum tooth inclination of incisors (degrees) Yes\nCanine maximum Absolute distance between tooth front point and jaw arch along jaw Yes\nprominence occlusal plane for canines (mm)\nIncisor maximum Absolute distance between tooth front point and jaw arch along jaw Yes\nprominence occlusal plane for incisors (mm)\nMolar attachment count Total number of attachments on molars (integer) Yes\nMolar average crown height Average height of crown height of molars (mm) Yes\nMolar maximum prominence Absolute distance between tooth front point and jaw arch along jaw Yes\nocclusal plane for molars (mm)\nPassive aligner count Number of passive aligners (integer) Yes\nFinal premolar crowding Final premolar crowding minus sum of collision depths for all teeth Yes\npairs between first premolars of given jaw\nInitial premolar crowding Initial premolar crowding minus sum of collision depths for all teeth Yes\npairs between first premolars of given jaw\nPremolar attachment count Total number of attachments on premolars (integer) Yes\nPremolar avg. crown height Average height of crown height of premolars (mm) Yes\nPremolar max angulation Maximum tooth angulation of premolars in one or more axes Yes\n(degrees)\nIncisor max inclination Maximum tooth inclination of premolars (degrees) Yes\nIntermolar distance Distance between leftmost and rightmost back molars (mm) Yes\nSpee curve for molars Spee curve depth for molars Yes”

The Spee curve, also known as the Curve of Spee, is a possible characteristic that can be extracted. It is the curvature in the mandibular-occlusal plane that begins at the premolar and continues to the terminal molar. The Spee curve, which is also known as the “Spee curve”, is an anatomic curvature that affects the occlusal alignment. It begins at the lower incisor and continues to the anterior border. This curvature can be measured by either finding a circle in 2D space in the sagittal plane or a sphere 3D space that fits the best a set tip points of the lower jaw. Curvature may be measured by the radius and angle between segments connecting the center circle to the tip of the terminal molar or the first incisor. The radius and angle of the circle are both important indicators of curvature.

The curvature of each jaw arch can be measured separately to determine the Spee curve in 2D space. The tip points can be projected onto a jaw-midline plane (e.g. where the x coordinate equals zero). This may solve the problem of finding the center of the circle and its radius that best fits all points:

“Euclidean distance (between the points and circle);”

“At block 222, processing logic can generate an embedding to each digital design of an aligner using a subset the characteristics that have been determined. In some cases, the embedding can be structured or tabular in data format.

“An alternative embodiment may not perform block 218 or 220. One or more height maps can be generated using the digital design of the aligners, e.g. from the 3D digital model or the aligner. You can project the 3D digital model onto different planes using multiple perspectives to create height maps. Block 222 may contain embeddings that are generated from the multiple height maps and a digital design.

“At block 224, processing logic creates a training databank that includes digital designs for a number of aligners. In an embodiment, the training dataset could include embeddings created at block 221. Each embedding can be associated with metadata that indicates whether the aligner associated to the embedding has been damaged or undamaged. Preferably, the training dataset contains thousands, tens or thousands, hundreds, of thousands, or more data points. Each data point is data (e.g. an embedding) associated to a different alignment. Combining digital designs of aligners and associated points of harm (as determined by real-world data) with digital designs of aligners that have associated probabilities of damage (as calculated by a numerical simulation), can be combined to create a robust machine learning model capable of predicting the probable points damage of new aligners based on digital models of these aligners in certain embodiments. In some embodiments, the machine learning model and statistical model can also be used to classify damage types, degrees of damage, or other information about aligners.

“At block 226, processing logic creates a machine-learning model from the training dataset. The machine learning model can be trained to process data (e.g. an embedding) from an orthodontic design and to output a probability of damage to the aligner. This may include shipping and handling, manufacturing, clinical usage, and any other possible outcomes. The machine learning model may be trained to produce false positives at a desired target rate, such as 2% or lower.

A machine learning model can refer to a model created by a trained engine using a training data set (e.g. training input, target outputs, labels). A set of training data may include at least one of the following: a) Digital designs of a first group of aligners, with labels that indicate whether each aligner has suffered one or more points or damage; b) Digital designs of a second group of aligners, with labels that indicate whether each aligner in the second set is likely to sustain damage. A machine learning model can be composed of one level of linear or unlinear operations (e.g. a support vector machine, or a neural network), or it may include multiple levels of nonlinear operations. Deep networks and neural networks can include convolutional neural network and/or hidden layers recurrent neural networks. One type of neural network is one that has interconnected nodes. Each node receives inputs from another node and performs one or several operations before sending the output to another node for further processing.

Convolutional neural networks are architectures that can provide image recognition. Convolutional neural network architectures may contain several layers of convolutional and subsampling filters that apply filters to certain portions of text to detect specific features (e.g. points of damage). A convolutional neural system includes a convolution operation. This multiplies each fragment of an image by filtering (e.g., matrixes) element-by?element, and then sums the results in a similar location in an output image.

“Recurrent neural network may propagate data forwards and backwards from later processing stages to early processing stages. Recurrent neural networks have the ability to store and process information from previous computations, as well as processing information sequences. Recurrent neural networks could also have a “memory”

“In certain embodiments, the machine-learning model could be a random forest classification. A random forest classifier uses an ensemble learning method to classification. It creates multiple decision trees (e.g. hundreds to thousands) and then trains them to make classification decisions based upon input data. The outputs of a random forest classifier are calculated by combining the decisions from multiple decision trees. In some embodiments, different decision trees of the random forest classifier can be trained using different parts or the entire training dataset. Each decision tree could be a predictive model that makes observations about input data to draw conclusions about the input data (represented by branches of the decision trees). Each decision tree could be trained to identify a digital design for an alignmenter. A training algorithm, such as feature bagging, which selects a random subset from the features at each candidate split during the learning process, may be used to train the random forest classifier. A trained random forest classifier can be used to determine why a classification was made. This is possible by using processing logic, or following the branches of one or more decision trees that led to the classification decision.

“In certain embodiments, the machine-learning model could be an XGBoost classification. An XGBoost classification is a implementation of a gradient boosted deci tree. Other gradient-boosted decision trees can be used in other ways to implement the machine learning algorithm. Boosting, an ensemble technique that adds new models to correct errors in existing models, is known as boosting. The models are added in a sequential fashion until they become obsolete. Gradient boosting refers to the creation of new models that can predict errors or residuals from existing models. The results of several models are combined to form a final prediction. Gradient boosting is a method that minimizes the loss from adding new models. The machine-leaning model could be a logistic regression model in some instances.

“In embodiments where a gradient boosted decision tree or random forest classifier (e.g. XGBoost) is trained on characteristics extracted digitally from dental arches and aligners, the machine-learning model can be trained to express joint effect of these characteristics and identify aligners most likely to be damaged.

“In certain embodiments, the machine-learning model can be periodically retrained with updated training datasets. As new patients are being treated, data may continue to be generated about manufactured aligners. Processing logic can repeat the training of the machine-learning model on a periodic or ongoing basis, such as every six months. The machine learning model can be retrained regularly to reflect new techniques and/or methods. This could include updated software or updated manufacturing flows. Some embodiments allow for continuous or ongoing training of the machine-learning model based on continuous data inflow. Different machine learning models may be trained in different ways for aligners made from different materials and manufactured using different manufacturing processes. A first machine learning model could be trained to predict possible points of damage to an aligner manufactured by thermoforming it on a mold. A second machine learning model might be trained to predict likely point of damage to an aligner directly printed using 3D printing, or other rapid prototyping methods.

“FIG. “FIG. Processing logic on a computing device performs one or more of the operations in method 230. Processing logic can include hardware (e.g. circuitry, dedicated logic or microcode), as well as software. Software (e.g. instructions executed by a processor device), firmware, and/or a combination thereof. A processing device that executes the aligner design analysis module 1450 in FIG. may perform one or more of the operations of method 230. 14. The method 230 can be used for any unique aligner in a patient’s treatment plan or at key points of the treatment plan. Method 230 may also include operations that can be performed in block 104 of FIG. 1A.”

“At block232 of method230, processing logic can perform an analysis of a digital design for an orthodontic aligner (e.g. a polymeric aligner) by using a trained machine-learning model. This model may have been trained according to method 200. The analysis of the digital design may be performed using the trained model. This could include applying (block 234) a digital design to the model. The trained machine-learning model can also be used to analyze the digital design for the aligner. This may involve generating (block 236) an output that indicates whether the aligner’s digital design includes any of the probable points of damage. The output of the trained model could identify the location of any damage if the digital design of an aligner contains the one or multiple points of potential damage. Additionally, the output of the trained model could include suggestions for corrective actions. Alternativly, the output from the trained machine-learning model can be input to a second system or module (e.g. another trained machine-learning model) with the digital design for the aligner. Based on the digital design and predicted damage points, the further system or module might recommend corrective actions.

“After the trained machine-learning model has determined that one or several points of probable damage have been predicted, the digital digital design of an aligner may be processed to include the one or two probable points. A numerical simulation of the digital polymeric aligner may also be performed to confirm that the one or multiple points of possible damage are present in certain embodiments. Any of the numerical simulations discussed herein may be used for the numerical simulation. The numerical simulation can be used to simulate the removal of an orthodontic aligner from a mold for a patient’s dental arch or simulate loading around weak spots within the orthodontic aligner. The numerical simulation can be more computationally costly than the processed digital model of an aligner. It also requires much more resources than the model created using the trained machine-learning model. The trained machine-learning model can be used to first process the digital model of an aligner, then limit the use of the numerical simulator to testing digital models for aligners for which it predicted a point. This will reduce resource consumption (e.g. memory and/or processor utilization). In some cases, the trained machine-learning model can determine the existence of one or more points of potential damage but not the exact location. The digital models of aligners that the trained machine-learning model predicted would be damaged using numerical simulation may be processed to identify the location of such points. Corrective actions may also be taken in certain embodiments.

“FIG. “FIG. Damage during clinical use, shipping damage, and so forth are examples of post-manufacturing damages. Method 200 may have been used to train the machine learning model. The processing logic on a computing device performs one or more of the operations in method 240. Processing logic can include hardware (e.g. circuitry, dedicated logic or microcode), as well as software. Software (e.g. instructions executed by a processor device), firmware, and/or a combination thereof. A processing device that executes the aligner design analysis module 1450 in FIG. may perform one or more of the operations in method 240. 14. The method 240 can be used for any unique aligner in a patient’s treatment plan or at key points of the treatment plan. Method 230 may also include operations that can be performed in block 104 of FIG. 1A.”

“At block 242 in method 240 processing logic may extract geometrical, treatment-related characteristics and/or clinical characters from a digital model of an orthodontic aligner in accordance with method 200. Processing logic may choose a subset from the characteristics at block 244. The selected subset of characteristics may be the same as those used to train the machine-learning model. Block 246, processing logic might generate an embedding to the digital design based upon the subset.

“At block 248, processingLogic processes data from the digital orthodontic aligner design using the trained machine-learning model. In an embodiment, the data from the digital design could include the embedding generated in block 246. Alternately, or in addition, data from the digital designer may include a digital three-dimensional model of the aligner, or a digital three-dimensional model of a mold or dental arch to be used for manufacturing the aligner. Alternately, or in addition, data from the digital designer may contain one or more height maps. These are created by projecting the three-dimensional digital model of the aligner or dental arch onto one or several planes.

“Block 250 is where the trained machine-learning model outputs a probability of damage to the aligner as a result of manufacturing or later use. This probability can be anywhere from 0 to 1. A 1 represents a 100% chance of the aligner being damaged, while a 0 indicates that there is 0% chance.

“In one embodiment, block 252 of the machine learning model outputs information that identifies the likelihood that certain points or locations on the orthodontic aligner will become damaged. A separate probability value may be output for each point on the orthodontic aligner, such as 0-1.

“In one embodiment, block 254 processing logic determines if the probability that the orthodontic aligner is damaged is below a threshold (or whether all points of the orthodontic aligner are damaged are below this threshold). If the probability that an orthodontic aligner will be damaged falls below the first threshold, then the process continues to block 254. This allows the algorithm to determine if the aligner is low-risk. FIG. As in FIG. 1B, an initial manufacturing flow for low-risk aligners can then be chosen for the aligner.”

Block 254 determines that the probability of an orthodontic aligner being damaged is greater than the first threshold. The method then moves to block 256. Block 256 is where processing logic determines if the probability that the orthodontic aligner will be damaged exceeds a second threshold. The second threshold could be higher than the first. The first threshold could be 0.2%, 0.5% or 1%. Or it may be 2%, 1% and 2%. The second threshold can be 15%, 20%. 25%. 30%. 40%. The method will continue to block 260 if the likelihood of the aligner being damaged exceeds the second threshold. The method will continue to block 258 otherwise.

“At block 258, it may be determined that the aligner meets standard risk requirements. As shown in FIG. As in FIG. 1B, an additional manufacturing flow may be chosen for standard risk aligners.

“At block 266, processing logic determines that the alignment is high-risk. As shown in FIG. As shown in FIG. 1B, an additional manufacturing flow may be used for high-risk aligners. One embodiment of block 262 processing logic might output a notification containing a location of at most one point and a probability for damage that is at or above the second threshold. This notification could be generated, for example, when the machine learning model outputs data that indicates the locations of points on an aligner and the probabilities of those points being damaged.

“In some embodiments (as shown in FIG. “In some embodiments, as shown in FIG. 2C, there are three possible classifications for an aligner. These are based on the likelihood that the aligner will become damaged during manufacturing or afterwards. These could include a low, medium, or standard risk classification and a high-risk classification. Other embodiments may place aligners in a binary classification. This includes standard risk (or no damage forecast) and high risk or damage prediction. Blocks 254 and256 can be skipped in such embodiments.

“FIG. 2D shows a flow diagram of a method 264 for determining if an orthodontic aligner in a set associated with a treatment program for a patient will become broken (e.g.) during or after the manufacturing of the set. This is in accordance to one embodiment. Method 200 may have been used to train the machine learning model. A computing device’s processing logic performs one or more of the operations in method 264. Processing logic can include hardware (e.g. circuitry, dedicated logic or microcode), as well as software. Software (e.g. instructions executed by a processor device), firmware, and/or a combination thereof. A processing device that executes the aligner design analysis module 1450 in FIG. may perform one or more of the operations in method 264. 14. The method 264 can be used for any treatment plan or the upper and lower dental arch for any treatment plan.

“At block 266, processing logic determines an alignment set consisting of digital designs for orthodontic aligners that are associated with a patient’s treatment plan. A treatment plan might divide the treatment into stages. Each stage may require a different aligner. One treatment plan can include as many stages as necessary and may also include digital designs for aligners. For example, 50 stages may be included. Separate digital designs may be created for the upper and bottom dental arches. One embodiment of an aligner set contains all digital designs for the upper or lower dental archs associated with a patient’s treatment plan. The aligner set can include all digital designs for the upper and lower dental archs associated with a treatment plan.

“At block 268, processing may extract geometrical characteristics, treatment-related characteristics and/or clinical characteristics. This is done from the digital design for each aligner. It is set in accordance with method 200. Processing logic can select a subset from the characteristics at block 270. A subset of the characteristics may be selected to correspond to the same characteristics used in training the machine learning model. Block 272 may allow processing logic to generate an embedding for each design in the aligner sets based on the relevant subset of characteristics.

“At block 274, processing logic processes the data from the digital designs for the orthodontic aligners using a trained machine learning model. In an embodiment, the data from the digital designs could include embeddings that were generated at block 246, Alternately, or in addition, data from digital designs could include digital three-dimensional models of aligners, digital three-dimensional models of a mold or dental arch to be used for manufacturing aligners, and digital three-dimensional models of these models. Alternately, or in addition, data from digital designs could include one or several height maps. These are created by projecting three-dimensional digital models of the aligner or dental arch onto one or multiple planes.

“Block 276, the trained machine-learning model outputs, each digital design of an alignmenter in the set, a probability of damage to the aligner (e.g. during manufacturing or later use). A probability value can range from 0 to 1. 1 could indicate a 100% chance of the aligner being damaged, while 0 indicates a 0% chance.

“In one embodiment, the block 278 processing logic determines if the probability that any orthodontic aligner is damaged is below a threshold (or whether all points of the aligners are at risk of being damaged). If the probability of any orthodontic aligners being damaged falls below the first threshold, then the block 278 processing logic continues, and it may be determined that the set is low-risk. As shown in FIG. As in FIG. 1B, an initial manufacturing flow for a low-risk aligner set can then be chosen for the aligner sets.”

“If the probability of at least one orthodontic aligner being broken is higher than the first threshold, then the block 278 block is completed. Block 282 is where processing logic determines if the probability that at least one orthodontic alignmenter will be damaged is at or beyond a second probability threshold. This means that the probability of any points being damaged on at least one orthodontic alignmenter are at or above this second threshold. The second threshold could be higher than the first threshold. The first threshold could be 0.2%, 0.5% or 1%. Or it may be 2%, 1% and 5%. The second threshold can be 15%, 20%. 25%. 30%. 40%. The method will continue to block 286 if the likelihood of any aligner being damaged in the set is greater than the second threshold. Otherwise, the method will continue to block 286.

“Block 284 may allow for the determination that the aligner set has a standard risk. As shown in FIG. As in FIG. 1B, an additional manufacturing flow may be used for standard risk aligner sets.

“At block 286, processing logic determines the aligner set to be high-risk. FIG. As in FIG. 1B, an additional manufacturing flow may be used for aligner sets with high risk components.

“Some embodiments (as shown in FIG. “In some embodiments, as shown in FIG. These could include a low, medium, or standard risk classification and a high-risk classification. Other embodiments may classify aligner sets into a binary classification. This includes standard risk (or no damage forecast) and high risk or damage prediction. Blocks 254 and 254 may not be used in such cases.

“FIG. 3A shows a flow diagram of a method 300 for performing analysis on a digital aligner design (e.g., polymeric aligner) using numerical simulating, according to one embodiment. Processing logic on a computing device performs one or more of the operations in method 300. Processing logic can include hardware (e.g. circuitry, dedicated logic or microcode), as well as software. Software (e.g. instructions executed by a processor device), firmware, and/or a combination thereof. A processing device that executes the aligner design analysis module 1450 in FIG. may perform one or more of the operations of method 300. 14. The method 300 can be used for any unique aligner in a patient’s treatment plan or at key points of the treatment plan. Method 300 may also include operations that can be performed in block 104 of FIG. 1A.”

“At block 302, processing logic can perform an analysis on the digital model of the aligner by numerical simulation. This is associated with the removal of the polymeric alignmenter from tooth-like and dental arch-like structures, such as the mold or patient’s dentition. Finite element, finite difference, finite volume, smoothed particle, combination of these methods or similar methods may be used in the numerical simulation. Finite element analysis, also known as finite element method, is a numerical method to solve partial differential equations. It can also be used to analyze the structural properties of aligners. In this example, the geometry of the structure (or aligner) is discretized to a few points or elements in a domain. This allows you to solve partial differential equations that characterize the constitutive relationships of the aligner material. The solutions are then explored in the finite dimensional functional area. Finite difference method can be used to describe a numerical method of solving differential equations. This involves approximating them using difference equations, and then calculating approximate values at discrete locations. A finite volume method can be used to represent and evaluate partial differential equations using algebraic equations. Finite volume method can also be used to calculate values (e.g. strain, stress, and force) at discrete points on the meshed geometry of a digital alignment. ?Finite volume? The small volume around each point in a mesh may be called?Finite volume? Meshfree methods can refer to methods that do not require interaction between nodes or points and all the neighbors. Meshfree methods don’t require connections between nodes in the simulation domain. There are meshfree methods such as the smoothed-particle Galerkin and hydrodynamics.

“Block 304 may contain processing logic that simulates one or more forces or displacements on the digital model of the aligner. These are related to the removal of the aligner (e.g. mold or dental arch) from the dental arch-like structural structure. Operations at blocks 306, 308., 310 and 312 can be used to simulate the one or more forces or displacements on the aligner’s digital design. Block 306 may be used to gather material property information (also known as material property information). Material properties can include the amount of stress or strain the material is capable of sustaining before it cracks, breaks, deforms, warps, etc. The Young’s Modulus is one example of a material property. Some embodiments allow the material properties to remain the same between digitally designed aligners. This is because they are made from the same material (e.g. polymeric). In some embodiments, material properties can be included in the configuration of the alignment design analysis module 1450.

“At block 308. Processing logic can gather the first geometry of an aligner using the digital design. The patient’s dental arch may determine the first geometry. You can generate the first geometry by creating a digital model (e.g. of a mold, or dental arch) of the aligner. The patient’s dental arch may be represented by the digital model of the arch-like dental structure. To create the digital design of an aligner, the digital model of the mold can be offset. The digital design of an aligner can include cavities that are designed to receive teeth (also called tooth-receiving cavity or caps). Attachments to the teeth and/or the patient may also be included in the digital design.

“At block310, processing logic can gather a second geometry from the dental archlike structure using a digital model (e.g. mold). A digital model of the dental arch-like structure can be created from data obtained from an intraoral scan and/or a treatment plan. The patient’s dental arch may be digitalized via scanning and used as a model to create the mold. The second geometry could include information about the patient’s dental arch such as their tooth size, shape, distance between teeth and attachments, the upper or lower dental arch, and so on.

Block 312, processing logic can simulate the removal the aligner with the one or two material properties and the first geometric from the dental arch-like structural having the second geometry. This is done by applying one or more loads to a set points on the digitally designed aligner. A series of partial differential equations may be used to simulate the numerical simulation. These equations model the application of one or more loads (e.g. forces and/or displacements), to the aligner with the material properties and to the first geometry in order to remove it from the dental arch-like structure that has the second geometry. The partial differential equations can calculate a stress value or strain value at each of the points on the aligner’s digital design. A determination of whether the point is likely to sustain damage may be made on the basis of this stress value or strain value. Elastostatic and elastodynamic partial equations can be used to calculate strain or stress states in the digital aligner design. This allows for the prediction of breakage, warpage or deformation. High polymeric strains/stresses can cause crack initiation or breakage as well as warpage and deformation in polymeric aligners. These partial differential equations can be described as:

“Find ui (u? 2) Such that:

“With boundary conditions:nu I(x.t)=u I g(x.t) at x???? u i\n?ij n j =t i(x,t) at x ??? t i?

“And initial conditions:nu I(x,0),=u i0 (x) at x???? u inv.i(x.0)=v. i0(x).????? u i”

Summary for “Aligner damage prediction, mitigation”

“In some embodiments processing logic selects between two manufacturing flows. The operations of block 112 are skipped. The method proceeds from block 114 to block 120.”

“Euclidean distance (between the points and circle);”

“At block 258, it may be determined that the aligner meets standard risk requirements. As shown in FIG. As in FIG. 1B, an additional manufacturing flow may be chosen for standard risk aligners.

“Block 284 may allow for the determination that the aligner set has a standard risk. As shown in FIG. As in FIG. 1B, an additional manufacturing flow may be used for standard risk aligner sets.

“At block 286, processing logic determines the aligner set to be high-risk. FIG. As in FIG. 1B, an additional manufacturing flow may be used for aligner sets with high risk components.

“Find ui (u? 2) Such that:

“With boundary conditions:nu I(x.t)=u I g(x.t) at x???? u i\n?ij n j =t i(x,t) at x ??? t i?

“And initial conditions:nu I(x,0),=u i0 (x) at x???? u inv.i(x.0)=v. i0(x).????? u i”

Click here to view the patent on Google Patents.

How to Search for Patents

A patent search is the first step to getting your patent. You can do a google patent search or do a USPTO search. Patent-pending is the term for the product that has been covered by the patent application. You can search the public pair to find the patent application. After the patent office approves your application, you will be able to do a patent number look to locate the patent issued. Your product is now patentable. You can also use the USPTO search engine. See below for details. You can get help from a patent lawyer. Patents in the United States are granted by the US trademark and patent office or the United States Patent and Trademark office. This office also reviews trademark applications.

Are you interested in similar patents? These are the steps to follow:

1. Brainstorm terms to describe your invention, based on its purpose, composition, or use.

Write down a brief, but precise description of the invention. Don’t use generic terms such as “device”, “process,” or “system”. Consider synonyms for the terms you chose initially. Next, take note of important technical terms as well as keywords.

Use the questions below to help you identify keywords or concepts.

What is the purpose of the invention Is it a utilitarian device or an ornamental design?
Is invention a way to create something or perform a function? Is it a product?
What is the composition and function of the invention? What is the physical composition of the invention?
What’s the purpose of the invention
What are the technical terms and keywords used to describe an invention’s nature? A technical dictionary can help you locate the right terms.

2. These terms will allow you to search for relevant Cooperative Patent Classifications at Classification Search Tool. If you are unable to find the right classification for your invention, scan through the classification’s class Schemas (class schedules) and try again. If you don’t get any results from the Classification Text Search, you might consider substituting your words to describe your invention with synonyms.

3. Check the CPC Classification Definition for confirmation of the CPC classification you found. If the selected classification title has a blue box with a “D” at its left, the hyperlink will take you to a CPC classification description. CPC classification definitions will help you determine the applicable classification’s scope so that you can choose the most relevant. These definitions may also include search tips or other suggestions that could be helpful for further research.

4. The Patents Full-Text Database and the Image Database allow you to retrieve patent documents that include the CPC classification. By focusing on the abstracts and representative drawings, you can narrow down your search for the most relevant patent publications.

5. This selection of patent publications is the best to look at for any similarities to your invention. Pay attention to the claims and specification. Refer to the applicant and patent examiner for additional patents.

6. You can retrieve published patent applications that match the CPC classification you chose in Step 3. You can also use the same search strategy that you used in Step 4 to narrow your search results to only the most relevant patent applications by reviewing the abstracts and representative drawings for each page. Next, examine all published patent applications carefully, paying special attention to the claims, and other drawings.

7. You can search for additional US patent publications by keyword searching in AppFT or PatFT databases, as well as classification searching of patents not from the United States per below. Also, you can use web search engines to search non-patent literature disclosures about inventions. Here are some examples:

Add keywords to your search. Keyword searches may turn up documents that are not well-categorized or have missed classifications during Step 2. For example, US patent examiners often supplement their classification searches with keyword searches. Think about the use of technical engineering terminology rather than everyday words.
Search for foreign patents using the CPC classification. Then, re-run the search using international patent office search engines such as Espacenet, the European Patent Office’s worldwide patent publication database of over 130 million patent publications. Other national databases include:

European Patent Office (EPO) provides esp@cenet to access a network of Europe’s patent databases with access to machine translation of European patents.
Japan Patent Office (JPO) – with access to machine translations of Japanese patents.
World Intellectual Property Organization (WIPO) offers PATENTSCOPE with a full-text search of published international patent applications and machine translations for some documents, as well as a list of international patent databases.
Korean Intellectual Property Rights Information Service (KIPRIS)
State Intellectual Property Office (SIPO) with machine translation of Chinese patents.
Other International Intellectual Property Offices with online patent databases include Australia, Canada, Denmark, Finland, France, Germany, Great Britain, India, Israel, Netherlands, Norway, Sweden, Switzerland, and Taiwan.

Search non-patent literature. Inventions can be made public in many non-patent publications. It is recommended that you search journals, books, websites, technical catalogs, conference proceedings, and other print and electronic publications.

To review your search, you can hire a registered patent attorney to assist. A preliminary search will help one better prepare to talk about their invention and other related inventions with a professional patent attorney. In addition, the attorney will not spend too much time or money on patenting basics.

Download patent guide file – Click here

(888) 723 8320

Software Technology for “Aligner damage prediction, mitigation”

3D Printing – Yuxiang Wang, Rohit Tanugula, Reza Shirazi AGHJARI, Andrew JANG, Chunhua Li, Jun Sato, Luyao Cai, Viktoria Medvinskaya, Arno Kukk, Andrey Cherkas, Anna Akopova, Kangning Su, Align Technology Inc

Search Patent for “Aligner damage prediction, mitigation”

Abstract for “Aligner damage prediction, mitigation”

Background for “Aligner damage prediction, mitigation”

Summary for “Aligner damage prediction, mitigation”

Get A Free Consultation

Address

Phone

E-mail

Quick Links

Software Technology for “Aligner damage prediction, mitigation”

3D Printing – Yuxiang Wang, Rohit Tanugula, Reza Shirazi AGHJARI, Andrew JANG, Chunhua Li, Jun Sato, Luyao Cai, Viktoria Medvinskaya, Arno Kukk, Andrey Cherkas, Anna Akopova, Kangning Su, Align Technology Inc

Search Patent for “Aligner damage prediction, mitigation”

Abstract for “Aligner damage prediction, mitigation”

Background for “Aligner damage prediction, mitigation”

Summary for “Aligner damage prediction, mitigation”

Get A Free Consultation

Submit your email and get to download this post as a PDF file

Address

Phone

E-mail

Quick Links

Submit your email and get to
download this post as a PDF file