Neural network-based generation and placement of tooth restoration dental appliances
Techniques are described for automating the design of dental restoration appliances using neural networks. An example computing device receives transform information associated with a current dental anatomy of a dental restoration patient, provides the transform information associated with the current dental anatomy of the dental restoration patient as input to a neural network trained with transform information indicating placement of a dental appliance component with respect to one or more teeth of corresponding dental anatomies, the dental appliance being used for dental restoration treatment for the one or more teeth, and executes the neural network using the input to produce placement information for the dental appliance component with respect to the current dental anatomy of the dental restoration patient.
1. A computing device comprising:
an input interface configured to receive one or more three-dimensional (3D) tooth meshes associated with a current dental anatomy of a dental restoration patient and a 3D component mesh representing a generated geometry for a dental appliance component; and
a neural network engine configured to:
provide the one or more 3D tooth meshes and the 3D component mesh received by the input interface as inputs to a neural network trained with training data comprising ground truth dental appliance component geometries and 3D tooth meshes of corresponding dental restoration cases; and
execute the neural network using the provided inputs to produce an updated model of the dental appliance component with respect to the current dental anatomy of the dental restoration patient.
2. The computing device of claim 1 , wherein the neural network is a generative adversarial network (GAN) comprising a generator network and a discriminator network,
and wherein to execute the neural network to produce the updated model of the dental appliance component with respect to the current dental anatomy of the dental restoration patient, the neural network engine is configured to provide the one or more 3D tooth meshes and the 3D component mesh received by the input interface as input to the generator network of the GAN.
3. The computing device of claim 2 , wherein the neural network engine is configured to train the GAN by:
executing the generator network of the GAN to generate an updated geometry of the dental appliance component using the one or more 3D tooth meshes and the 3D component mesh;
providing the updated geometry of the dental appliance component, the one or more 3D tooth meshes, and the 3D component mesh as inputs to the discriminator network of the GAN; and
executing the discriminator network to output a probability of the updated geometry representing a ground truth geometry.
4. The computing device of claim 1 , wherein the neural network is further trained with placement information of the ground truth dental appliance component geometries as part of the training data.
5. The computing device of claim 1 , wherein the dental appliance component comprises one or more of:
a mold parting surface,
a gingival trim,
a door,
a window,
a facial ribbon,
an incisal ridge,
a lingual shelf,
a diastema matrix,
a case frame,
a part label;
a door hinge,
a door snap,
a door vent,
a snap clamp, or
a center clip.
6. A method comprising:
receiving, at an input interface, one or more three-dimensional (3D) tooth meshes associated with a current dental anatomy of a dental restoration patient and a 3D component mesh representing a generated geometry for a dental appliance component;
providing, by a neural network engine communicatively coupled to the input interface, the one or more 3D tooth meshes and the 3D component mesh received by the input interface as inputs to a neural network trained with training data comprising ground truth dental appliance component geometries and 3D tooth meshes of corresponding dental restoration cases; and
executing, by the neural network engine, the neural network using the provided inputs to produce an updated model of the dental appliance component with respect to the current dental anatomy of the dental restoration patient.
7. The method of claim 6 , wherein the neural network is a generative adversarial network (GAN) comprising a generator network and a discriminator network, and wherein executing the neural network to produce the updated model of the dental appliance component with respect to the current dental anatomy of the dental restoration patient comprises providing, by the neural network engine, the one or more 3D tooth meshes and the 3D component mesh received by the input interface as input to the generator network of the GAN.
8. The method of claim 7 , further comprising training, by the neural network engine the GAN by:
executing the generator network of the GAN to generate an updated geometry of the dental appliance component using the one or more 3D tooth meshes and the 3D component mesh;
providing the updated geometry of the dental appliance component, the one or more 3D tooth meshes, and the 3D component mesh as inputs to the discriminator network of the GAN; and
executing the discriminator network to output a probability of the updated geometry representing a ground truth geometry.
9. The method of claim 6 , wherein the neural network is further trained with placement information of the ground truth dental appliance component geometries as part of the training data.
10. The method of claim 6 , wherein the dental appliance component comprises one or more of:
a mold parting surface,
a gingival trim,
a door,
a window,
a facial ribbon,
an incisal ridge,
a lingual shelf,
a diastema matrix,
a case frame,
a part label;
a door hinge,
a door snap,
a door vent,
a snap clamp, or
a center clip.
11. A non-transitory computer-readable storage medium encoded with instructions that, when executed, cause one or more processors of a computing system to:
receive, via an input interface of the computing system, one or more three-dimensional (3D) tooth meshes associated with a current dental anatomy of a dental restoration patient and a 3D component mesh representing a generated geometry for a dental appliance component;
provide the one or more 3D tooth meshes and the 3D component mesh received via the input interface as inputs to a neural network trained with training data comprising ground truth dental appliance component geometries and 3D tooth meshes of corresponding dental restoration cases; and
execute the neural network using the provided inputs to produce an updated model of the dental appliance component with respect to the current dental anatomy of the dental restoration patient.
12. The non-transitory computer-readable storage medium of claim 11 , wherein the neural network is a generative adversarial network (GAN) comprising a generator network and a discriminator network,
and wherein the instructions that cause the one or more processors to execute the neural network to produc the updated model of the dental applicance component with respect to the current dental anatomy of the dental restoration patient comprises instructions that, when executed, cause the one or more processors to provide the one or more 3D tooth meshes and the 3D component mesh received via the input interface as input to the generator network of the GAN.
13. The non-transitory computer-readable storage medium of claim 12 , further encoded with instructions that, when executed by the one or more processors, train the GAN by:
executing the generator network of the GAN to generate an updated geometry of the dental appliance component using the one or more 3D tooth meshes and the 3D component mesh;
providing the updated geometry of the dental applicance component, the one or more 3D tooth meshes, and the 3D component mesh as inputs to the discriminator network of the GAN; and
executing the discrominator network to outpout a probability of the updated geometry representing a ground truth geometry.
14. The non-transitory computer-readable storage medium of claim 11 , wherein the neural network is further trained with placement information of the ground truth dental appliance component geometries as part of the training data.
15. The non-transitory computer-readable storage medium of claim 11 , wherein the dental appliance component comprises one or more of:
a mold parting surface,
a gingival trim,
a door,
a window,
a facial ribbon,
an incisal ridge,
a lingual shelf,
a diastema matrix,
a case frame,
a part label;
a door hinge,
a door snap,
a door vent,
a snap clamp, or
a center clip.