Photo realistic rendering of smile image after treatment
A method may include: receiving facial image of the patient that depicts the patient's teeth; receiving a 3D model of the patient's teeth; determining color palette of the depiction of the patient's teeth; coding 3D model of the patient's teeth based on attributes of the 3D model; providing the 3D model, the color palette, and the coded 3D model to a neural network; processing the 3D model, the color palette, and the coded 3D model by the neural network to generate a processed image of the patient's teeth; simulating specular highlights on the processed image of the patient's teeth; and inserting the processed image of the patient's teeth into a mouth opening of the facial image.
1. A method of orthodontically treating a patient's teeth comprising:
receiving facial image of the patient that depicts the patient's teeth;
receiving a 3D model of the patient's teeth;
determining color palette of the depiction of the patient's teeth;
coding 3D model of the patient's teeth based on attributes of the 3D model;
providing the 3D model, the color palette, and the coded 3D model to a neural network;
processing the 3D model, the color palette, and the coded 3D model by the neural network to generate a processed image of the patient's teeth;
simulating specular highlights on the processed image of the patient's teeth; and
inserting the processed image of the patient's teeth into a mouth opening of the facial image.
2. The method of claim 1 , comprising:
forming spline at the edge of the inner lips to define the mouth opening of the facial image.
3. The method of claim 1 , comprising:
training the neural network using facial images of people that depict their teeth.
4. The method of claim 1 , comprising:
blurring the processed image of the patient's teeth.
5. The method of claim 4 , wherein the blurring occurs after inserting the processed image of the patient's teeth into the mouth opening of the facial image.
6. The method of claim 5 , wherein the blurring is alpha channel blurring.
7. The method of claim 1 , wherein generating the color palette comprises:
blurring the depiction of the patient's teeth from the facial image.
8. The method of claim 7 , wherein the blurring is a Gaussian blur.
9. The method of claim 1 , wherein the facial image is a 2D facial image.
10. The method of claim 1 , wherein coding the 3D model comprises coding a color channel of a plurality of pixels of a 2D rendering of the 3D model with attributes of the 3D model or the facial image.
11. The method of claim 10 , wherein the attributes are one or more of the brightness of the patient's teeth at each pixel location, the angle of the surface of the 3D model with respect to the facial plane at each pixel location, and the dental structure type of the 3D model at each pixel location.
12. The method of claim 11 , wherein the brightness of the patient's teeth location at each pixel location is determined based on the brightness of a blurred depiction of the patient's teeth from the facial image.
13. The method of claim 11 , wherein the dental structure is one or more of an identity of each tooth or the gingiva in the 3D model at each pixel location.
14. The method of claim 1 , wherein the processed image of the patient's teeth is a 2D image.
15. The method of claim 1 , wherein simulating specular highlights on the processed image of the patient's teeth includes:
identifying bright locations on the patent's teeth in the facial image;
determining the surface for a plurality of locations on the patient's teeth in the facial image;
identifying a set of normals at the bright locations; and
applying simulated specular highlights in the processed image based on the set of normals.
16. The method of claim 15 , wherein the set of normals are an average normal of the normals at the specular highlight locations.
17. The method of claim 15 , wherein the set of normals a plurality of the most common normals identified.
18. The method of claim 15 , wherein the set of normals includes an average normal plus or minus 1 degree, 2 degrees, 3 degrees, 4, degrees, or 5 degrees.
19. The method of claim 15 , wherein the set of normals includes each node of a multimodal distribution of normals.
20. The method of claim 15 , wherein the set of normals includes the normals at which the specular highlights occur.
21. The method of claim 15 , wherein applying the simulated specular highlights in the processed image based on the set of normals includes:
determining an average normal from the set of normals at the bright locations.
identifying second normals at each location of the teeth in the processed image;
determining the cosine similarity between the average normal and the second normals at each location;
determining a difference in brightness at each location of the teeth in the processed image based on the cosine similarity at each respective location; and
applying the difference in brightness to a brightness channel pixels of the processed image at each respective location.
22. The method of claim 15 , wherein determining the average normal from the set of normals at the bright locations includes using outlier rejection.
23. The method of claim 15 , further comprising:
determining the difference in brightness at each location based on a brightness, a specular highlight brightness, a relative shininess of the teeth, and the cosine similarity at each location.
24. The method of claim 23 , wherein determining the difference in brightness at each location based on a brightness a specular highlight brightness, a relative shininess of the teeth, and the cosine similarity at each location includes using a Blinn-Phong shading model.
25. The method of claim 15 , wherein the processed image uses the L*a*b* color space.
26. The method of claim 1 , further comprising:
brightening the patient's teeth in the processed image.
27. The method of claim 26 , wherein brightening the patient's teeth in the processed image comprises:
masking the patient's teeth by created a tooth mask;
increasing the luminance channel of the image in the tooth mask; and
increasing the blue-yellow color channel of the image in the tooth mask.
28. The method of claim 26 , wherein increasing the luminance channel of the image in the tooth mask includes increasing the value of a L* channel by 1.
29. The method of claim 26 , wherein increasing the blue-yellow color channel of the image in the tooth mask includes increasing the value of a b* channel by 3.
30. The method of claim 26 , further comprising:
converting the processed image to L*a*b* color space.