Metal artifact reduction in computed tomography
A computer-implemented method for modifying X-ray projection images of a subject region includes: generating a set of combined two-dimensional (2D) projections of a subject region, wherein each combined 2D projection includes one or more mask-bordering pixels and one or more mask-edge pixels; forming a three-dimensional (3D) matrix of the set of combined 2D projections; based on the 3D matrix, generating a linear algebraic system for determining pixel values for pixels indicated in a set of 2D projection metal masks, wherein a first change in slope of pixel value associated with a mask-edge pixel of a combined 2D projection is constrained to equal a second change in slope of pixel value associated with a mask-bordering pixel of a combined 2D projection; determining values for a variable vector of the linear algebraic system; and generating a set of inpainted 2D projections by modifying initial 2D projections with values for the variable vector.
1. A computer-implemented method for modifying X-ray projection images of a subject region, the method comprising:
generating a set of combined two-dimensional (2D) projections of a subject region, based on a set of initial 2D projections of the subject region and a set of 2D projection metal masks for the subject region, wherein each combined 2D projection includes one or more mask-bordering pixels and one or more mask-edge pixels;
arranging the set of combined 2D projections sequentially to form a three-dimensional (3D) matrix;
based on the 3D matrix, generating a linear algebraic system for determining pixel values for pixels indicated in the set of 2D projection metal masks, wherein a first change in slope of pixel value associated with a mask-edge pixel of a combined 2D projection is constrained to equal a second change in slope of pixel value associated with a mask-bordering pixel of a combined 2D projection, and wherein the mask-bordering pixel is adjacent to the mask-edge pixel;
determining values for a variable vector of the linear algebraic system, wherein each element of the variable vector represents an inpainted pixel value; and
generating a set of inpainted 2D projections by modifying the initial 2D projections with values for the variable vector.
2. The computer-implemented method of claim 1 , wherein the mask-border pixel is included in a first combined 2D projection in the set of combined 2D projections and the mask-edge pixel is included in a second combined 2D projection in the set of combined 2D projections that is adjacent to the first combined 2D projection in the 3D matrix.
3. The computer-implemented method of claim 1 , wherein the mask-border pixel is included in a first combined 2D projection in the set of combined 2D projections and the mask-edge pixel is included in the first combined 2D projection and is adjacent to the mask-border pixel within the first combined 2D projection.
4. The computer-implemented method of claim 1 , wherein generating the linear algebraic system comprises applying a harmonic function to a domain represented by pixels included in the 3D matrix of the set of combined 2D projections.
5. The computer-implemented method of claim 4 , wherein the pixels included in the 3D matrix form a 3D matrix of image data.
6. The computer-implemented method of claim 4 , wherein generating the linear algebraic system further comprises solving a discretized form of the harmonic function via a finite differences method.
7. The computer-implemented method of claim 1 , wherein determining values for the variable vector of the linear algebraic system comprises computing the values for the variable vector via a conjugate gradient method.
8. The computer-implemented method of claim 1 , wherein generating the set of combined 2D projections of the subject region comprises modifying the initial 2D projections with location information from the set of 2D projection metal masks that indicates one or more mask pixels.
9. The computer-implemented method of claim 1 , wherein the set of 2D projection metal masks includes location information for one or more mask pixels and one or more mask-edge pixels.
10. The computer-implemented method of claim 9 , wherein modifying the initial 2D projections with values for the variable vector comprises applying the values for the variable vector to pixels indicated by the location information.
11. The computer-implemented method of claim 1 , wherein the set of initial 2D projections of the subject region include X-ray projection images of the subject region.
12. The computer-implemented method of claim 1 , further comprising, generating a reconstructed volume based on the set of inpainted 2D projections.
13. The computer-implemented method of claim 1 , further comprising:
generating a non-binary mask based on a final 3D metal object mask and a reconstructed volume of the subject region; and
generating a final reconstructed digital volume of the subject region based on the non-binary mask, the reconstructed volume of the subject region, and a low-artifact reconstructed volume of the subject region.
14. The computer-implemented method of claim 13 , wherein the reconstructed volume of the subject region is generated using the set of initial 2D projections of the subject region.
15. The computer-implemented method of claim 13 , wherein generating the final reconstructed digital volume based on the non-binary mask comprises, for each edge voxel of a metal object in the reconstructed volume, determining a value that is based on an image value from the reconstructed volume, an image value from the low-artifact reconstructed volume, and a value from the non-binary mask.
16. The computer-implemented method of claim 15 , wherein the final 3D metal object mask includes location information associated with each edge voxel of the metal object.
17. The computer-implemented method of claim 13 , further comprising generating the low-artifact reconstructed volume based on low-artifact 2D projections of the subject region.
18. The computer-implemented method of claim 17 , wherein the low-artifact 2D projections are based on the set of inpainted 2D projections.
19. The computer-implemented method of claim 13 , wherein:
the reconstructed volume includes a first visual artifact caused by the presence of a metal object in the reconstructed volume;
the low-artifact reconstructed volume includes a second visual artifact that corresponds to the first artifact; and
the second visual artifact is reduced in visual prominence relative to the first visual artifact.
20. A non-transitory computer readable medium storing instructions that, when executed by a processor, cause the processor to perform the steps of:
generating a set of combined two-dimensional (2D) projections of a subject region, based on a set of initial 2D projections of the subject region and a set of 2D projection metal masks for the subject region, wherein each combined 2D projection includes one or more mask-bordering pixels and one or more mask-edge pixels;
arranging set of combined 2D projections sequentially to form a three-dimensional (3D) matrix;
based on the 3D matrix, generating a linear algebraic system for determining pixel values for pixels indicated in the set of 2D projection metal masks, wherein a first change in slope of pixel value associated with a mask-edge pixel of a combined 2D projection is constrained to equal a second change in slope of pixel value associated with a mask-bordering pixel of a combined 2D projection, and wherein the mask-bordering pixel is adjacent to the mask-edge pixel;
determining values for a variable vector of the linear algebraic system, wherein each element of the variable vector represents an inpainted pixel value; and
generating a set of inpainted 2D projections by modifying the initial 2D projections with values for the variable vector.