Method, system and devices for instant automated design of a customized dental object
Embodiments of the present invention provide a method, system, devices and software which provide customized, clinically relevant designs, e.g. treatment plans/solutions for at least single tooth replacement therapy in real-time as needed to enable immediate confirmation of the validity and execution of patient-specific treatment solutions. A Machine Learning algorithm is used in a method or system for computer implementation of automatic restoration design. The restoration design includes the design of restorative elements such as crowns and abutments. Software is provided for carrying out the methods when executed on a digital processor. Manufacturing of the restorative elements is also included.
1 . A computer-implemented method for providing a representation of a 3D shape of a restorative dental object for a patient, the method comprising:
training, by one or more computing devices and using a plurality of preexisting treatment 3D data sets, a trained machine learning system in an end-to-end manner, wherein the plurality of preexisting treatment 3D data sets comprises 3D images of patients' dentitions and 3D shapes of representations of patients' restorative dental objects and the trained machine learning system includes a convolutional neural network,
inputting a representation of a 3D scan of at least one portion of a patient's dentition to the trained machine learning system, the representation of a 3D scan defining at least one implant position of an implant, the trained machine learning system being installed on one or more computing devices,
accessing a set of parameters used to generate a parametric model, each parameter of the set of parameters having a corresponding range, and each corresponding range having a set of bins having a limited resolution, wherein the set of parameters relate to a characteristic of a tooth surface anatomy, a tooth dentition, or a restoration type,
estimating, by the convolutional neural network, a correct bin of the set of bins for each corresponding range of each parameter of the set of parameters, to obtain optimized parameters for a restorative dental object to correspond with the 3D scan of the at least one portion of the patient's dentition, and
identifying, using the trained machine learning system and based on the optimized parameters of the parametric model, the representation of the 3D shape of the restorative dental object for the implant in an end-to-end manner.
2 . A computer-implemented method according to claim 1 , comprising applying computer based end-to-end machine learning models.
3 . A computer-implemented method according to claim 1 , wherein an output of the trained machine learning system is a representation of a 3D shape that is expressed as a set of parameters which defines an abutment or crown.
4 . A computer-implemented method according to claim 3 , wherein the set of parameters is a part of a parametric modelling.
5 . A computer-implemented method according to claim 1 wherein the trained machine learning system is adapted to use a discriminative machine learning algorithm.
6 . A computer-implemented method according to claim 1 , wherein the restorative dental object is a crown, or an abutment for the implant.
7 . A computer-implemented method according to claim 1 , wherein point samples are extracted on a 3D surface of a jaw scan geometry adapted for input to the convolutional neural network.
8 . A computer-implemented method according to claim 1 , for a new clinical case of an implant-based restoration, the method comprising receiving input of a user design preference, patient jaw 3D scans, and an implant position and orientation from 3D scan data of a feature location object.
9 . A computer-implemented method according to claim 1 , further comprising manufacturing of a customized restorative element made from the representation of a 3D shape of the restorative dental object, the restorative dental object being a custom abutment or crown for an implant.
10 . A computer-implemented system for providing a representation of a 3D shape of a restorative dental object for a patient, the system comprising:
means for training, by one or more computing devices and using a plurality of preexisting treatment 3D data sets, a trained machine learning system in an end-to-end manner, wherein the plurality of preexisting treatment 3D data sets comprises 3D images of patients' dentitions and 3D shapes of representations of patients' restorative dental objects and the trained machine learning system includes a convolutional neural network,
means for inputting a representation of a 3D scan of at least one portion of a patient's dentition, a representation of a 3D scan defining at least one implant position of an implant, the inputting being to the trained machine learning system installed on one or more computing devices,
means for accessing a set of parameters used to generate a parametric model, each parameter of the set of parameters having a corresponding range, and each corresponding range having a set of bins having a limited resolution, wherein the set of parameters relate to a characteristic of a tooth surface anatomy, a tooth dentition, or a restoration type,
means for estimating, by the convolutional neural network, a correct bin of the set of bins for each corresponding range of each parameter of the set of parameters, to obtain optimized parameters for a restorative dental object to correspond with the 3D scan of the at least one portion of the patient's dentition, and
means for identifying, using the trained machine learning system and based on the optimized parameters of the parametric model, the representation of the 3D shape of the restorative dental object for the implant in an end-to-end manner.
11 . A computer-implemented system according to claim 10 , comprising means for applying computer based end-to-end machine learning models.
12 . A computer-implemented method for training a machine learning system, the machine learning system being installed on one or more computing devices, the method comprising:
training the machine learning system offline with a plurality of preexisting treatment 3D data sets in an end-to-end manner, the plurality of preexisting treatment 3D data sets comprising 3D images of patients' dentitions and 3D shapes of representations of patients' restorative dental objects, wherein the plurality of preexisting treatment 3D data sets comprises 3D images of patients' dentitions and 3D shapes of representations of patients' restorative dental objects and the machine learning system includes a convolutional neural network, the machine learning system used to generate a 3D shape of a restorative dental object, wherein the 3D shape was generated by a parametric model that is defined by a set of parameters,
wherein a range of each parameter value of the set of parameters is divided among a set of bins each having a limited range thus providing a desired resolution, and the machine learning system, after training is adapted to estimate a correct bin where a specific parameter belongs to obtain optimized parameters for a restorative dental object,
wherein a parameter of the set of parameters relates to a characteristic of a tooth surface anatomy.
13 . A computer-implemented system for training a machine learning system, the machine learning system being installed on one or more computing devices, the computer-implemented system comprising:
means for training the machine learning system offline with a plurality of preexisting treatment 3D data sets in an end-to-end manner, the plurality of preexisting treatment 3D data sets comprising 3D images of patients' dentitions as inputs to the machine learning system and 3D shapes of representations of patients' restorative dental objects as outputs of the machine learning system, wherein the plurality of preexisting treatment 3D data sets comprises 3D images of patients' dentitions and 3D shapes of representations of patients' restorative dental objects and the machine learning system includes a convolutional neural network, the machine learning system used to generate a 3D shape of a restorative dental object, wherein the 3D shape was generated by a parametric model that is defined by a set of parameters,
wherein a range of each parameter value of the set of parameters is divided among a set of bins each having a limited range thus providing a desired resolution, and the machine learning system, after training, is adapted to estimate a correct bin where a specific parameter belongs to obtain optimized parameters for a restorative dental object,
wherein a parameter of the set of parameters relates to a characteristic of a tooth surface anatomy.