Automated detection, generation and/or correction of dental features in digital models

1. A non-transitory computer readable medium comprising instructions that, when executed by a processing device, cause the processing device to perform operations comprising:

generating a three-dimensional model of a dental site from scan data of the dental site, the three-dimensional model comprising a representation of a preparation tooth;

receiving or generating an image of the preparation tooth;

processing data from the image using a trained machine learning model that has been trained to identify margin lines of preparation teeth, wherein the trained machine learning model outputs a probability map comprising, for each pixel in the image, a probability that the pixel depicts a margin line;

determining, for each point of a plurality of points on the three-dimensional model that maps to a pixel in the image of the preparation tooth, a probability that the point depicts the margin line using the probability map;

computing the margin line by applying a cost function to the plurality of points on the three-dimensional model based on probabilities associated with the plurality of points; and

updating the three-dimensional model of the dental site by marking the margin line on the representation of the preparation tooth.

2. The non-transitory computer readable medium of claim 1 , wherein the image of the preparation tooth is generated by projecting at least a portion of the three-dimensional model onto a two-dimensional surface.

3. The non-transitory computer readable medium of claim 1 , wherein the scan data is intraoral scan data, and wherein the image of the preparation tooth is an intraoral image included in the intraoral scan data, and wherein the image comprises a height map.

4. The non-transitory computer readable medium of claim 1 ,

wherein the cost function selects points that together form a contour having a combined minimal cost, and wherein for each point a cost of the point is related to an inverse of the probability that the point depicts the margin line.

5. The non-transitory computer readable medium of claim 4 , the operations further comprising:

determining whether the combined minimal cost exceeds a cost threshold; and

responsive to determining that the combined minimal cost exceeds the cost threshold, determining that the computed margin line has an unacceptable level of uncertainty.

6. The non-transitory computer readable medium of claim 4 , the operations further comprising:

computing separate costs for different segments of the computed margin line;

determining that a segment of the computed margin line has a cost that exceeds a cost threshold; and

determining that the segment of the computed margin line has an unacceptable level of uncertainty.

7. The non-transitory computer readable medium of claim 6 , the operations further comprising:

highlighting the segment of the computed margin line having the unacceptable level of uncertainty in the three-dimensional model.

8. The non-transitory computer readable medium of claim 6 , the operations further comprising:

locking regions of the three-dimensional model comprising segments of the computed margin line having acceptable levels of uncertainty;

receiving a new intraoral image depicting the segment of the computed margin line with the unacceptable level of uncertainty; and

updating the three-dimensional model using the new intraoral image to output an updated three-dimensional model, wherein a first region comprising the segment of the computed margin line with the unacceptable level of uncertainty is replaced using information from the new intraoral image, and wherein locked regions of the three-dimensional model comprising the segments of the computed margin line having the acceptable levels of uncertainty are unchanged during the updating.

9. The non-transitory computer readable medium of claim 8 , wherein the scan data comprises a plurality of blended images of the dental site, wherein each blended image of the plurality of blended images is based on a combination of a plurality of images, and wherein receiving the new intraoral image comprises:

accessing a plurality of individual intraoral images used to generate at least some of the plurality of blended images;

identifying a subset of the plurality of individual intraoral images that depict the segment of the computed margin line; and

selecting the new intraoral image from the subset of the plurality of individual intraoral images, wherein the new intraoral image comprises an improved depiction of the margin line as compared to the image of the preparation tooth.

10. The non-transitory computer readable medium of claim 9 , the operations further comprising:

generating a plurality of different versions of the updated three-dimensional model, wherein each of the plurality of different versions is based on a different individual intraoral image from the subset of the plurality of individual intraoral images; and

receiving a user selection of a particular version of the updated three-dimensional model corresponding to the new intraoral image.

11. The non-transitory computer readable medium of claim 8 , the operations further comprising:

generating a projected image of the first region by projecting at least a portion of the updated three-dimensional model onto an additional two-dimensional surface;

processing data from the projected image using the trained machine learning model, wherein the trained machine learning model outputs an additional probability map comprising, for each pixel in the projected image, a probability that the pixel depicts the margin line; and

further updating the updated three-dimensional model of the dental site by marking the margin line in the first region based on the additional probability map.

12. The non-transitory computer readable medium of claim 1 , wherein the trained machine learning model further outputs an indication for at least a section of the margin line as to whether the section of the margin line depicted in the image is a high quality margin line or a low quality margin line.

13. The non-transitory computer readable medium of claim 1 , the operations further comprising:

determining that the margin line is indeterminate in at least one section of the margin line associated with the image;

processing data from the image or a new image generated from the three-dimensional model using a second trained machine learning model that has been trained to modify images of teeth, wherein the second trained machine learning model outputs a modified image; and

updating the three-dimensional model of the dental site using the modified image, wherein an updated three-dimensional model comprises an updated margin line with an increased level of accuracy.

14. The non-transitory computer readable medium of claim 13 , wherein the modified image comprises a plurality of pixels that are identified as part of the margin line.

15. The non-transitory computer readable medium of claim 13 , wherein the image comprises a depiction of an interfering surface that obscures the margin line, wherein at least a portion of the depiction of the interfering surface is removed in the modified image, and wherein a portion of the margin line that was obscured in the image is shown in the modified image.

16. The non-transitory computer readable medium of claim 15 , wherein the interfering surface comprises at least one of blood, saliva, soft tissue or a retraction material.

17. The non-transitory computer readable medium of claim 1 , wherein the image is a monochrome image that contains height information, and wherein an input to the machine learning model comprises the data from the image and additional data from a two-dimensional color image that lacks height information.

18. A system comprising:

a memory; and

a computing device operatively connected to the memory, the computing device configured to:

generate a three-dimensional model of a dental site from scan data of the dental site, the three-dimensional model comprising a representation of a preparation tooth;

receive or generate an image of the preparation tooth;

process data from the image using a trained machine learning model that has been trained to identify margin lines of preparation teeth, wherein the trained machine learning model outputs a probability map comprising, for each pixel in the image, a probability that the pixel depicts a margin line;

determine, for each point of a plurality of points on the three-dimensional model that maps to a pixel in the image of the preparation tooth, a probability that the point depicts the margin line using the probability map;

compute the margin line by applying a cost function to the plurality of points on the three-dimensional model based on probabilities associated with the plurality of points; and

update the three-dimensional model of the dental site by marking the margin line on the representation of the preparation tooth.

19. The system of claim 18 ,

wherein the cost function selects points that together form a contour having a combined minimal cost, and wherein for each point a cost of the point is related to an inverse of the probability that the point depicts the margin line.

20. The system of claim 18 , wherein the trained machine learning model is further to output an indication for at least a section of the margin line as to whether the section of the margin line depicted in the image is a high quality margin line or a low quality margin line.