Systems and methods for generating multi-view synthetic dental radiographs for intraoral tomosynthesis
Systems and methods for generating multi-view synthetic dental radiographs for intraoral tomosynthesis. In some embodiments, the method includes generating or receiving two-dimensional (2D) projection images, manipulating pixel values contained in each of the plurality of 2D projection images, reconstructing a three-dimensional (3D) image space from information available in the plurality of 2D projection images, the 3D image space comprising voxel values, manipulating the voxel values in the 3D image space using one or more tunable weighting algorithms that can be adjusted to emphasize one or more features of interest of each image in the 3D image space, generating a plurality of synthetic dental radiographs from multiple views using information available in the 3D image space, and displaying one or more of the synthetic dental radiographs. In some embodiments, the system includes a display in communication with an image processing system comprising one or more processors for performing the method.
1. A method for generating one or more multi-view synthetic dental radiographs using a chain of interdependent image processing steps, the method comprising:
generating or receiving a plurality of two-dimensional (2D) projection images;
manipulating pixel values contained in each of the plurality of 2D projection images by:
segmenting artifact-producing features in each of the plurality of 2D projection images; and
assigning pixel values to segmented regions by inward interpolation from surrounding pixel values for each of the plurality of 2D projections, thereby reducing imaging artifacts caused by the artifact-producing features;
reconstructing a three-dimensional (3D) image space from information available in the plurality of 2D projection images, the 3D image space comprising voxel values;
manipulating the voxel values in the 3D image space using one or more tunable weighting algorithms that can be adjusted to emphasize one or more features of interest of each image in the 3D image space;
generating a plurality of synthetic dental radiographs from multiple views using information available in the 3D image space;
fusing the artifact-producing features segmented from the plurality of 2D projection images into the plurality of synthetic dental radiographs to produce a set of final synthetic dental radiographs that comprise the artifact-producing features; and
displaying one or more synthetic dental radiographs of the set of final synthetic dental radiographs that comprise the artifact-producing features;
wherein the artifact-producing features are segmented in each of the plurality of 2D projection images prior to reconstructing the 3D image space.
2. The method of claim 1 , further comprising manipulating pixel values in the plurality of synthetic dental radiographs.
3. The method of claim 1 , wherein the artifact-producing features are formed by metal objects in a subject's teeth.
4. The method of claim 3 , wherein the metal objects comprise screws, fillings, surgical implants, and/or amalgam.
5. The method of claim 1 , wherein the 3D image space is generated using analytical or iterative reconstruction algorithms customized to intraoral tomosynthesis.
6. The method of claim 1 , further comprising identifying and/or enhancing features of interest, using filters and/or deep-learning techniques, features of interest including dental caries or dental fractures in the 3D image space.
7. The method of claim 1 , wherein the plurality of synthetic dental radiographs represent a range of viewing perspectives that may or may not depict a same angle from which the original x-ray projections were acquired.
8. The method of claim 1 , wherein different weighting algorithms are applied to the 3D image space to enhance features of interest, such as caries or fractures, with each weighting algorithm producing a unique set of multi view synthetic dental radiographs.
9. The method of claim 8 , wherein enhancing fractures comprises emphasizing higher-frequency components of an image and enhancing caries comprising emphasizing lower-frequency components of the image.
10. The method of claim 1 , further comprising optimizing the plurality of synthetic dental radiographs using filters customized to dental imaging.
11. The method of claim 1 , wherein reconstructing the 3D image space comprises using one or more of the following reconstruction techniques: filtered back projection (FBP), simultaneous iterative reconstruction technique (SIRT), simultaneous algebraic reconstruction technique (SART), or maximum likelihood expectation maximization (MLEM).
12. A method for generating one or more multi-view synthetic dental radiographs, the method comprising:
positioning an intraoral x-ray detector in a subject's mouth;
determining a position of the intraoral x-ray detector relative to one or more x-ray source;
capturing one or more x-ray projections from multiple viewing angles relative to the intraoral x-ray detector;
transferring the one or more x-ray projection images to one or more processors;
manipulating, by the one or more processors, pixel values contained in the one or more x-ray projection images by:
segmenting artifact-producing features in each of the plurality of 2D projection images; and
assigning pixel values to segmented regions by inward interpolation from surrounding pixel values for each of the plurality of 2D projections, thereby reducing imaging artifacts caused by the artifact-producing features;
reconstructing a 3D image space from information available in the one or more x-ray projection images, the 3D image space comprising voxel values;
manipulating the voxel values in the 3D image space using one or more tunable weighting algorithms that can be adjusted to highlight specific image features in each image in the 3D image space;
generating a plurality of synthetic dental radiographs from multiple views using information available in the 3D image space;
fusing the artifact-producing features segmented from the plurality of 2D projection images into the plurality of synthetic dental radiographs to produce a set of final synthetic dental radiographs that comprise the artifact-producing features; and
displaying one or more synthetic dental radiographs of the set of final synthetic dental radiographs that comprise the artifact-producing features;
wherein the artifact-producing features are segmented in each of the plurality of 2D projection images prior to reconstructing the 3D image space.
13. A system for generating one or more multi-view synthetic dental radiographs comprising:
an image processing system comprising one or more processors; and
a display in communication with the image processing system;
wherein the image processing system is configured to:
receive a plurality of two-dimensional (2D) projection images;
manipulate pixel values contained in each of the plurality of 2D projection images by:
segmenting artifact-producing features in each of the plurality of 2D projection images; and
assigning pixel values to segmented regions by inward interpolation from surrounding pixel values for each of the plurality of 2D projections, thereby reducing imaging artifacts caused by the artifact-producing features;
reconstruct a three-dimensional (3D) image space from information available in the plurality of 2D projection images, the 3D image space comprising voxel values;
manipulate the voxel values in the 3D image space using one or more tunable weighting algorithms that can be adjusted to emphasize features of interest of each image in the 3D image space;
generate a plurality of synthetic dental radiographs from multiple views using information available in the 3D image space;
fuse the artifact-producing features segmented from the plurality of 2D projection images into the plurality of synthetic dental radiographs to produce a set of final synthetic dental radiographs that comprise the artifact-producing features; and
display one or more synthetic dental radiographs of the set of final synthetic dental radiographs, which comprise the artifact-producing features, on the display;
wherein the artifact-producing features are segmented in each of the plurality of 2D projection images prior to reconstructing the 3D image space.
14. The system of claim 13 , wherein the artifact-producing features are formed by metal objects in a subject's teeth.
15. The system of claim 14 , wherein the metal objects comprise screws, fillings, surgical implants, and/or amalgam.
16. The system of claim 13 , wherein the 3D image space is generated using analytical or iterative reconstruction algorithms customized to intraoral tomosynthesis.
17. The system of claim 13 , wherein the image processing system is further configured to identify and/or enhance features of interest, using filters and/or deep-learning techniques, including dental caries or dental fractures in the 3D image space.
18. The system of claim 13 , wherein the plurality of synthetic dental radiographs represent a range of viewing perspectives that may or may not depict a same angle from which the original x-ray projections were acquired.
19. The system of claim 13 , wherein the image processing system is further configured to apply different weighting algorithms to the 3D image space to enhance features of interest, such as caries or fractures, with each weighting algorithm producing a unique set of multi-view synthetic dental radiographs.
20. The system of claim 19 , wherein the image processing system is configured to:
enhance fractures by emphasizing higher-frequency components of an image; and
enhance caries by emphasizing lower-frequency components of an image.
21. The system of claim 13 , wherein the image processing system is further configured to optimize the plurality of synthetic dental radiographs using filters customized to dental imaging.
22. The system of claim 13 , wherein the image processing system is configured to reconstruct the 3D image space using one or more of the following filter techniques: filtered back projection (FBP), simultaneous iterative reconstruction technique (SIRT), simultaneous algebraic reconstruction technique (SART), or maximum likelihood expectation maximization (MLEM).